Guidebook
Uploaded by
Tejaswi NisanthGuidebook
Uploaded by
Tejaswi NisanthTopic
Mathematics
The Mathematics
of Everyday Life
Course Guidebook
Mohamed Omar
Harvey Mudd College
LEADERSHIP
President & CEO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PAUL SUIJK
Chief Financial Officer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BRUCE G. WILLIS
Chief Marketing Officer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CALE PRITCHETT
SVP, Marketing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JOSEPH PECKL
VP, Customer Engagement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . KONSTANTINE GELFOND
VP, Technology Services. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MARK LEONARD
VP, Product Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JASON SMIGEL
VP, General Counsel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DEBRA STORMS
VP, People . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AUDREY WILLIAMS
Sr. Director, Creative & Production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . KEVIN BARNHILL
Sr. Director, Content Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . KEVIN MANZEL
Director, Business Operations & Planning. . . . . . . . . . . . . . . . . . . . . . . . . GAIL GLEESON
Director, Editorial & Design Services. . . . . . . . . . . . . . . . . . . . . . . . . . . FARHAD HOSSAIN
Director, Content Research & Alternative Programming. . . . . . . . . WILLIAM SCHMIDT
Director, Creative. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OCTAVIA VANNALL
PRODUCTION
Studio Operations Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JIM M. ALLEN
Video Production Director. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ROBERTO DE MORAES
Technical Engineering Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SAL RODRIGUEZ
Quality Assurance Supervisor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JAMIE MCCOMBER
Senior Post-Production Manager. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PETER DWYER
Production Manager. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RIMA KHALEK
Producer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ADAM VOGTMAN
Content Developer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SUSAN DYER
MATTHEW LAING
Associate Producer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JULIET RILEY
Graphics Manager. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . JAMES NIDEL
Graphic Artist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DANIEL RODRIGUEZ
Managing Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OWEN YOUNG
Sr. Editor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ANDREW VOLPE
Editor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ART JARUPHAIBOON
Assistant Editor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHARLES GRAHAM
Audio Engineer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHRIS HOOTH
Director. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ALEX TAFRESHI
Camera Operator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HALEY SAUNDER
Production Assistant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DEVON SEDRICK
EDITORIAL & DESIGN
Sr. Managing Editor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BLAKELY SWAIN
Sr. Graphic Designer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . KATHRYN DAGLEY
Proofreader. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ASHLEY GALLAGHER
Transcript Editor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . WILLIAM DOMANSKI
Mohamed Omar
Harvey Mudd College
Mohamed Omar is an Associate Professor of Mathematics and the Joseph
B. Platt Chair in Effective Teaching at Harvey Mudd College. He earned
his bachelor’s and master’s degrees in Mathematics from the University of
Waterloo and his PhD in Mathematics from the University of California,
Davis. Before arriving at Harvey Mudd College, he was a postdoctoral
fellow at the California Institute of Technology.
Mohamed’s mission is to change the world from math-phobic to math-
loving, and he has fiercely devoted his life to inclusion in mathematics.
In 2018, he won the Mathematical Association of America’s Henry L.
Alder Award, an early-career undergraduate teaching award given to no
more than three educators annually. He has been featured in Forbes and
Scientific American, and he is the author of more than 30 peer-reviewed
articles in internationally recognized journals in his research specialty,
applications of algebra to discrete mathematics.
i
Table of Contents
Introduction
About Mohamed Omar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Guides
1 The Magic of Compound Interest . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 How Mortgages Really Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3 How to Get Out of Credit Card Debt . . . . . . . . . . . . . . . . . . . . . . . 24
4 The Real Cost of Commuting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5 Understanding Your Health Data . . . . . . . . . . . . . . . . . . . . . . . . . 45
6 The Math of Environmental Friendliness . . . . . . . . . . . . . . . . . . 55
7 Getting Wise about Health and Life Insurance . . . . . . . . . . . . . 66
8 Optimizing Your Diet through Math . . . . . . . . . . . . . . . . . . . . . . 76
9 Making Great Estimates with Little Data . . . . . . . . . . . . . . . . . . 86
10 The First-Digit Law and Fraud Spotting . . . . . . . . . . . . . . . . . . . . 99
11 Voter Math and Gerrymandering . . . . . . . . . . . . . . . . . . . . . . . . 110
12 Dividing a Cake or an Inheritance Fairly . . . . . . . . . . . . . . . . . 126
Supplementary Material
Quiz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
ii
The Mathematics of Everyday Life
Math builds critical thinking skills. And it’s needed for advancement in
college programs. But what is math actually good for in everyday life?
This course is dedicated to answering this age-old question by diving into
several facets of daily life and showing how math can help tremendously.
The first four lessons focus on math and your money. These lessons
illustrate the magic of compound interest and how important investing
early is. This serves as the foundation for explaining several financial
aspects of everyday life, starting with an in-depth analysis of mortgages
and how they really work. Many typical concerns with mortgages are
addressed, including how to decide between renting or buying, how to
determine whether to take a 15-year or 30-year mortgage, and what the
true impact of interest rates are. Next comes an analysis of debt, focusing
on credit cards. Then, common problems that arise with credit cards are
discussed, including how to decide whether a credit card is worth having
for the points and what order in which you should pay off debts. Next, the
lifelong costs associated with commuting are assessed, and the pros and
cons of buying an expensive home close to work versus a less expensive
home far away are addressed.
Then, the course investigates how math can help you understand
your health data. Typical biomarkers like BMI and blood pressure are
explained, and you’ll even learn how to use mathematics to navigate
treatment options for ailments.
The next topic is math and environmental friendliness. You’ll discover
how your everyday activities impact your carbon footprint and the
environment in general.
1
Scope
Next, you’ll explore the mathematics of health, starting with life
insurance, where you’ll be mathematically introduced to what insurance
is. The course then dives into case studies, analyzing what factors
influence life insurance choices. Then, you’ll discover how health
insurance functions and how to mathematically make choices between
health insurance options.
Continuing with the mathematics of health, the course then addresses
how to use math to optimize your diet. Whether your goal is weight loss or
weight gain, you’ll learn how to use mathematics to tailor your nutrition
and diet.
The next topic is the mathematics of data, starting with the introduction
of useful estimation tools you can use in your daily life. Then, the course
moves on to the mathematics of fraud detection in a lesson on Benford’s
law, where you’ll learn how it’s possible to detect fraud with the naked
eye and discover how this could have caught one of the largest accounting
scandals in US history.
The course then addresses the role of mathematics in voting
empowerment, with an in-depth mathematical investigation of
gerrymandering. Here, you’ll be exposed to the mathematical tools used
to assess whether political entities manipulate boundaries of electoral
constituencies in their favor.
The course ends with a lesson on sharing fairly. A focal point is a case
study of how to use mathematics to fairly divide assets between heirs who
have competing interests.
2
1
TABLE OF
CONTENTS
The Magic of
Compound
Interest
3
Lesson 1 | The Magic of Compound Interest
O
e of the most powerful financial
n
phenomena in our economic world is
completely governed by mathematics:
compound interest. Understanding
compound interest can help you
tremendously with some of the most
important financial decisions of
your life.
4
Lesson 1 | The Magic of Compound Interest
Daily Double
Suppose you’re given a choice between the following two options:
1 You are given $1,000,000 at the end of a 30-day month.
2 You are given $0.01 on day 1, and every day the amount of money you
had the previous day is doubled.
Which option would you choose?
Let’s say you choose option 2. By day 10, you only have $5.12—which
doesn’t seem very promising.
5
Lesson 1 | The Magic of Compound Interest
But if you keep doubling the amount you started with, by day 30,
you’ll end up with $5,368,709.12! As you can see, the effect of doubling
upon doubling upon doubling is enormous! This is the power of
compounding.
Simple versus
Compound Interest
Let’s say you deposited $1,000 in a bank account. That initial invested
amount is called the principal in this transaction. Giving your money to
the bank is technically giving the bank a loan. They have $1,000 of your
money to use for their daily operations. So you’d hope that the bank
would give you some extra money as a thank you for loaning them that
amount.
6
Lesson 1 | The Magic of Compound Interest
Let’s say the bank promises 5% additionally on top of the principal for
the loan at the end of each year. This means that at the end of the first
year, you would earn 5% of $1,000, which is $50. This is your interest.
And the interest rate—in this case, 5%—is the percentage that you earn
on the principal that you loan the bank.
What should you do with this extra $50? You could take out that $50 and
spend it, and your principal remains at $1,000. At the beginning of next
year, then, you’d be back to a $1,000 principal, and earning 5%, you’d
earn another $50.
If this continued for 10 years, how much profit would you earn at the
end of the process?
Each year, you would earn $50 of profit because you always start with
a $1,000 principal. So at the end of 10 years, you’d make 10 times that
amount, or $500.
A formula emerges for calculating the total profit you earn through this
process: The principal (in this case, $1,000) multiplied by the rate (in
this case, 5%, or 0.05) multiplied by the amount of time invested (in this
case, 10 years) gives you a total of $500 profit:
1,000 × 0.05 × 10 = 500
So at the end of 10 years, you would have a total of $1,500.
This phenomenon—earning a specific interest amount based on your
principal and your rate—is known as simple interest.
In the case of simple interest, you took your $50 annual profit and
withdrew it to spend on whatever you wanted. But what if you instead
left that $50 in the bank account to grow even more?
7
Lesson 1 | The Magic of Compound Interest
Reinvesting your profits is the backbone of compound interest. Instead
of earning on the same principal each year, you earn interest on your
principal and on top of your interest, so your principal itself constantly
increases!
Say you deposited $1,000 into a bank account like you did before. At the
end of the first year, you’d have a profit of $50, for a total of $1,050. But
this time, if you decided to leave the $50 in the bank account, how much
money would you have at the end of year two?
You would earn 5% of the $1,050:
0.05 × 1,050 = 52.50
That’s an extra $2.50 more than you would have earned than if you
spent the $50 profit from year one.
The next year, your principal is now $1,102.50, because you are
reinvesting the extra $52.50. Earning 5% on that gives you a profit of
about $55:
0.05 × 1,102.50 = 55.125
So how much money would you have after 10 years? Let’s revisit the
calculations to make things a bit simpler.
You started with $1,000 invested. Your profit is 5%, so at the end of year
one, you will have 105% of your principal (100% being the principal
itself and 5% being the profit). So the total amount you would have at
the end of year one is $1,050:
1.05 × 1,000 = 1,050
8
Lesson 1 | The Magic of Compound Interest
This amount, 1.05 × 1,000, is your new principal. The next year, you
again will be left with 105% of the new principal at the end of the year
(100% of the new principal itself and 5% profit on that new principal).
So your total newer principal is 1.05 times your principal at the
beginning of the year, which was itself 1.05 × 1,000. So your principal at
the end of year two is this:
1.05 × (1.05 × 1,000) = 1.052 × 1,000
Continuing this process, at the end of 10 years, your total amount will
be almost $1,630:
1.0510 × 1,000 = 1,628.89
In the simple interest scenario, where you spent your profits each year,
your total earnings were $1,500 instead. That’s quite a difference!
This is again because with compound interest, you earn interest on
previously earned interest, and this inflates the amount you earn in
total. This is the power of compound interest! Making interest off
extra interest has a multiplicative effect that makes your money go a
long way.
To see how dramatic this effect can be, let’s imagine that instead of
investing for 10 years, you invested for 20 years.
With simple interest, you would earn 5% each year on your principal,
giving you a total of $1,000 of profit (20 × 1,000 × 0.05). So you’d end up
with $2,000 in the end.
With compound interest, you would end up with $2,653.30. That’s a
massive difference!
1,000 × 1.0520 = 2,653.30
9
Lesson 1 | The Magic of Compound Interest
When you invested for 10 years, the difference you made by not
spending your profits each year—in other words, when you built
compound interest—was $128.89. But when you invested for 20 years,
the difference in profit was $653.30. That’s five times more profit than
investing for 10 years!
So compounding over a long period of time has a drastic effect on your
outcome.
To make this mathematical, let’s say that P is the principal ($1,000), r
is the interest rate as a decimal (0.05), and t is the amount of time you
invest in years (10). To calculate the total principal at the end of 10 years
of investment, you went through the following process: At the end of
year one, you still have your original principal, P = 1,000, and your
profit is 0.05 × 1,000, or r × P. This gives you a total of P, your principal,
plus r × P, your interest:
1,000 + (0.05 × 1,000)
P + (r × P)
To simplify this, you can factor out the P to get (1 + r) × P. So your total
after year one is 1.05 × 1,000.
At the beginning of year two, your principal—the original amount you
have—is (1 + r) × P, which is 1.05 × 1,000:
(1 + r) × P
1.05 × 1,000
You earn another 105% on this total, giving you the following at the end
of year two:
(1 + r) × (1 + r) × P, or
(1 + r)2 × P
10
Lesson 1 | The Magic of Compound Interest
Following this pattern, at the end of t years, you have (1 + r)t × P dollars:
(1 + r)3 × P
(1 + r)4 × P
(1 + r)t × P
This is called the compound interest formula.
Compound interest has a staggering
effect if you have time on your side.
This is why the best investment
advice is to invest early.
How Much Should You
Save for Retirement?
Saving money for retirement is an essential aspect of your financial
life. Let’s investigate how compound interest affects your nest egg.
Suppose you have $100,000 in a retirement account at age 35. The S&P
500, an index fund tracking major companies in the United States,
averaged an annual return of roughly 8% from 1957 to 2018. If this is the
case, how much would your $100,000 nest egg be expected to be worth
when you are 65 years old?
11
Lesson 1 | The Magic of Compound Interest
The principal in your investment, P, is $100,000, and the interest rate, r,
is 0.08. You turn 65 in 30 years, so the time invested, t, is 30. According
to the compound interest formula, your expected investment value at 65
would be more than a million dollars!
P × (1 + r)t = 100,000 × (1 + 0.08)30 = 1,006,265.69
That’s a significant increase—all because of compound interest.
People often rely on retirement calculators online to figure out how
much they should save for retirement. But different calculators give
different answers. Instead, you can empower yourself to be able to
calculate this on your own so that you can make informed decisions
based on your individual circumstances.
If you take into consideration things like inflation and the fluctuation
of the average from year to year, to be very conservative, you can
assume a 4% annual rate of return.1 This is a rate that many financial
forecasters use.
Let’s say you’re 35 years old and hoping to retire at 65 years old. At that
age, you’d like to live comfortably off your retirement nest egg, so you
don’t want to make a dent on the principal. Since your assumed annual
interest rate is 4%, this would mean that you would live off the 4%
return we get from our nest egg each year. Let’s also assume that the
amount you have for retirement right now is $100,000. What will this
turn into at retirement?
1 Many investment forecast tools assume an average rate of return. But
averages don’t tell the full picture. The markets fluctuate year by year, and
the degree of this fluctuation impacts the final return on your investment. So
you need to be more conservative in your estimates than what investment
calculators predict for you.
12
Lesson 1 | The Magic of Compound Interest
Your principal (P) is 100,000; rate (r) is 4%, or 0.04; and time (t) is your
retirement age of 65 minus your current age of 35, so 30 years. So your
total at retirement will be almost $325,000:
P(1 + r)t = 100,000 × (1.04)30 = 324,339.75
But you haven’t taken into consideration what you invest every month
from the time you are 35 to the time you are 65. To account for this
money, you can use the following general formula:
(1 + 𝑟𝑟)! # $ − 1
𝑆𝑆 = 𝑃𝑃 × ,
𝑟𝑟
where P is the amount you contribute each month, r is the monthly
interest rate, and n is the number of months you will be investing. This
formula tells you the amount—or sum, S—you’ll have when you put in
a specific amount of money each month at a specific monthly interest
rate over a period of months.
If you add together these monthly contributions to the amount you
made off your original $100,000 in your retirement, you’ll get a total
retirement amount. You can play with these formulas to get a feel
for how much you should be saving each month to reach your goal
retirement amount.
Reading
Thomsett and Thomsett, “The Mathematics of Investing.”
US Department of Agriculture, “Calculators and Counters.”
US Securities and Exchange Commission, “Compound Interest
Calculator.”
13
2
TABLE OF
CONTENTS
How Mortgages
Really Work
14
Lesson 2 | How Mortgages Really Work
T
ere are many mortgage calculators
h
available online, but what’s really
going on behind the mysterious
mathematics of mortgages? And
how can you adapt these calculators
to your own specific situations and
make informed financial decisions?
Understanding how these calculators
work has so many benefits. It can
allow you to understand the true
impact of fluctuations in mortgage
costs based on term length and
interest rate, and it can empower
you to determine how fluctuations in
mortgage terms that aren’t covered by
online calculators actually affect your
mortgage.
15
Lesson 2 | How Mortgages Really Work
The Basics of
Mortgage Calculation
How much will your monthly payment on a mortgage be? Let’s develop
a formula that will calculate this based on some key data:
{ the amount you borrow from the bank, P (the principal);
{ the monthly interest rate of the mortgage, r; and
{ the life of the loan, n (15 versus 30 years, or 180 versus 360 months).
Let’s say you have a $200,000 mortgage, so P is 200,000; an annual
interest rate of 6%, so your monthly interest rate r is 0.06/12, or 0.005; and
a loan of 30 years, so n is 360.1
Let’s look at what happens in the first few months of the loan. You start
with P = 200,000, and after the first month, you owe interest of 0.005
times that number, so your loan has an amount of 200,000 + (200,000 ×
0.005), or 200,000 × 1.005:
200,000 + (200,000 × 0.005) = 200,000(1.005)
You’re trying to determine your monthly payment, so let’s give it a
variable of M for monthly payment. You pay M on the total after one
month, so at the end of the first month, you have a mortgage balance of
200,000(1.005) − M.
1 The interest on mortgages are sometimes calculated daily instead of
monthly.
16
Lesson 2 | How Mortgages Really Work
Now you apply the process again. Your interest now is based on
this new mortgage balance. You pay an interest of r% on it, so your
mortgage jumps to
(200,000(1.005) − M) × (1.005),
which is the quantity you had after the first month, or 200,000(1.005) −
M, multiplied by the additional interest of 1.005.
But you pay down the mortgage by your monthly payment M. This gives
you a balance at the end of month two of what looks like this:
(200,000(1.005) − M) × (1.005) − M
= 200,000(1.005)2 − M(1.005) − M
You start with your mortgage amount at the end of month one, multiply
it by 1.005 to account for the interest that month, and then subtract your
monthly payment of M. This gives you a total of
200,000(1.005)2 − M(1.005) − M.
When you do this again, month after month, you start to notice a
pattern. And after 360 months, you get the following balance:
200,000(1.005)360 − M(1.005)359 − M(1.005)358 − … − M(1 + r)1 − M
This is a hefty formula—but luckily it can be simplified, to something
like this:
(1.005)!"# − 1
200,000(1.005)!"# − 𝑀𝑀
0.005
This expression is the balance on your mortgage after 360 months.
17
Lesson 2 | How Mortgages Really Work
But how much should your mortgage payment, M, be?
At the end of the lifetime of your mortgage, you should end up with an
outstanding loan of $0. So your monthly mortgage balance (M) satisfies
the following equation:
(1.005)!"# − 1
200,000(1.005)!"# − 𝑀𝑀 =0
0.005
If you rearrange for M, you get the following:
200,000 × 0.005(1.005)!"#
M= = $1,199.10
(1.005)!"# − 1
So in this scenario, your monthly mortgage payment would be roughly
$1,200.
You can now use this formula to compute the monthly payment on a
mortgage. If your principal of $200,000 is replaced with a general
principal P, your monthly interest rate of 0.5% is replaced with r, and
your mortgage time of 360 months (or 30 years) is replaced with an n,
then you’d have the following general formula for the monthly payment
on your mortgage:
𝑃𝑃𝑃𝑃(1 + 𝑟𝑟)!
𝑀𝑀 =
(1 + 𝑟𝑟)! − 1
18
Lesson 2 | How Mortgages Really Work
And because of this, you can start to get some insight on how different
parameters, such as mortgage interest rate2 and mortgage term length,
affect the amount you pay on the life of the loan.
This monthly mortgage payment formula can also be used to figure out
how much home you can afford given your desired monthly payment.
To figure out what your initial mortgage principal would be, you can
plug in numbers for M, r, and n in the formula and determine P by
rearranging the equation to solve for P.
And the best part about this is you don’t have to consult a mortgage
calculator to figure this out!
15-Year versus 30-Year
Mortgages
Deciding the life of a mortgage—typically deciding between a 15-year
and 30-year mortgage—is difficult. A 30-year mortgage usually results
in lower monthly payments so that you can afford more home, but a
15-year mortgage pays off the loan more quickly and reduces the total
interest you pay.
But how much would you actually save by going with the 15-year over
the 30-year mortgage? You can use the monthly mortgage payment
formula to get a sense of how much of a difference it makes to have one
versus the other.
2 A small change in interest rate can have a drastic effect on the amount you
pay over the life of your mortgage.
19
Lesson 2 | How Mortgages Really Work
Using the same example numbers—a mortgage loan of $200,000 at an
annual interest rate of 6%—it turns out that a 15-year mortgage at a 6%
annual interest has about the same return as a total mortgage payment
than a 30-year mortgage at 3% annual interest rate in the long run.
But there’s a catch: 15-year mortgages carry a much higher monthly
payment, and that might not be tolerable for a household. It’s difficult
to make the right decision, but at least you are now empowered to be
able to make the decision in an informed way.
So in this example, you’ve figured out the savings you get on a 15-year
versus a 30-year mortgage if you assume an annual interest rate of 6%.
But what about a general interest rate r? How do the total payments
compare between the two types of mortgages?
20
Lesson 2 | How Mortgages Really Work
Let’s say you have an interest rate of r and your mortgage amount is P.
Using the monthly mortgage payment formula, on a 15-year mortgage,
you have the following monthly payment:
𝑃𝑃𝑃𝑃(1 + 𝑟𝑟𝑟!"#
(1 + 𝑟𝑟𝑟!"# − 1
And on a 30-year mortgage, you have the following monthly payment:
𝑃𝑃𝑃𝑃(1 + 𝑟𝑟𝑟!"#
(1 + 𝑟𝑟𝑟!"# − 1
So the total mortgage payments on the entire life of the 15-year and
30-year mortgages are the following:
𝑃𝑃𝑃𝑃(1 + 𝑟𝑟𝑟180
180 ×
(1 + 𝑟𝑟𝑟180 − 1
𝑃𝑃𝑃𝑃(1 + 𝑟𝑟𝑟360
360 ×
(1 + 𝑟𝑟𝑟360 − 1
)
After doing some algebra, you get the ratio between the total loan
amount on a 15-year mortgage versus a 30-year mortgage:
1
2
1
× 1 + !"#
𝑥𝑥
)
21
Lesson 2 | How Mortgages Really Work
Let’s then look at the difference between these two mortgage terms.
{ If you had a 6% annual interest rate, the formula would give you a
ratio that is roughly 0.7. So the total amount you pay on a 15-year
mortgage is a little more than 2/3 the amount of a 30-year.
{ If you had a 12% annual interest rate, the formula would give you a
ratio that is roughly 0.58. So the total amount you pay on a 15-year
mortgage is a little more than half the amount of a 30-year.
22
Lesson 2 | How Mortgages Really Work
In both of these scenarios, that’s a huge savings, considering that
mortgages are hundreds of thousands of dollars! So in general, a
15-year mortgage is a very smart move for a mortgage if the monthly
payment is manageable.
Reading
Moneysmart.gov, https://moneysmart.gov.au/#tools-and-resources.
NerdWallet.com, “Mortgage Calculator.”
Thomsett and Thomsett, “The Mathematics of Investing.”
US Department of Transportation, “Ride Sharing Programs.”
US Securities and Exchange Commission, “Compound Interest
Calculator.”
23
3
TABLE OF
CONTENTS
How to Get
Out of Credit
Card Debt
24
Lesson 3 | How to Get Out of Credit Card Debt
D
bt, in its many forms—from business
e
loans, to student loans, to credit
cards and other commercial credit
products—is a major contributor to
the backbone of our society. From
a mathematical viewpoint, how
much do different debt instruments
actually cost? Are they really worth
it? The answers depend on many
factors, but the goal of this lesson is
to empower you, regardless of the
situation you are in, to feel confident
in addressing them.
25
Lesson 3 | How to Get Out of Credit Card Debt
Credit cards are commonplace in our
society, and the credit card industry earns
quite a chunk of capital. In 2019 alone,
VISA’s revenue was more than $23 billion.
Mathematical Underpinnings
of Credit Cards
Credit cards give you access to a specific amount of money that you can
borrow from a credit card company, bank, or general lender.
The money borrowed from the company must be paid back, subject to
certain time restrictions, but you pay a premium for having borrowed
money in the first place.
Credit cards usually quote an APR (annual percentage rate), which
is the fee for using the credit card, including the interest you pay the
bank for borrowing specific amounts of money, together with some
additional fees.1 However, on credit cards, these fees are usually not
substantive. The bulk of the payment is in the interest.
1 In addition to APR, many credit cards have an annual fee that is paid once
every year.
26
Lesson 3 | How to Get Out of Credit Card Debt
Suppose a credit card has an 18% APR. This means you are charged at
an annual percentage rate of 18%. But the balance on your credit card
fluctuates day to day. So how are you charged each day?
The amount of interest you are charged for a specific day is the APR
divided by the number of days in the year, which can be assumed to
be 365. 2
daily interest % = APR/365
Suppose you had an outstanding balance of $1,000 on a credit card one
day. If your APR is 18%, that day your interest charge would be 0.18 (the
interest rate) divided by 365 (the number of days in a year) times $1,000,
for a total of about 49 cents.
APR 1,000 × 0.18
= = 0.49
365 365
This doesn’t seem like much, but things add up day to day, especially as
outstanding balances accrue. How much interest you end up paying on
a credit card depends on the balances you keep.
One of the difficulties with credit cards in general is that the amount
you owe increases rapidly because of compound interest. The annual
interest rate of 18% is very high but quite typical of credit cards.
Compare that with current mortgage rates and you see a significant
difference. Moreover, with credit cards, you aren’t left with an asset
like you are when you take out mortgages.
2 Note that this ignores leap years.
27
Lesson 3 | How to Get Out of Credit Card Debt
You can see where even more dire issues come in with this. Here, it was
assumed that no money was spent beyond the $1,000 borrowed. But it is
common to slip deeper and deeper into credit card debt. As this occurs,
you increase your amount owed and have an even heftier and heftier
charge.
So what can you do to mitigate these charges? One option is to stay clear
of credit cards. Another option is to pay off credit cards early enough,
during the interest grace period,3 so that debt doesn’t accrue.
Paying Off Credit Card Debt
How should someone who has accrued lots of credit card debt attack
their credit cards to pay them off?
There are two factors to consider: one is mathematical, and another is
motivational. What may work best mathematically for clearing debt
may not be the best for maintaining personal motivation to stick to
paying off the debt.
Suppose you have two credit cards that you want to pay off: One has a
balance of $2,000 at an APR of 20% and another has a balance of $10,000
at an APR of 12%.
What is mathematically the best way to pay these off? This is a real-life
question that mathematics can help with and that online calculators
can’t answer directly.
3 Many credit cards have a grace period before interest is actually charged on
them. The period varies wildly depending on the credit card but definitely
influences how much interest is charged.
28
Lesson 3 | How to Get Out of Credit Card Debt
The first thing you should do is pay the minimum required payment on
each credit card. Credit card companies usually require a minimum
payment on a credit card as an enforcement that you are paying toward
the balance.
Minimum payments are usually based on a percentage of the balance
owed—typically somewhere between 1% and 5%. If minimum payments
are not paid, there are penalty fees that are so heavy that interest owed
on the cards typically pale in comparison.
But ignoring minimum payments and focusing on the balances and
owed amounts, the mathematical strategy is to attack the debt with the
highest interest rate as much as possible. That way, the total interest
you pay on the combined debt is less and less over time.
But paying off credit card debt is much more than a mathematical
phenomenon. People’s emotions and circumstances play important
roles in decision-making. Sometimes, even though the mathematical
way to go is financially optimal, sticking with methods that are more
motivational might work better in eliminating the debt.
This type of decision might
be fruitful for students in a There are many online
scenario like the following:
calculators that
having $2,000 in credit
card debt and a $40,000 can calculate credit
student loan. Paying off the card payments for
small debt quickly provides
you, but it’s helpful
motivation to continue paying
aggressively, even though the to understand how
mathematically sound way to appropriately
to go is to pay off the $40,000
complete online fields
loan first.
and how the fields
affect the outcome.
29
Lesson 3 | How to Get Out of Credit Card Debt
The Financial Side of College
Student loans play a huge role in the economics of the United States. As
of 2020, roughly 10% of the nation’s debt is in federal student loans—a
steep rise even from a decade ago. With the rising costs of tuition,
students have graduated with more than twice as much debt as students
did even 15 years ago on average.
What are the true costs of student loans over a long period? And are
student loans really worth it? It is very difficult to give a one-size-fits-all
answer to the latter question. There are many factors that mathematics
can’t account for.
One example is the opportunity potential. Certain institutions may
provide students with opportunities that vastly outweigh the debt
accrued for attending the institution. It’s very difficult to assess to
what extent such opportunities are likely to come to fruition for any
particular student.
However, it is worth assessing the mathematics of the situation. This
is especially true for federal student loans, which are in general not
forgivable in a bankruptcy.
There are many financial vehicles for offering student loans, but
subsidized federal student loans—which are given to those who
demonstrate financial need, based on responses to the Free Application
for Federal Student Aid (FAFSA)—are the most generous of the loans,
and also quite ubiquitous.
30
Lesson 3 | How to Get Out of Credit Card Debt
Comparing Student Loan Types
There are many financial vehicles for offering student loans, but here are
the three common ones:
Subsidized federal student loans are given to those who demonstrate
financial need based on responses to the FAFSA. The government pays
your interest while students are in school; interest starts accruing and
loan payments start six months after graduation.
Unsubsidized federal student loans are given to students who do not
demonstrate financial need according to the FAFSA. Loan payments
don’t start until six months after graduation, but the government does
not cover interest, which starts building up as soon as the funds are
dispersed to the student in school.
Private student loans are given directly from financial institutions,
generally at higher interest rates than federal student loans, and must
be paid while a student is in school.
INTEREST STARTS PAYMENT BEGINS
ACCRUING
SUBSIDIZED 6 months after 6 months after
FEDERAL LOAN graduation graduation
UNSUBSIDIZED immediately 6 months after
FEDERAL LOAN graduation
PRIVATE LOAN immediately immediately
31
Lesson 3 | How to Get Out of Credit Card Debt
With this type of loan, the government pays your interest while you are
still enrolled in school.4 After graduation, you are given a six-month
grace period, after which interest will start accruing and you will have
to start making loan payments.
Suppose that you took out a total of $40,000 in federal student loans
through your college career. What will the cost of such a loan look like,
assuming an annual interest rate of 5%?
Interest on student loans is calculated just like credit cards, in that
interest is charged daily and compounded based on what you owe.
So, for any given day in this scenario, you are charged 5/365% interest
each day of the loan.
As with credit cards, this type of compounding adds up very quickly.
The difference here, though, is that the principal changes over time
because you are required to pay off the loan by a certain period. This
is reminiscent of what happens with mortgage payments. You have a
term in which the mortgage needs to be paid, so as you pay monthly,
the principal on the mortgage decreases, which then means the interest
decreases.
As you learned in the previous lesson, the monthly payment of a
mortgage that has a principal of P dollars and an interest rate of r
compounded monthly that’s paid over 30 years is the following:
(1 + 𝑟𝑟)!"#
𝑀𝑀 = 𝑃𝑃𝑃𝑃 ×
(1 + 𝑟𝑟)!"# − 1
4 Certain enrollment requirements for certain institutions must be met.
32
Lesson 3 | How to Get Out of Credit Card Debt
The 360 here was obtained by calculating the number of compounded
months. For student loans, the interest is compounded daily, so for
a typical payment period for a student loan, which is 10 years, the
number of days of compounding is 10 × 365. So the formula for the
amount you pay daily on a 10-year student loan with a daily interest rate
of r and principal P is the following:
(1 + 𝑟𝑟)3650
D = 𝑃𝑃𝑃𝑃 ×
(1 + 𝑟𝑟)3650 − 1
If you plug in P = 40,000, your principal, and r = 0.05/365, your daily
interest rate, this gives you the daily owing amount. And if you multiply
by 30, this will give you the monthly payment, which works out to about
$420 per month. To get a quick estimate, you can also use a student loan
calculator online.
Paying $420 per month for 10 years is a hefty amount. The total
payment would be $50,136, which is $10,136 in interest alone!
But more importantly, that $420 per month for 10 years could have gone
into a retirement account and generated interest over several decades
afterward.
So taking out a student loan doesn’t just cost you the significant
interest, but it takes away your potential to invest early, which has a
staggering effect on your finances at retirement. Student loans that are
especially large result in issues with being able to save for investments
like a down payment for a home.
33
Lesson 3 | How to Get Out of Credit Card Debt
For many individuals, going to college is one of the most important
investments in their lives. If you know that you are going to take out loans,
here are just a few strategies, informed by mathematics, you can employ:
Pay aggressively early. The more you pay on the principal early on, the
less principal you will have, and the less interest will accrue. This
results in a lower overall payment for the loan, and a quicker time
period to get out of the loan.
Aggressively apply for grants or scholarships rather than loans. Spending
a significant time before college searching swaths of scholarship
databases may be worth the time based on the amounts that the
scholarships render.
Deeply consider the return on investment for the program of consideration.
Asking the institution for average early- and mid-career salaries of
graduates of the given program and comparing this data with other
programs with similar credentials can help inform the extent to
which your particular college and program of choice truly provide
opportunities for a sound financial future.
Reading
Moneysmart.gov, https://moneysmart.gov.au/#tools-and-resources.
Thomsett and Thomsett, “The Mathematics of Investing.”
US Securities and Exchange Commission, “Compound Interest
Calculator.”
34
4
TABLE OF
CONTENTS
The Real Cost
of Commuting
35
Lesson 4 | The Real Cost of Commuting
H
w much money do you truly save by
o
buying a more affordable home away
from work? What is the toll on your
time in the long run of commuting?
What are the benefits of taking public
transportation if it’s an option? All of
these questions can be tackled with
the help of mathematics.
36
Lesson 4 | The Real Cost of Commuting
According to the US Census Bureau,
in 2013 86% of all workers drove to
work by private vehicle, only 9.4% of
which carpooled. This is a significant
change from 1980, when about 2/3
of the population commuted alone
and only about 1/5 carpooled.
37
Lesson 4 | The Real Cost of Commuting
The Mathematics behind
Commuting Costs
A common narrative in American culture is that living in a less
expensive neighborhood that requires a somewhat significant commute
is financially more sound than living closer to work. But is this
actually true?
It’s hard to tell in general, and everyone’s life situation is different.
In some cases, the option might not even be possible; you might be
completely priced out of living closer to work.
More importantly, there are aspects of life that matter to people just as
much or more than the difference in cost. The community your family
lives in, school districts for children, and quality of life are just a few
such examples.
The biggest driver for misconceptions that people have about the cost
of commuting is that they tend to think of their estimated weekly,
biweekly, or monthly costs instead of the lifetime costs of commuting.
Imagine that your household has two income earners and two children.
You have the option of buying either of the following:
1 a $320,000 home that
Home purchases are
is a 30-mile commute
from work or embedded in this process,
but much of this analysis
2 a $480,000 home that
can be carried over
is a 5-mile commute
from work. to the scenario where
a family rents for an
extended period of time.
38
Lesson 4 | The Real Cost of Commuting
Let’s assume that the mortgages on both houses have a 6% annual
interest rate, which works out to a 0.5% monthly interest rate. If
you plug this into the monthly mortgage payment calculator from
lesson 2, the monthly payments on the houses are $1,918.56 on home 1
and $2,877.84 on home 2. This results in a total house payment of
$690,682.20 on home 1 and $1,036,023.30 for home 2. That’s a substantial
difference!
COMMUTE MORTGAGE INTEREST MONTHLY TOTAL
PAYMENT PAYMENT
HOME 30 miles 320,000 6% annual $1,918.56 $690,682.20
1 0.5% monthly
HOME 5 miles 480,000 6% annual $2,877.84 $1,036,023.30
2 0.5% monthly
Another huge cost is the cost of commuting. Let’s start conservatively
by assuming that the couple in the household travel together to work
in a midsized sedan and do not need to worry about transportation for
their children to and from school. How many miles of commuting will
the couple complete in the 30-year period?
Let’s use 240 days as an estimate for the number of days a year they
commute.1 If the couple chooses home 1, they commute 60 miles per
day: 30 to work and 30 home. If they choose home 2, they commute 10
miles per day.
1 Many jobs have many different structures and vacation times. Some jobs
allow for working from home at different periods. You’d need to adjust for
your specific situation.
39
Lesson 4 | The Real Cost of Commuting
So the total number of miles the couple spends commuting in a year is
14,400 for home 1 and 2,400 for home 2. What is the difference in fuel
cost from commuting over the 30-year period?
Let’s assume an average fuel efficiency of 27 miles per gallon.2 In a
year, commuting would require 533.3 gallons of gas for home 1 and 88.9
gallons for home 2.
To estimate gas prices,3 let’s take the 2019 average gas rate, $2.60 per
gallon, and apply an average inflation of 2% to it over a 30-year period.
Gas in the first year will cost $2.60. In the second year, it will cost
2.60 × 1.02; in the third, it will cost 2.60 × 1.022; and continuing this, in
the 30 th year, it will cost 2.60 × 1.0229, or about $4.60. So the average rate
will be the average of these 30 numbers.
This works out to about $3.50 per gallon in an average year. So, over
a 30-year span, commuting costs due to gas if home 1 is purchased
will add up to $55,996.50 and to $9,334.50 if home 2 is chosen. This is
a savings of more than $45,000 on gas alone if home 2 is bought over
home 1.
However, the total mortgage cost for home 1 is roughly $350,000 less
than that of home 2. So you might think the gas costs in commuting are
not nearly enough to make home 2 a worthy purchase.
But there are a few things to consider in this situation. First, there
are many other costs associated with owning a car besides gas.
Maintenance needs to be done on cars. And as cars get used over time,
they need to be replaced.
2 Currently in the United States, the average is 24.9, but cars tend to become
more efficient over time, so 27 is conservative.
3 This is actually quite difficult because oil prices fluctuate drastically due to
environmental and geopolitical reasons.
40
Lesson 4 | The Real Cost of Commuting
Maintenance and
Replacement of Cars
AAA Newsroom estimates an annual cost of $1,350 for maintenance
per 15,000 miles driven.4 If you purchase home 1, you’d be commuting
14,400 miles per year, and if you purchase home 2, you’d be commuting
2,400 miles per year, a difference of 12,000 miles.
So each year, your maintenance costs would likely estimate to an
additional $1,080 per year.
12,000
× $1,350 = $1,080
15,000
That’s a difference of more than $32,000 in maintenance costs! And this
doesn’t even take inflation into consideration.
Another maintenance cost that you should consider is purchasing
vehicles. According to Consumer Reports, the average lifespan of a
vehicle in the United States is 150,000 miles. Purchasing home 1 adds
an additional 12,000 miles of driving per year, or 360,000 miles. This
amounts to about two to three more cars needed over the 30-year span
if buying home 1 instead of home 2.
This adds an additional cost of between $30,000 and $50,000 in the most
generous of circumstances, not even considering that if you decide to
buy a cheaper vehicle, the maintenance costs might increase more.
4 AAA Newsroom estimates maintenance and repair costs at an average of
about 9 cents per mile.
41
Lesson 4 | The Real Cost of Commuting
So taking into consideration the additional maintenance costs and
car-buying costs due to commuting, the roughly $350,000 savings in
mortgage payments from buying home 1 is reduced by almost $100,000
to make a total ballpark savings of $250,000.
It seems to be the case, then, that buying home 1 is still the best
financial bet overall.
And there are some potential additional advantages that home 1 has—
the most direct of which is the extra cash availability to the family
every month over the 30-year period. If that money is invested in the
stock market over the 30-year period, it has the potential to earn a
significant swath of cash, meaning a significant bonus for home 1!
The mortgage savings might also be used to invest in a college
fund. This is a fantastic family investment that offers children an
opportunity that might not be possible if home 2, with the larger
mortgage, is opted for.
But accounting solely for car maintenance, mileage, and ownership
costs, home 2 is by far the better choice, with its short 5-mile commute,
as opposed to the 30-mile commute for home 1.
And there are other advantages of home 2 that aren’t as apparent at
face value. The first is that if two cars are used for commuting instead
of just one, then all the additional commuting costs associated with
the purchase of home 1 are now doubled! This makes the advantage for
home 2 more stark.
Furthermore, a likely scenario here is that home 1 is a suburban home,
whereas home 2 is closer to the heart of a city. As a consequence,
home 2 might have a much higher chance of appreciating rapidly.
42
Lesson 4 | The Real Cost of Commuting
The Time Cost of Commuting
One very important aspect of commuting that has been omitted so far
is the time cost. How much time in our lives is spent driving in our
vehicles? Let’s compare between home 1 and home 2.
Home 1 requires 14,400 annual miles of commuting, and home 2 is
a commute of 2,400 miles per year, for a difference of 12,000 miles.
Even if this mileage is driven at a high speed of 65 miles per hour,
eliminating traffic as a factor (which in reality it is), this is at the very
least an extra 184.6 hours of driving per year. Over the span of 30 years,
this is an extra 230.8 continuous days driving—almost an additional
eight continuous months of your life driving!
This is a significant time cost that many people overlook, and it’s a
very important one. Eight months is a significant chunk of one’s life,
and this is a very conservative estimate. The reality is more likely
somewhere between 12 and 18 months of continuous additional time
spent solely on commuting. This is time that could have been spent
doing other things, such as spending time with family, focusing on
health and fitness, or engaging in personal interests.
There are many situations that can influence whether you buy one
home or another. While a concrete decision hasn’t been made here
between home 1 or home 2, the mathematics have been presented
to help you justify the decision for yourself, based on your specific
situation.
43
Lesson 4 | The Real Cost of Commuting
Strategies for Saving
Let’s say that after a careful analysis, you make the decision to buy a home
with a longer commute. What can you do to ease your commuting costs?
Vehicle choice: Select vehicles that generally require less maintenance
and consume less, such as energy-efficient vehicles.
Carpooling: Commuting to work with a group of people cuts costs
dramatically. On top of the sheer savings from not driving your car
as often or as far, many companies and organizations offer financial
incentives or benefits for commuting.
Public transportation: Depending on your specific life situation, city,
family setup, and geography of home and work, public transportation
might be a reasonable option to consider for a commute.
Reading
McKenzie, “Who Drives to Work?”
Stine, “AIDS Update 1999.”
US Department of Energy, “Find and Compare Cars.”
44
5
TABLE OF
CONTENTS
Understanding
Your Health Data
45
Lesson 5 | Understanding Your Health Data
H
alth is a pivotal aspect of life, and it
e
is important that you take charge of
the health issues you face. But you’re
often bombarded with so much data
and jargon that it’s hard to discern
what’s correct from what’s incorrect.
This lesson will empower you to
look at health data with a fresh lens,
with the goal of understanding what
various aspects of health data really
mean so that you can make more
informed decisions.
46
Lesson 5 | Understanding Your Health Data
Body Mass Index
Body mass index (BMI) is a quantity that compares your height and
your weight. BMI is meant to tell you whether you are, by a doctor’s
consideration, in one of the following categories:
{ underweight, which is a BMI less than 18.5;
{ normal weight, which is a BMI between 18.5 and 24.9;
{ overweight, which is a BMI between 25 and 29.9; or
{ obese, which is a BMI of 30 or more.
There are plenty of apps online that you can use to calculate your BMI,
but what’s behind those calculations? Here is the formula that
calculates your BMI:
(weight in pounds)
BMI = 703 ×
(height in inches)2
According to doctors, a high BMI is an indication of obesity, whereas a
very low BMI suggests that an individual is underweight. But the issue
is this: What is actually being measured when BMI is calculated is
based only on a person’s height and weight.
Consider, for example, bodybuilders, who are typically lean and have
visible musculature. Bodybuilders likely eat well-balanced diets,
with healthy vegetables and lean sources of protein to maintain
muscle mass.
According to a doctor’s order, the BMI of a bodybuilder might be
considered overweight. But their physical condition doesn’t match with
the health complications and images you might associate with being
overweight.
47
Lesson 5 | Understanding Your Health Data
Lean body mass and fat percentage are examples of the myriad issues
with using BMI to measure health indicators.
So why do doctors use BMI? For the general population, it works well as
an indicator. But keep in mind that the only input in the BMI formula
is height and weight and that the content of your weight—fatty tissue or
muscle mass—matters a lot in using BMI as an indicator for health issues.
Blood Pressure
Blood pressure measurements are usually listed as two numbers:
{ systolic blood pressure, which indicates the pressure your blood
exerts on your artery walls during the time that your heart is
beating; and
{ diastolic blood pressure, which indicates the pressure your blood
exerts on your artery walls during the time that your heart is resting
between beats.
Both of these units of pressure are measured in millimeters of mercury
(mm Hg).1 The higher the millimeters of mercury, the more pressure is
exerted on your artery walls.
The systolic pressure is typically much higher than the diastolic
pressure. This makes sense because the pressure exerted when the
heart is beating is higher than when the heart is resting.
1 The name for this unit comes from the fact that it stands for the amount
of pressure the base of a tube of mercury 1 millimeter high exerts when
subjected to the force of gravity.
48
Lesson 5 | Understanding Your Health Data
A blood pressure reading of 120 mm Hg/80 mm Hg or slightly lower is
considered normal, whereas a blood pressure reading above 140 mm
Hg/90 mm Hg is considered hypertension.
Blood pressure is used by doctors as an indicator of risk factors for
heart disease. The more pressure is put on the walls of your arteries,
the more work the body has to exert.
But blood pressure is a measure of pressure on artery walls. Though
this generally might be associated with risk factors like a sedentary
lifestyle and habits like smoking, you need to take the data for what it
is—a measure of actual pressure on arterial walls.
This pressure can be caused by a myriad of factors. A simple example
of how a misreading can occur is if an individual has heightened stress
when visiting a doctor. Stress is known to be correlated with increased
blood pressure. When stressed, you often feel an increase in your heart
rate, and thus a subsequent pressure is added to your arterial walls.
Taking blood pressure measurements over a long span of time gives
a more accurate sense of what your average everyday blood pressure
is, so this is one way to combat issues like taking blood pressure
measurements only at doctor’s office visits.
But more importantly, if a physician makes a conclusion about your
health based on this statistic, mathematics says that you need to look
into this more.
To get a better sense of your overall health composition, it is better to
consider many different body measurements, blood pressure being one
factor in this.
49
Lesson 5 | Understanding Your Health Data
When getting a numerical reading of any
kind, you should ask what the units of
measurement are. Those units tell you
what is truly being measured. Though
the output, such as a high BMI or high
blood pressure, might correlate with
various health factors, you have to do
due diligence to find out what is really
going on in your individual situation.
Health Statistics
Another important aspect of taking control of your health is accurately
understanding health statistics. You are often presented with health
statistics, and conclusions are drawn based on those findings. But
you really need to look at the details of the data to make accurate
inferences.
One very common way this plays out is with reporting relative risks of
diseases or treatment without looking at the bigger picture.
50
Lesson 5 | Understanding Your Health Data
Suppose your doctor tells you that you have contracted a rare bacterial
infection, and there are two possible treatments. The doctor then says,
“Treatment 2 has a 50% higher probability of inducing a heart attack, so
I highly recommend treatment 1.”
What would your response be if you were given the information
that treatment 1 costs $5,000 and treatment 2 costs $50? If you could
afford it, you likely would much prefer to take treatment 1 after the
information you were given about the increased heart attack risk of
treatment 2.
However, there is a very important piece of information here that might
inform a decision: What if the chance of a heart attack under treatment
1 is two out of every 1,000,000 cases and under treatment 2 is three out
of every 1,000,000 cases? Then, on average, for every three heart attacks
on treatment 2, there are two heart attacks on treatment 1.
So, treatment 2 has a 3/2 = 1.5 times chance—i.e., a 50% greater chance—
of causing a heart attack. But the chances of a heart attack with either
treatment is extraordinarily rare!
Basically, the information you were originally given about percentages
is extremely misleading in the situation. But it could have truly affected
the decision you made afterward. This fictitious scenario is not far
from what happens in real life.
So if a doctor suggests that one treatment is much more effective than
another, or that one procedure is much riskier than another, consider
asking more about the mathematics of the situation. If both treatments
have a very low absolute risk, then comparing them to each other is
likely misleading.
51
Lesson 5 | Understanding Your Health Data
Medical Test Results
Another great time to ask your physician some questions is when you
receive results from a medical test.
Suppose you test positive for a disease. You might think that this
indicates a strong likelihood that you have the disease—that the
chances of testing positive and not having the disease are slim. But you
need to think twice about this, and mathematics can help you see why.
Testing positive for a disease when you do not have it is known as a false
positive. And false positives are much more common than you might
think they are.
Suppose a fictitious disease is present in only 0.1% of the population. Of
the people who have the disease, 99.9% of them test positive for it. Of
the people who do not have the disease, only 0.1% of them test positive.
Suppose you tested positive for the disease. What’s the likelihood that
this was an erroneous test—that it was a false positive?
Let’s imagine a population of 1,000 people and analyze what happens to
the people in this group who test positive.
Of the 1,000 people, 0.1% have the disease—or, in other words, one
person has the disease. This person has a 99.9% chance of testing
positive, so it is safe to say that this person tests positive. Of the
remaining 999 people—i.e., the ones who do not have the disease—0.1%
of them test positive. That’s roughly one person who tests positive.
So of the entire population of 1,000 people, two test positive. But only
one has the disease. That’s a false positive rate of 50%!
52
Lesson 5 | Understanding Your Health Data
If you had a population of 1,000,000 instead, 1,000 would have the
disease, and subsequently 999 of them would have tested positive. But
999,000 people would not have the disease, and 999 of those would test
positive, so again there is a 50% false positive rate: 999 tested positive
and have the disease, and 999 test positive and don’t.
TOTAL POPULATION = 1,000,000
TESTED TESTED
POPULATION POSITIVE NEGATIVE
HAVE 1,000 999 1
DISEASE
DO NOT 999,000 999 998,001
HAVE
DISEASE
This is very drastic for the 999 people who test positive and don’t have
the disease! Even though this scenario is a fictitious one, it is reflective
of real-life situations that can have drastic consequences for the lives of
individuals.
It is imperative that when you test positive for a disease, you do due
diligence to determine the efficacy of the testing procedures. It is
important for you to do this alongside doctor recommendations because
doctors may not take the time to analyze the numbers with you.
Understanding false positive rates by looking at the data gives you a
more informed opinion about the efficacy of diagnostic testing.
53
Lesson 5 | Understanding Your Health Data
How do you equip yourself against falling for these errors?
It’s too challenging to expect a doctor to be informed about every study
on every medicine they prescribe or every procedure they suggest.
But it is your duty to be proactive, and knowing these mathematical
principles can help you detect misleading information about your
health, even when it comes from a doctor.
Ask about the facts. Ask about details. Empower yourself with the
mathematics to do independent research through scientific journals.
All of this will enable you even more to make the best decision for you.
Reading
Charig, Webb, Payne, and Wickham, “Comparison of Treatment of
Renal Calculi.”
Lawlor, Smith, and Ebrahim, “Commentary.”
Naylor, Chen, and Strauss, “Measured Enthusiasm.”
US Department of Transportation, “Ride Sharing Programs.”
Vigen, “Spurious Correlations.”
54
6
TABLE OF
CONTENTS
The Math of
Environmental
Friendliness
55
Lesson 6 | The Math of Environmental Friendliness
T
e regular activities of humans—
h
including driving, creating waste, and
using paper—have an impact on the
well-being of Earth’s environment. But
which factors have more impact than
others? Which factors should you focus
on most? A mathematical perspective
can give you a better idea of where
to put your attention when thinking
about the health of the planet.
56
Lesson 6 | The Math of Environmental Friendliness
Carbon Footprint
Our carbon footprint is the total greenhouse gas emissions caused by
our day-to-day activities. From driving cars to recycling, many of our
activities have an impact on how much greenhouse gases we emit and
subsequently how much we impact our environment.
Greenhouse gases are gases in the atmosphere that absorb and reemit
heat to the Earth. As a consequence, these gases keep the Earth warmer
than it would be otherwise. On Earth, the primary greenhouse gases
include water vapor, carbon dioxide, methane, and nitrous oxide.
Greenhouse gases naturally occur in the atmosphere, but the
burning of fossil fuels and other similar human activities force the
concentration of greenhouse gases in the atmosphere to increase
unnaturally. The consequences are increases in temperatures on Earth.
Carbon footprint measures the amount of these greenhouse gases
released in the atmosphere by these human activities.
Though there are many greenhouse gases that are emitted by human
activity, the one that is most common is carbon dioxide. Because of
this, the scientific community has agreed upon measuring the emission
of any greenhouse gas based on the amount of carbon dioxide it would
take to generate the same emission.
The generally agreed upon standard measure for quantifying carbon
footprint—the amount of greenhouse gases released by human
activity—is carbon dioxide equivalent, or CO2e. For any type of
greenhouse gas and any specific amount of it, the CO2e indicates the
amount of carbon dioxide that would have an equivalent amount of
emission.
57
Lesson 6 | The Math of Environmental Friendliness
Suppose that in a particular industrial region on Earth, 10 tons of
methane and 1 ton of nitrous oxide are released into the atmosphere. It
turns out that an equal weight of methane has about 80 times as much
emission as carbon dioxide would, and an equal weight of nitrous oxide
has about 300 times.
So, the release of these gases has a CO2 equivalent of 1,100 tons:
10 × 80 + 300 = 1,100
So, although you often hear that carbon footprint is the amount of
carbon generated by human activity, it is actually more of an aggregate
of the amount of common greenhouse gases generated by human
activity, weighted by the gases’ relative impact on heating as compared
to carbon dioxide.
Human Activity
Based on data from the Environmental Protection Agency (EPA), the
total US greenhouse gas emissions in 2018 can be broken down by
sector as follows.
58
Lesson 6 | The Math of Environmental Friendliness
But if we’re looking to change our behavior to help the environment,
many of these contributions to greenhouse gas emissions are not in
our control.
The website epa.gov has a carbon footprint
calculator to help you make informed choices in
your daily life to reduce your carbon footprint.
Carbon Footprint of Waste
What’s the impact of waste on carbon footprint? In 2018, the average
carbon waste per person per year was 692 pounds of CO2e. This means
that the greenhouse gas emissions are the equivalent of the emissions
that would result from using 692 pounds of carbon.
How does recycling affect this? Here is a list of commonly recycled
items in US households, from least to most impactful on carbon
footprint.
RANK ITEM PER PERSON/YEAR
5TH glass 25 lbs. CO2e
4TH magazines 27 lbs. CO2e
3RD plastics 35 lbs. CO2e
2ND metals 90 lbs. CO2e
1ST newspapers 110 lbs. CO2e
59
Lesson 6 | The Math of Environmental Friendliness
There are a few things to notice about how math informs recycling
priorities:
1 If you want to get started with recycling, you can see where you would
have the most impact. If you get a newspaper daily, start there. The
average household would reduce their carbon footprint by more than
four times if they focused on recycling newspapers over recycling
plastics instead!
2 It makes a huge difference to have recycling as a part of our regular
practice. Without it, our waste is 692 pounds of CO2e per person per
year. If we recycle as the average person in the US does and do it
consistently in every category, we would be down to about 400 pounds
of CO2e per person per year.1
Carbon Footprint of Electricity
Electricity, natural gas, and fuel oil are used to power and heat our
households. Which contributes more to our carbon footprint?
RANK ITEM PER PERSON/YEAR
3RD natural gas 3,071 lbs. CO2e
2ND fuel oil 4,848 lbs. CO2e
1ST electricity 5,455 lbs. CO2e
1 The assumptions here are that you recycle 100% of the items in each of
these categories, though this might be unrealistic.
60
Lesson 6 | The Math of Environmental Friendliness
The interesting thing to notice about these numbers is not their relative
order but how they compare to the carbon footprint of waste.
The carbon footprint of all forms of household waste combined is 692
pounds of CO2e per person per year. That’s eight times less than the
carbon footprint of electricity. This might tell you that significant
reductions in electricity consumption might go a much longer way than
reductions in waste.
Empowered with mathematics, you would have to do some more
digging to arrive at such a conclusion, though. For instance, you would
need to know how much of the carbon footprint from electricity is truly
in your control.
Perhaps the basic necessities for operating the electricity of a
household might take up a bulk of the carbon footprint already, leaving
you only with a little bit that you have control to reduce.
So what can you do to find out how much of your carbon footprint from
electricity is truly in your control?
One way to gauge this is by looking at electricity bills. According to the
EPA, the average cost in the US in 2018 for the usage of 1 kilowatt-hour 2
was a little less than 12 cents, and the average monthly consumption of
electricity per household was 943 kilowatt-hours. So the average
monthly electricity bill per household was $112.03.
$0.118
943 kWh × kWh = $112.03
2 Electricity usage is measured in kilowatt-hours (kWh). One kilowatt-hour is
the energy needed to power 1,000 watts of energy for 1 hour.
61
Lesson 6 | The Math of Environmental Friendliness
If the carbon footprint of electricity is 5,455 pounds of CO2e per person
per year, and the amount spent on electricity per household per month
is $112, it would make sense to convert this dollar figure to the amount
spent per person per year.
There are an average of 2.57 people per household, and there are 12
months in a year. So if the cost of electricity is $112 per household per
month, each individual on average pays 2.57 times fewer than that,
resulting in $43.59 per person per month. Because there are 12 months
in a year, the amount paid per year is 12 times this amount.
0.12
943 × = $43.59
2.57
This is $523.08 per person per year.
Since the average carbon footprint is 5,455 pounds of CO2e per person
per year, this says that every dollar that’s charged for electricity
translates to about 10.5 pounds of CO2e per person per year.
You can now see what fluctuations in your electricity bill actually
directly convert to in terms of your carbon footprint.
For example, recall that waste accounts for an average of 692 pounds of
CO2e per person per year. If saving $1 on your annual electricity costs
translates to an average of about 10.5 pounds of CO2e per person per
year saved, that means that shaving off about $70 annually on your
electricity bill converts to reducing your carbon footprint by the entire
amount of household waste produced by one of your household
members in a year!
692
= 70
10.5
62
Lesson 6 | The Math of Environmental Friendliness
With math, you can compare the relative effects of different types of
conservation and make informed choices because of it.
Of course, in reality, you shouldn’t have to choose between reducing
waste or reducing electricity use. You really should do both. But at least
you can be empowered in understanding the impacts of each.
Carbon Footprint of
Transportation
Gasoline use contributes an average of 19.6 pounds of CO2e per gallon.
The average miles driven per gallon on a US vehicle is 24.7. So each mile
of driving translates to 0.79 pounds of CO2e.
19.6
= 0.79
27.7
The average miles traveled per vehicle is 11,398, so this gives an annual
average of 9,044 pounds of CO2e per car per year.
0.79 × 11,398 = 9,044
That outweighs the pounds of CO2e per person per year from electricity
by almost a factor of two—and outweighs waste by a factor of more
than 13!
63
Lesson 6 | The Math of Environmental Friendliness
Cars make a massive difference on carbon footprint. You can now use
these points of data to determine the effect of conservative measures
like carpooling on your carbon footprint.
Let’s say you and a work colleague live close together but work 25 miles
away from your homes. How can you estimate the carbon footprint
saved by driving together?
Assume that each person works an average 250 days a year.3 Since life
events happen, let’s use the conservative estimate that you carpool
together 200 days of the year.
If each person drove separately, they would drive 50 miles each day for
a combined total of 100 miles. If they drove together, the total would be
just 50 miles. This is a reduction of 50 miles of driving combined each
day. Over 200 days, this is a total of 10,000 fewer miles driven.
50 miles/day × 200 days = 10,000 miles
Because the average carbon footprint for driving is 0.79 pounds of CO2e
per mile, this amounts to 7,900 pounds of CO2e combined.
Per person, that’s almost 4,000 pounds of CO2e saved by carpooling.
That’s almost the equivalent of the carbon footprint of the yearly
waste of six people! It also offsets about 72% of each person’s
individual contribution to their household’s carbon footprint through
electricity use.
3 This takes into account weekdays, national holidays, and a two-week
vacation.
64
Lesson 6 | The Math of Environmental Friendliness
Paying attention to transportation has serious
potential in reducing your carbon footprint. From
public transportation to carpools, doing what
you can to drive less has an impact that goes
a very long way as compared to reduction in
your waste and in your household energy use.
Flying puts a big dent into your carbon footprint. According to
carbonfootprint.com, a round trip flight from LAX to JFK in economy
class adds 1,278 pounds of CO2e to your carbon footprint.
That’s almost twice the amount generated per person in a household
from waste over the span of two years! In fact, it’s more than the carbon
footprint of the average person over an entire year in more than 30
countries in the world!
Since flying adds quite substantially to your carbon footprint, it is good
to keep this in mind if you have other alternatives for certain trips you
might be taking.
Reading
Brander, “Greenhouse Gases, CO2, CO2e, and Carbon.”
Environmental Protection Agency, “Greenhouse Gas Emissions.”
GlobalChange.gov, “National Climate Assessment.”
US Department of Energy, “Find and Compare Cars.”
US Department of Transportation, “Ride Sharing Programs.”
65
7
TABLE OF
CONTENTS
Getting Wise
about Health
and Life
Insurance
66
Lesson 7 | Getting Wise about Health and Life Insurance
H
alth is one of the most important
e
facets of your life. But navigating
the health system is often extremely
complicated, with a multitude of
choices presented to you. This
lesson aims to fight this information
overload and empower you with the
mathematical tools necessary to
assess some of these pivotal choices in
your life.
67
Lesson 7 | Getting Wise about Health and Life Insurance
Consult with an insurance specialist
and a financial advisor before
making any specific decisions for
yourself or your loved ones.
Life Insurance
Insurance is a contract in which an individual or entity is given
financial protection against losses. Life insurance is a specific type
of insurance in which you pay a premium (monthly or annually) to
have an insurance company cover the potential loss of your income, or
perceived contribution to your household, upon your death. This is paid
to those you deem to depend on you, who are called beneficiaries.
There are many types of life insurance instruments, but two of the
largest and most common ones are term life insurance and whole life
insurance.
Term life insurance pays your beneficiaries in the case of death (subject
to terms on your life insurance contract) if you pass away during a
specific period stated in your contract.
You typically pay a premium for a specific number of years, called
a term—for example, 10 or 20 years. From the start date of the
policy, if you pass away during the term of the policy, you earn the
stated death benefit, and it is paid out to whomever you specify your
beneficiaries to be.
68
Lesson 7 | Getting Wise about Health and Life Insurance
One particular situation in which a 20-year term life insurance policy
could be advantageous is for a family with young children. In that 20-
year period, if something were to happen to a parent or guardian, they
could ensure that the children are taken care of regardless of when the
parent or guardian died, up until the children are in their early 20s.
Term life insurance policies are priced based on many factors,
including
{ length of term, where shorter term policies generally cost less;
{ health and family history of health, such as current and past health
conditions, family history of disease, and health lifestyle factors;
{ age, where premiums typically rise substantially with age;
{ gender, where premiums typically rise substantially if you are male
instead of female; and
{ hobbies, where a skydiving habit might lead to more expensive
insurance.
Consider Keisha, a 35-year-old shopping for term life insurance. She has
two children, ages 3 and 6, and a husband, Dwayne, who is also 35 years
old. A 20-year term policy that provides a $250,000 coverage would cost
her about $19.05 per month for 20 years, whereas a $1,000,000 coverage
would cost her about $52.16 per month.
Suppose Keisha earns $85,000 per year and Dwayne is a stay-at-home
dad who does not earn an income. It would be extremely beneficial to
Dwayne—and give him peace of mind—if he had life costs covered for
a substantial period of time after Keisha passed away, if that were to
happen. Covering her salary for 10 years would amount to $850,000;
with inflation, this could easily amount to more than $1,000,000.
69
Lesson 7 | Getting Wise about Health and Life Insurance
How much would a $1,000,000 policy cost to cover such a situation? The
premium is $52.16 per month, so if the policy lasted the full 20 years
(240 months), this would amount to $12,518.40 paid. That’s extremely
small compared to the $1,000,000 coverage provided!
Let’s now assume that instead of being married with kids, Keisha and
Dwayne are engaged 25-year-olds who each earn roughly $50,000 per
year and intend to one day have children.
If they are interested in term life insurance to cover lost wages in case
they pass away when they have children, should they buy insurance
now or wait?
Of course, there are many factors to consider, but suppose that the
couple decided to wait until they had children to purchase term life
insurance and that they had their first child at 40 years old. What would
the difference in term insurance costs be?
Age is a primary factor to consider with term life insurance. The
policy is priced based on the age you are when your insurance starts; it
typically does not increase every year of the policy.
Suppose that Keisha and Dwayne take out a joint life insurance policy at
age 25 with a payout of $1,000,000 and that under their insurance plan,
this costs $80 per month for a 20-year policy. Over the 20-year period,
this would be an expenditure of $19,200.
80 × 240 = 19,200
If they started at 40 years old and if insurance premiums for term life
insurance increased 6% for each year that you wait to start, how much
of an increase would this be?
70
Lesson 7 | Getting Wise about Health and Life Insurance
According to the compound interest calculator from lesson 1, the
premium would be $191.72 per month.
80 × (1.06)15 = 191.72
Over the life of the insurance policy, the cost will be $46,022—an
increase of $26,812.80 from the total cost of the insurance policy if
started at 25 years old.
So, starting early might seem like an advantage for the couple. But if
they are 25 and have a 20-year term policy, they would only have the
policy until they were 45. But they started having children at 40, so
their oldest child is only five years old. So they would have paid 15 years
into a 20-year policy and would probably want to add a newer 20-year
policy at a very high rate starting at 45 years of age.
One thing they could consider is starting the insurance policy young
and extending the life of the insurance, opting for a 30-year term rather
than a 20-year one.
Typically, 30-year term policies are more expensive, but not by a drastic
amount. If the couple pays $80 monthly on a 20-year term starting at 25
years old, the 30-year term might cost in the vicinity of $100 monthly.
Of course, many factors affect these numbers, but assuming this
ballpark estimate, the couple will pay a total of $36,000 on the 30-year
term insurance. This is still $10,000 cheaper than the 20-year policy
started at age 40!
So it seems that the 30-year term insurance policy is the winner—but it
starts at age 25 and thus ends at age 55, whereas the 20-year term starts
at age 40 and ends at age 60. This would leave ages 55 to 60 open if the
couple opts for the 30-year term.
71
Lesson 7 | Getting Wise about Health and Life Insurance
The decision on what to do here varies based on individual
circumstances. The key is to consider different term lengths and do the
long-term calculation to get a sense of what works best for you.
As opposed to term life insurance, whole life insurance is an insurance
policy that lasts for a policyholder’s lifetime. This insurance is paid
out to the beneficiary upon the policyholder’s death, provided that no
premium payments were skipped.
The premium you pay on a whole life insurance policy is typically
divided into two chunks: one that goes into a savings account within the
policy and one that goes toward the policy itself. You have the option of
borrowing against that savings account while you’re alive.
Why bother with term insurance if whole life insurance covers you
for life?
The difference in premium costs are staggering! If a 30-year-old male
sees a $53.91 monthly charge on a 30-year term life insurance policy
that pays out $1,000,000 upon death, that same male may see a whole
life policy premium of roughly $700 per month.
What if the 30-year-old decided to take the 30-year term life insurance
policy and invest the difference of roughly $646 per month over the
entire 30 years?
Assuming an average return of 7% each year, using the compound
interest calculator, this would amount to $788,211.06. But the difference
is this money is completely available to him at 60 years of age and can
still compound. If he needs access to it, he can use it. If he decided
instead to invest it for another 20 years until he’s 80, assuming the
same average rate of return, he will end up with $3,519,944.14. That’s
about $2.5 million more than he would have left his family under the
insurance policy!
72
Lesson 7 | Getting Wise about Health and Life Insurance
Furthermore, upon the death of the policyholder, the savings
component of the whole life insurance policy is absorbed into the
policy’s death benefit—so in a sense, the savings component disappears.
So while there are disadvantages to whole life insurance, there are
also advantages that whole life insurance may have over term life
insurance, depending on individual circumstances. For instance, whole
life insurance is a lifelong insurance, so it has a guaranteed payout in
the end. This might be more comfortable for individuals who wouldn’t
feel comfortable investing in the stock market after their term life
insurance has ended.
Health Insurance
As with life insurance, there are many different instruments that
provide health insurance and many different types of health insurance
programs.
Two specific types are HMO (Health Maintenance Organization) and
PPO (Preferred Provider Organization) plans. The major differences
between the two are as follows:
{ cost: HMO premiums are usually comparable to PPO premiums, but
HMOs usually have the lowest out-of-pocket costs, including low
deductibles.1
{ size of plan network: HMOs usually provide a list of doctors that are
in-network that can be seen for specialty services, whereas PPOs
typically offer a much wider list of options.
1 These are the amounts you are required to pay before insurance kicks in.
73
Lesson 7 | Getting Wise about Health and Life Insurance
{ ability to see a specialist: HMOs usually require a referral from a
primary care doctor to see a specialist; with PPOs, an appointment
can usually be booked with a specialist directly.
{ coverage of out-of-network services: The largest advantage that PPOs
have over HMOs is the flexibility to see out-of-network providers.2
These costs are even sometimes partially covered.3
Suppose Indira is single and earns $120,000 annually. She has the
option of choosing between PPO insurance and HMO insurance offered
by her company. Here are the details of both programs:
{ The HMO has a monthly premium of $53, whereas the PPO is $60.
{ The HMO has a max out-of-pocket limit of $6,500.
{ The HMO has a $20 copay for visits with a primary care doctor.
Copays vary for other services, depending on the specialist.
{ The deductible on the PPO is $1,500 for in-network services and
$3,000 for out-of-network services. These deductibles are separate;
one does not contribute to the other.
{ After the $1,500 deductible for in-network services, the PPO requires
Indira to pay 20% of medical care costs up to $3,500 for in-network
services and $6,500 for out-of-network services. After that, all
services are covered at 100%.
2 This does not mean that PPO-insured individuals can see any out-of-network
provider. There are many health providers who do not accept any type of
insurance.
3 There may be a separate out-of-network deductible, though.
74
Lesson 7 | Getting Wise about Health and Life Insurance
Suppose Indira is very healthy, has no existing health issues, has no
evidence of genetic predisposition to illness, and rarely goes to the
doctor, except for two to three annual checkups. She also wants to have
hospital coverage in case of an emergency.
If she went with the HMO plan, she would be paying the monthly
premium plus two or three copays for her annual checkups. She would
also have to pay her hospital copay if an emergency were to happen.
If she went with the PPO plan, it is likely that the entirety of her
medical care would not meet her $1,500 deductible, so she would likely
have to pay for her medical care out of pocket entirely.
In this scenario, then, Indira might prefer an HMO.
But suppose instead that Indira has a rare health condition that
requires a substantial amount of medication and weekly doctor visits.
A $20 to $50 copay each week on an HMO plan can add up. But more
importantly, it is likely Indira will need to see specialists, who likely
won’t be an option to visit under an HMO.
Reading
Kaiser Family Foundation, “Average Monthly and Annual Premiums for
Covered Workers.”
Moneysmart.gov, https://moneysmart.gov.au/#tools-and-resources.
75
8
TABLE OF
CONTENTS
Optimizing
Your Diet
through Math
76
Lesson 8 | Optimizing Your Diet through Math
T
e diet industry in America is a
h
multibillion-dollar industry. It’s an
industry that seems to provide people
with thousands of options for how to
achieve their goals but no definitive
way to decide which pathway is best
for any particular individual. Everyone
has a different health goal. Whether
your goal is weight loss, muscle gain,
or maintenance, mathematics can help
you create an optimal diet to achieve
your goals.
77
Lesson 8 | Optimizing Your Diet through Math
Foundations of Nutrition
We all have different bodies, with different metabolisms based on
lifestyle, age, and other identity factors; illnesses; and hormones.
Energy is required in every bodily function—from the functioning of
our brains to the use of our muscles. And the calorie is the basic unit
that’s used to measure that energy.
But what exactly is a calorie? For example, if you bought a drink and
saw that it had 170 calories in it, what does that actually mean?
Since a calorie is a unit of energy, this tells you the amount of new
energy you can expend after drinking this drink. Another way to view
this is that it tells you how much fuel your body gets—in other words,
how much energy it gets—by drinking this drink.
When you hear a phrase like “eat fewer calories,” this converts more
literally to “obtain less energy,” whereas the statement “burn calories
with exercise” means you will expend energy when you exercise.
Weight Loss
If your goal is to lose 10 pounds and you have a sedentary lifestyle, how
would you accomplish this without exercising much?
In order to figure this out, you need to calculate how many calories
you use up in a day—in other words, how much energy you expend in a
day. This is referred to as your calories out. Then, you need to compare
78
Lesson 8 | Optimizing Your Diet through Math
this with how many calories you take in, which is referred to as your
calories in. The difference between these two numbers is your net
calories for the day.
Calories in depends on the food you intake—
both the quantity and content. Calories out
depends on how much energy you exert.
How many calories do you burn when your body is completely at rest?
Of course, your body is rarely completely at rest. You twitch your eyes
and wiggle your nose, and even simple activities like these require
energy to perform. But knowing how many calories your body burns
when you are completely at rest gives you a lower bound for the number
of calories you burn by default, which can in turn give you a cap on the
number of calories you should be taking in.
The number of calories your body burns when it’s completely at rest is
referred to colloquially as your resting metabolic rate and scientifically
as your basal metabolic rate (BMR).
The common method of calculating BMR is the Harris-Benedict
equation, which calculates BMR based on four factors: gender, height
(in pounds), weight (in inches), and age (in years):
(4.536 × weight) + (15.875 × height) ± (5 × age) – 161,
where you subtract 5 times your age in years if you are a woman and
add 5 times your age in years if you are a man.
79
Lesson 8 | Optimizing Your Diet through Math
For example, for a 47-year-old woman who weighs 165 pounds and is 5
feet 8 inches tall, her resting metabolic rate is about 1,432 calories.1
(4.536 × 165) + (15.875 × 68) – (5 × 47) – 161
= 748.44 + 1,079.5 – 235 – 161
= 1,431.94
= ~1,432
The next step is to factor in activity.
Your total energy expenditure (TEE) is the total amount of energy you
use up in a day, including all of your physical activities.
To calculate your TEE, you start with your basal metabolic rate. The
higher your activity level, the more energy you expend, so the higher
your TEE is. You can quantify what your TEE is as a function of your
BMR by using a calorie-tracking app that you might use on your weight
loss journey.2
Assuming a sedentary lifestyle, the TEE in a day for the 47-year-old
female would be 1,718 calories. That’s how much energy she expends in
any given day.
In order to lose weight, you need your net calories to be negative so that
you expend more energy than you take in. But how much more energy
do you need to expend to lose the weight you want to lose?
1 These are just estimates. There are other factors that affect BMR that can
make a significant difference. For example, the percentage of your body that
is lean muscle mass has been found in some studies to have an influence
on BMR.
2 Though this is not a scientific source, it can give you an estimate on where to
start. If there don’t seem to be any concrete results in the first month or so
of your weight loss journey, you can adapt your program accordingly.
80
Lesson 8 | Optimizing Your Diet through Math
The rule of thumb in the nutrition world is that a 1-pound loss typically
requires an estimated 3,500 calories of energy. So if you want to lose 10
pounds, you can do so in 10 weeks by having a 3,500-calorie deficit per
week. This works out to a 500-calorie deficit per day on average, or a net
calorie amount of −500 calories.
The 47-year-old woman in this scenario would need a net calorie intake
of 1,200 calories per day.
1,718 − 500 = 1,218
This is quite an aggressive regimen! But you can change your plan in
several ways to optimize for weight loss while not being so strict on
your calories.
You can make room for more calories by being more active. If you
become lightly active—exercising lightly three or four times a week—
your TEE factor jumps from 1.2 to 1.37.
So, the TEE of the 47-year-old woman would be 1,962 calories. To lose 10
pounds in 10 weeks, she’d need the 500-calorie-a-day deficit, resulting
in a total caloric intake of 1,462 calories.
Another strategy is to lose weight over a longer period of time. Instead
of losing 10 pounds in 10 weeks, you could aim to lose 10 pounds in 20
weeks. This is a long horizon, but it doesn’t require as drastic a change
to your daily caloric intake.
Things get a bit trickier if you have a larger weight goal. Your TEE is a
constant multiple of your BMR, based on your activity level, and your
BMR depends on your weight. But as you lose weight, your BMR will
change, which in turn changes your TEE, which in turn changes the
number of calories you need to consume in order to lose weight.
81
Lesson 8 | Optimizing Your Diet through Math
If you are aiming to lose 10 pounds, the change in your TEE is not as
noticeable as it is when you are aiming to lose 50 pounds, for example.
How would you adapt your situation, assuming a lightly active lifestyle,
to lose 50 pounds over 50 weeks?
One strategy is to recalibrate the caloric intake you need every 10 weeks
or so based on your weight.
NET CALORIES FOR
WEEK POUNDS BMR TEE WEIGHT LOSS
0 230 1,991 2,728 2,228
10 220 1,946 2,666 2,166
20 210 1,901 2,604 2,104
30 200 1,855 2,541 2,041
40 190 1,810 2,480 1,980
50 180 1,764 2,417 1,917
As you continue moving forward along your plan, the number of
calories you can consume in order to maintain your weight loss goal
drops each day. This is because you lose weight along the way.
Keeping overall adherence to your general plan and crunching the
numbers beforehand can make a significant difference. It can give
you a general guideline to follow and a way to keep on track and be
accountable.
82
Lesson 8 | Optimizing Your Diet through Math
There are many apps available that give you a
target for net calories based on your specific goals.
If you input your starting weight, goal weight,
and amount of time you hope to accomplish the
weight loss, these apps provide targeted daily
calories in order to achieve your goals. They
even adapt your daily target as you continue
to lose weight so that you don’t have to think
about updating daily targets accordingly.
Weight Gain
Instead of losing weight, you may have an interest in gaining weight. You
can do so by reversing all of the net calorie losses to net calorie gains.
For example, instead of having a net calorie deficit of 500 calories a day,
you could alter this to a net calorie increase of 500 calories a day and
subsequently gain roughly 1 pound per week.
The tricky aspect of gaining weight is that for some people, the real goal
is to gain lean muscle mass. This requires a focus on looking not only at
how much you eat but also the content of what you eat.
The content of your food is actually important even if your goal is to
lose weight, because different food items have different nutritional
content in different densities.
83
Lesson 8 | Optimizing Your Diet through Math
So it’s not just about the calories you consume but also the content of
those calories. You’ve probably heard the statement “not all calories are
equal.” This essentially means that there are different types of calorie
sources your body receives.
When you eat any type of food, you obtain many different nutrients.
Micronutrients are vital for your health and mostly consist of minerals
and vitamins. These don’t play a large factor in your caloric intake.
What does play a large factor are macronutrients, which are required
in large amounts in your everyday diet. The three most prominent
macronutrients are fats, proteins, and carbohydrates.
One gram of each of these provides the body with a certain amount of
energy it can expend: One gram of either protein or carbohydrate gives
you 4 calories of energy, whereas fat gives you 9 calories of energy—
more than two times as much!
In planning any diet and optimizing it for your needs, you should take
into consideration that fats provide much more energy than carbs and
protein. This is why you might hear the recommendation to avoid fat or
eat fat-free food; per gram, fat provides many more calories.
CALORIES
1/4 CUP PEANUTS 207
2 APPLES 180
12 CUCUMBERS 180
84
Lesson 8 | Optimizing Your Diet through Math
If your goal is to gain weight—specifically, to gain lean muscle mass—
how should you tailor your diet?
The production of muscle is supported by the intake of a considerable
amount of protein. In terms of your diet, this would mean increasing
your consumption of high-protein foods, such as meats, poultry, fish,
beans and bean products, legumes, and nuts. Supplements like protein
powders and shakes would help, too. Keep in mind, though, that
regardless of your goals, it’s important to eat a balanced diet.
Empowered with this information, do an audit
of your caloric intake—individually tracking
carbohydrates, proteins, and fats—and analyze
the mathematics. Adjust your intake to suit
your goals and desires. You might be surprised
by your results after just a few months!
Reading
Mifflin, St. Jeor, Hill, Scott, Daugherty, and Koh, “A New
Predictive Equation for Resting Energy Expenditure.”
US Department of Agriculture, “Calculators and Counters.”
US Department of Health and Human Services, “Current
Dietary Guidelines.”
85
9
TABLE OF
CONTENTS
Making Great
Estimates with
Little Data
86
Lesson 9 | Making Great Estimates with Little Data
M
thematical thinking can be
a
powerful for helping you make
estimates quickly in situations
where you might not have a sense
of what reasonable estimates even
are. This lesson will equip you
with mathematical tools that will
allow you to make estimations that
are more informed than random
guesses.
87
Lesson 9 | Making Great Estimates with Little Data
Fermi Problems
Fermi1 problems are usually extreme scientific calculations that
can be estimated well if the calculations are broken down into
manageable steps.
For example, how many gallons of finished motor gasoline were
consumed in the United States in 2019?
The key question in consideration is the order of magnitude. Is it
millions? Tens of millions? Hundreds of millions? Billions? More? To
consider this, you start by making very rough estimates.
Let’s assume that the average car drives roughly 15,000 miles per year.
Assuming a family of four as a typical household with 1.5 cars, that’s
22,500 miles per household. There are roughly 325,000,000 people in the
United States, so a very rough estimate on the number of households
is about 81,250,000. So the total number of miles traveled over all
households is more than 1.8 trillion miles!
22,500 × 81,250,000 = 1,828,125,000,000
Next, how do you figure out how much gasoline a car holds? You can
look to your own experience. Let’s say you drive a Honda Civic and a
tank of gas usually lasts about 350 miles. The tank is about 12 gallons.
This is a relatively small car, so let’s go with 14 gallons for an average
car. That equates to roughly 25 miles per gallon.
1 Fermi problems were named after Enrico Fermi, a scientist who had an
uncanny ability to make decent approximations through calculations, even
with barely any data at his fingertips.
88
Lesson 9 | Making Great Estimates with Little Data
Taking into account the total number of miles driven that were
calculated previously, this gives you a final estimate of more than 73
billion gallons consumed in 2019.
1
1,828,125,000,000 × = 73,125,000,000
25
The true number of gallons turned out to be about 142.23 billion
gallons. So that was a pretty good estimate!
Calculations like these are great for quickly assessing and estimating
the order of magnitude of quantities, which then allows you to make
informed decisions.
Fermi problems are just one example of an estimation tool that you can
use to make you feel much more confident about your guesses.
Bell Curves
Another mathematical tool that can help with estimating quantities is
bell curves, which can be used to cast predictions. So much data, when
plotted in a graph, comes very close to following the shape of a bell curve.
For example, consider the activity of rolling four dice. Each die is labeled
with a number between 1 and 6, so the lowest total you can get is
1 + 1 + 1 + 1 = 4,
and the highest total is
6 + 6 + 6 + 6 = 24.
89
Lesson 9 | Making Great Estimates with Little Data
You might have an intuition that it’s pretty difficult to get a 4 or a 24 but
probably easier to roll a number like 15.
When you simulate rolling four dice 10,000 times, here is the
distribution of outcomes in a bar graph:
Notice that the data follows the shape of a bell curve!
You can use the power of knowing that data comes in a bell-shaped
curved to be able to make predictions.
First, notice that the peak of the curve happens around 14. This is
actually the average of all the possible rolls!
4 + 5 + … + 24
= 14
21
The sum of the numbers from 4 to 24 divided by 21, which is the total
number of possible outcomes, is 14.
90
Lesson 9 | Making Great Estimates with Little Data
So, the theoretical average—the average of all possible outcomes of
rolling four dice—is 14, and this happens to coincide with the peak in
the bell curve.
Also notice that the bell curve is symmetric about this mean of 14. So,
even after running the simulation, where the outcomes seemingly
couldn’t have been predicted, the average of these outcomes matches
the theoretical mean, and the data is symmetric about this mean!
Notice that the bell curve has an interesting shape, too. Many of the
outcomes for the sum of the four dice are clustered in the center, near
the mean of the bell curve, whereas the proportion of the sums that are
low or high is quite slim.
This type of phenomenon happens very often in data that fits into
bell-shaped curves like this. Data like this is considered to be normally
distributed.
Data is normally distributed if, when it is graphed, it follows the shape
of the graph of this function:
91
Lesson 9 | Making Great Estimates with Little Data
Notice that the graph is symmetric about the constant μ (mu), which is
referred to as the mean, or average, of the distribution.
Also notice the parameter σ (sigma), which is known as the standard
deviation. By the way that σ presents itself in the function, it dictates
when certain spreads occur. The bell curve has a steep drop-off near
the center, and then it starts to tail off. The place where it starts to drop
off is exactly σ away from the central axis at μ.
If data is normally distributed, it won’t look exactly like a bell curve,
but it will approximately look like one. You can actually figure out the
graph of the bell curve that approximates your data by getting rough
estimates from the data itself.
First, you can estimate the axis of symmetry of the bell curve (μ) by
calculating the sample mean, which is the average of all the data you
have collected.
For example, in the experiment with 10,000 dice rolls, the sample mean
is the actual average of all the rolls, which is 14.049—which is quite
close to the expected theoretical mean, which was 14.
So for your bell curve, you will use the sample mean for what μ is.
From average adult blood pressure to
the percentage gains or losses in stocks
on the stock market in a given day,
many phenomena follow the shape of a
bell curve.
92
Lesson 9 | Making Great Estimates with Little Data
One way to estimate the spread of the data, which is measured by σ, is
to look at how far each data point is from the mean and get an average
sense of what that is.
That calculation would look something like this:
Σ(𝑥𝑥 − 𝜇𝜇𝜇! ,
$
𝑁𝑁
where Σ is the sum, N is the total number of data points, and x takes on
the value of each data point in the data set as you sum.
In practice, though, for many technical reasons, mathematicians use a
quantity that is slightly different and looks like the following:
Σ (x − x– )!
$
n–1
This is called the sample standard deviation, and it is used to estimate
what σ should be in order to fit the best bell curve to the given data.
So, you have a data set that you know has a bell curve shape, meaning
that you know it is normally distributed, and you want to fit a bell curve
onto it that best approximates the data.
You can do this by graphing the function, where the parameters
that govern the function, μ and σ, can be calculated directly from
the data—μ from the sample mean and σ from the sample standard
deviation.
93
Lesson 9 | Making Great Estimates with Little Data
Let’s do this with the sample of 10,000 dice rolls:
The bell curve seems to accurately represent the bar graph data.
But why should you bother fitting a strange and seemingly complicated
curve to this bar graph?
It turns out that in situations where you know beforehand that data is
normally distributed, you can make inferences and estimations based
on the bell curve!
In the example with 10,000 dice rolls, let’s say you wanted to roll the
dice one more time and estimate probabilities for what the sum of the
dice in the next roll will be. How could you do this?
The sample mean (μ) was 14.049 and the sample standard deviation (σ)
was 6.025.
94
Lesson 9 | Making Great Estimates with Little Data
Luckily, there is a well-known mathematical principle that says that
for any bell curve, the probability of being one standard deviation,
two standard deviations, and three standard deviations away from the
mean are roughly 68%, 95%, and 99%, respectively. This is sometimes
called the 68-95-99 rule.
For example, in the dice-roll simulation, the data indicates that the
probability that the next roll falls in the interval (μ − σ, μ + σ) is 68%,
whereas the probability that the next roll falls in the interval (μ − 2σ,
μ + 2σ) is 95%.
What is the probability that the roll value is at least μ − σ?
Because of the symmetry of the bell curve, the probability of being
in the interval (μ − σ, μ) is the same as that of being in the interval
(μ, μ + σ). Since the probability of being in the interval (μ − σ, μ + σ) is
68%, the probability of being in the interval (μ − σ, μ) is half that, or
34%. By the symmetry of the bell curve, the probability of being at
least μ is 50%. So, altogether, the probability of being greater than μ − σ
is 84%.
95
Lesson 9 | Making Great Estimates with Little Data
So, when data is normally distributed, you have predictive power and
get reasonable estimations!
If you know beforehand that the data will be normally distributed,
it turns out mathematically that the sample mean of 40 randomly
generated data points does a pretty good job of estimating the true
mean. This is also true for the spread of the data.
Statistical Tools
Instead of rolling four dice, this time let’s simulate the outcome of
rolling only one die 10,000 times and record the result. Here is the
outcome:
96
Lesson 9 | Making Great Estimates with Little Data
This is not a bell-shaped curve! In the previous examples, the
assumption was that the data fit a bell curve. However, there are many
phenomena in which this is not the case.
So how do you know whether to expect a bell-shaped curve or not?
There are many proven statistical tests that can be used to determine
whether or not a normal distribution is a reasonable predictive model
for a situation at hand. A formal statistics course would enable you to
understand more about the nuances of this.
But in your everyday life, it would be good to know whether or not a
particular phenomenon is well suited for approximation by a normal
distribution.
Sometimes, looking at swaths of past data can give a reasonable sense
of whether or not this is the case. However, it would be good to at
least have some situations where it is definitive that data will indeed
be normally distributed so that you know the bell-curve model will
actually work. And it turns out that there are situations where this does
happen.
A particular one of importance to highlight is when you take data that
is itself an average aggregated over groups of the same type and where
the numbers you are taking an average of don’t depend on each other.
Recall that when one die was selected, the distribution did not look like
it was bell-shaped. Suppose instead that you wanted the average value
when 25 dice are rolled. This is an average aggregated over a group—
namely, a group of dice—all of the same type, which means that the
likelihood of a die taking any value is the same regardless of which die
is chosen. Moreover, the numbers you are taking an average of don’t
depend on each other: If one die lands on a 6, it doesn’t influence the
result of any other die.
97
Lesson 9 | Making Great Estimates with Little Data
Let’s simulate this 10,000 times and plot.
Notice that you indeed have a bell curve this time! Taking an average
aggregated over groups of the same type will lead to a normally
distributed data set—no matter what. This is confirmed by a statistical
theorem known as the central limit theorem.
This is a remarkable tool from statistics that you can use to help make
predictions in your everyday life and—even more importantly—one that
allows you to better understand the statistics that fly around you every
day, from political polling to medical reporting and beyond.
Reading
HealthKnowledge.org.uk, “Standard Statistical Distributions.”
O’Neil, Weapons of Math Destruction.
University of Maryland Physics Education Research Group, “University
of Maryland Fermi Problems Site.”
98
10
TABLE OF
CONTENTS
The First-Digit
Law and Fraud
Spotting
99
Lesson 10 | The First-Digit Law and Fraud Spotting
D
ta is integrated into many facets of
a
your daily life, from reading news
articles to managing your health. But
how can you tell if data reported to
you is accurate? More importantly,
how can you tell if data is fabricated?
100
Lesson 10 | The First-Digit Law and Fraud Spotting
First-Digit Distributions
Suppose you are asked to record the population of every county in every
state in the United States—that’s a total of 3,142 counties. Suppose you
recorded the first digit of this population.
For example, San Bernardino County in California has a population of
roughly 2.157 million people, so you would record a 2. Kalawao County
in Hawaii has a population of 88, so you would record an 8.
If you recorded the first digit of the population of every one of the 3,142
counties, what fraction would start with a 1? What about a 2? Or a 3?
If you made a bar graph plotting the first digits of the populations, what
would it look like? It would seem reasonable to assume that the chances
are uniformly distributed between 1 and 9. In other words, it wouldn’t
seem like having a 2 as the first digit is any more likely than having an 8
as the first digit.
Here is the actual graph of the distribution of the first digit in US
county populations:
101
Lesson 10 | The First-Digit Law and Fraud Spotting
About 30% of the counties have a population whose first digit is a 1,
whereas only about 4% have a population whose first digit is a 9. Also,
as the digit increases from 1 to 2 to 3, and so on, the chance of having a
digit as a leading digit gradually decreases.
One feature to notice is that the numbers in this data set vary over
different orders of magnitude, with some numbers in the tens, the
hundreds, the thousands, etc., all the way up to the millions. This turns
out to influence why the distribution looks like it does.
The first-known reference to this interesting distribution was due to
American astronomer Simon Newcomb in 1881. In that time, scientists
and mathematicians often used logarithm tables to explicitly compute
scientific and mathematical quantities. These tables were usually pages
and pages of numbers, and he noticed after some time that the first few
pages of the book were much more worn out than the latter pages.
102
Lesson 10 | The First-Digit Law and Fraud Spotting
From this, he suspected that numbers in science and mathematics
tended to have a much higher probability of having a 1 as a first digit
as opposed to a 2, and a much higher probability of having a 2 as a first
digit as opposed to a 3, etc.
It wasn’t until almost six decades later that physicist Frank Benford
rediscovered this. He took copious amounts of data stretching over
vast swaths of science and mathematics and noticed this first-digit
distribution appearing time and time again.
He collected data in nature, such as river lengths; data from physics,
such as x-ray voltages; baseball data from the American League; digits
in mathematical functions; specific heat capacity of materials; and so
on. The result every time was a distribution that looked similar to the
one seen for county populations.
This distribution is theoretically modeled by what is now known as the
Benford distribution.
Benford’s Law
Benford’s distribution estimates the probability of the leading digit in a
data set being one of the numbers 1 through 9.
Suppose P(d) is the probability of having a leading digit d. A data set is
said to satisfy a Benford distribution if for any d chosen from the
numbers 1, 2, 3, … 9,
1
𝑃𝑃( 𝑑𝑑 ) = log!" ) 1 + - .
𝑑𝑑
103
Lesson 10 | The First-Digit Law and Fraud Spotting
Let’s examine the consequent probabilities DIGIT P(d)
in a data set that satisfies a Benford 1 0.3010
distribution. 2 0.1761
3 0.1249
If a bar graph of the probabilities shown at 4 0.0969
right is plotted, notice that the probabilities 5 0.0792
indeed match a similar phenomenon that 6 0.0669
was found in the populations example. 7 0.0580
8 0.0512
9 0.0458
Keep in mind, though, that not all data will satisfy a Benford
distribution! For example, the age of people in a population would not
likely satisfy a Benford distribution. For instance, you wouldn’t expect
that 30% of the population is either 1 year old or between 10 and 19
years old!
However, as is seen experimentally, many data sets do satisfy a
distribution like this, so it is important to understand when data sets
satisfy it so that you can use it to help detect fraudulent data.
104
Lesson 10 | The First-Digit Law and Fraud Spotting
Detecting Fraudulence
If you’re given a data set and you want to test it for fraudulence using
the Benford distribution, there are two questions that need to be asked:
{ Should this data set be expected to follow Benford’s distribution?
{ If it does, how can you test for fraudulence?
In real life, both of these require rigorous statistical analyses. You
would have to take a course on statistics to faithfully argue that a data
set follows Benford’s distribution and that fraudulence is taking place.
But there is a good rule of thumb for when the Benford distribution
applies: If the data is spread over multiple orders of magnitude, it is
more likely to have a Benford distribution.
For example, human adult height in inches is almost always a two-digit
number, so you wouldn’t expect the distribution of this data to follow a
Benford distribution.
On the other hand, if you examined the number of dollars across all
bank accounts held by a large financial institution, the numbers would
range from 1 digit to many digits, so you might expect the distribution
to follow a Benford distribution.
That being said, even if you don’t have a sense of the order of magnitude
of your data, and even if you don’t perform statistical tests, you can
still use the Benford distribution to get a sense of whether or not
fraudulence might be happening.
To get a litmus test for possible fraudulent activity, you can:
{ Look at known data sets that follow a Benford distribution.
{ See, then, if your data set satisfies a Benford distribution.
105
Lesson 10 | The First-Digit Law and Fraud Spotting
It is well known to the IRS that certain fields in a tax return follow
a Benford distribution. This has become evident from millions of
analyses, together with adherence to some statistics. So Benford’s law
can be used in these contexts.
Suppose you were working for the IRS and were responsible for writing
a computer program that detects whether or not a company should be
audited. In other words, you want to write a program that detects the
possibility, not the certainty, of fraudulence.
What you can do is look at the specific tax fields that are supposed
to follow a Benford distribution and see if they actually do satisfy a
Benford distribution.
So, for instance, if you did this and found that the data looked like the
data illustrated in the black line graph, then you would have a sense
that fraudulence is at play and flag it for an audit.
Why wouldn’t someone with a nefarious mind cook the books so that
their data satisfies a Benford distribution to begin with?
106
Lesson 10 | The First-Digit Law and Fraud Spotting
There are two tricky things with this.
{ A lot of taxation data is obtained from doing arithmetic with other
numbers in the data, so you can only be so crafty in the end.
{ It turns out that there is a Benford law for not only the first digit, but
also the second digit, third digit, fourth digit, etc. This table shows
the expected distribution of these various digits. Because of this, it’s
very challenging to create large sets of data that satisfy all of these
at once.
2ND PLACE 3RD PLACE 4TH PLACE
DIGIT 1ST PLACE (IF FIRST DIGIT 1) (IF FIRST DIGIT 1) (IF FIRST DIGIT 1)
0 0.11968 0.10178 0.10018
1 0.30103 0.11389 0.10138 0.10014
2 0.17609 0.19882 0.10097 0.10010
3 0.12494 0.10433 0.10057 0.10006
4 0.09691 0.10031 0.10018 0.10002
5 0.07918 0.09668 0.09979 0.09998
6 0.06695 0.09337 0.09940 0.09994
7 0.05799 0.09035 0.09902 0.09990
8 0.05115 0.08757 0.09864 0.09886
9 0.04576 0.08500 0.09827 0.09982
This became apparent in one of the largest financial scandals of the
modern era: the Enron crisis. Mark Nigrini studied revenue numbers
and earnings-per-share data in reported financial numbers Enron
released in 2001 and 2002, and also in their actual numbers.
107
Lesson 10 | The First-Digit Law and Fraud Spotting
Here is the distribution of the second digit from the actual numbers in
2001 and 2002 and what’s expected from the Benford distribution:
This looks consistent with the Benford distribution, as you should
expect.
Now let’s look at the count of earnings-per-share data when using the
last two digits, as reported in 2001 and 2002. For example, if earnings
per share was $34.58, then the last two digits would be 58.
108
Lesson 10 | The First-Digit Law and Fraud Spotting
This doesn’t look consistent with the Benford distribution at all!
Nigrini did a statistical analysis and saw that, in particular, the
expected number of zeros in the last digit significantly exceeded
expected proportions in 2001 and vastly so in 2002.
Just by knowing how the Benford distribution works and looking at this
data, you are able to at least suspect that Enron was cooking the books.
Of course, detailed statistical analysis would be needed to actually
prove this.
Enron1 is just one of many instances in which the Benford distribution
has revealed suspicious financial behavior. Benford’s distribution has
had sweeping implications in other fields, too. For instance, it was used
to detect fraud in the 2009 elections in Iran. Today, evidence based on
the Benford distribution is widely used in the US legal system at all
levels of government.
Reading
Collins, “Data Mining Your General Ledger with Excel.”
Nigrini, “An Assessment of the Change in the Incidence of Earnings
Management.”
——— , Benford’s Law.
O’Neil, Weapons of Math Destruction.
1 Enron was one of the five largest audit companies on Earth before its
collapse.
109
11
TABLE OF
CONTENTS
Voter Math and
Gerrymandering
110
Lesson 11 | Voter Math and Gerrymandering
W
en it comes to voting, political
h
borders can be manipulated to
unfairly advantage one party
over another. This is a practice
that has a bipartisan nature: In
the United States, both Democrat
and Republican parties have
participated in seemingly unfair
practices that mathematics could
have helped uncover.
111
Lesson 11 | Voter Math and Gerrymandering
Playing with Political Borders
Imagine a political area that consists of 50 different regions and in
which 60% of the regions historically vote for a purple party, whereas
40% historically vote for a yellow party. The political layout may look
something like this:
Imagine the powers that be need to divide these 50 different regions
into 5 districts and that a party—let’s say the purple party—wins a
district if a majority of the regions in that district vote purple.
For example, if one of the 5 districts has 12 regions in it and 7 of those
regions vote purple, then the purple party wins the district.
Using this voting scheme, you can see how wildly the outcomes of the
election can differ depending on how the 5 districts are shaped.
112
Lesson 11 | Voter Math and Gerrymandering
Consider the following shape for the 5 districts:
In this case, the purple party would win all 5 districts and hence would
dominate the election. So if the purple party is in charge, this might be
a great strategy for them to decide how to shape the districts.
However, if the yellow party is in charge, the following strategy might
be better:
113
Lesson 11 | Voter Math and Gerrymandering
If the 5 districts take on these shapes, then the yellow party will win
3 districts, whereas the purple party wins 2. So now the yellow party
wins the election!
So even though the votes in the individual regions did not change,
manipulating the shape of the districts made for completely different
voting outcomes.
The practice of intending to manipulate political boundaries to favor
one party over another is called gerrymandering.
How is gerrymandering detected in practice?
This map of North Carolina’s congressional districts in 2012 certainly
looks like it’s gerrymandered. District 3 is not even connected, while
district 12 is extremely thin.
114
Lesson 11 | Voter Math and Gerrymandering
The map of North Carolina congressional districts in 2016 also has
some issues, given the varied shapes of the regions, especially the one
that goes along the coastal stretches.
How can these irregularities be quantified?
One perspective is to suggest that districts shouldn’t be too far off from
being circular. Circles, by their very nature, capture the neighborhood of
a specific location and are equally inclusive of points near their centers.
Using a mathematical formula, Daniel D. Polsby and Robert Popper
introduced a method for scoring a region based on how close it is to
being circular.
115
Lesson 11 | Voter Math and Gerrymandering
The Polsby-Popper Score
If you have a circle of radius r, the area of the circle (A) is πr 2, whereas
the perimeter (P) is 2πr. The ratio of the area of the circle to the square
of the perimeter is a constant:
𝐴𝐴 1
=
𝑃𝑃! 4𝜋𝜋
If you multiply this quantity by 4π, you get 1:
4𝜋𝜋𝐴𝐴
=1
𝑃𝑃!
It turns out that if a district (D) is any 2D shape, or compilation of 2D
shapes, whose area is A(D) and whose perimeter is P(D), then
4𝜋𝜋 × 𝐴𝐴 (𝐷𝐷)
≤ 1.
𝑃𝑃(𝐷𝐷)!
So this number is always between 0 and 1, and it equals 1 only when
D is a perfect circle. This value for a district is called the Polsby-
Popper score.
Suppose a district D is a perfect rectangle with a perimeter of 400 miles.
If it were a square, then each side of the district would be 100 miles, so
the area of the district would be
A(D) = 10,000 mi2.
From this, the Polsby-Popper score of the district would be roughly
0.785.
4𝜋𝜋 × 10,000
= 0.785
160,000
116
Lesson 11 | Voter Math and Gerrymandering
This is a somewhat high score, as it is pretty close to 1.
If the district D were a rectangle with a length of 150 miles, then the
width would be 50 miles, resulting in an area of 7,500 square miles.
A(D) = 7,500 mi2
From this, the Polsby-Popper score would drop drastically to
roughly 0.589.
If the region were drastically changed so that it had a length of 190
miles, then the width would be only 10 miles. As a very thin rectangle,
this would look like a region that was intentionally gerrymandered.
The area would be 950 square miles, and the Polsby-Popper score would
be very low at 0.149.
So, for regions that are thinner and more spread out, the Polsby-Popper
score is very low, supporting the suspicion that such regions are
gerrymandered.
Gerrymandering is a compound word formed by
combining the words Gerry and salamander. The
practice was named after former governor of
Massachusetts Elbridge Gerry, who signed a bill
that redrew the state senate district to make a
new district that was shaped like a salamander.
117
Lesson 11 | Voter Math and Gerrymandering
In practice, the Polsby-Popper score has issues. For instance, it is
usually sensitive to regions that have very complicated coasts, because
the complication usually drastically adds to the perimeter.
Despite this, the score is used abundantly to assess potential
gerrymandering.1 But there are also other mathematical tools that are
used to assess gerrymandering.
The Reock Score
If districts whose shapes are far from being circular might serve as an
indicator for gerrymandering, then measuring how close a district is to
being circular could help identify this practice. To do this, you can try
to enclose the district by the smallest circle possible and see how much
space the district takes up in the circle.
For example, if you enclose the following region in the smallest circle
you can, the region barely takes up any space in the circle.
1 In 2000, proposition 106 was passed in Arizona, giving the power of
districting congressional and legislative boundaries to an independent
body that was bipartisan. This committee used the Polsby-Popper score to
evaluate if gerrymandering was taking place.
118
Lesson 11 | Voter Math and Gerrymandering
However, if you do that with this region, it takes up a lot of space.
And the first region looks a lot more suspect to gerrymandering than
the second.
For a district D, let M(D) be the area of the minimum bounding circle of
the district. Then, you are calculating this ratio:
A(D)
M(D)
This ratio is always between 0 and 1, because the district sits inside
the circle, so it always has a smaller area than the circle. Named after
Ernest Reock, this ratio is called the Reock score of a district.
As you can see, then, the North Carolina district shown on the previous
page would have a low Reock score because it barely takes up any area
compared to its enclosing circle.
And the Alabama congressional district shown above would have a
high one.
This matches the notion that the Reock score does a decent job of
detecting gerrymandering.
119
Lesson 11 | Voter Math and Gerrymandering
The Reock score is similar to many other methods for testing
gerrymandering. Some of these include bounding by a rectangle
instead of a circle and bounding by other polygons. But there are other
potential indicators of gerrymandering that are not intrinsically based
on the shape of districts.
The Efficiency Gap
Suppose you had a state that was square in shape and had voting
patterns that could be summed up in the following picture.
There are two political parties: the yellow party and the purple party.
The state happens to have its population evenly distributed throughout,
and each of the circles represents the same number of voters, all of
whom vote for the party of the circle’s color.
120
Lesson 11 | Voter Math and Gerrymandering
So there’s an almost even split in the voting in this scenario of the 36
circles: 19 are yellow and 17 are purple.
You are in charge of selecting the districts. Each district should contain
exactly 3 circles, and you will let them take on the following shapes:
Suppose you were interested in gerrymandering in favor of the yellow
party.
One way you could go about this is by creating districts that are filled
with purple circles. This way, you concentrate all of the purple party’s
voting power into a district and subsequently reduce their voting power
in other districts.
121
Lesson 11 | Voter Math and Gerrymandering
This phenomenon is known as packing. And you can apply it in favor of
the yellow party by drawing the districts like this:
Notice that there are a number of districts filled with purple blocks,
and as a result, the purple party’s power is concentrated in these
districts. The purple party’s voting power in other districts is diffuse.
Even though the purple party and yellow party have almost an equal
number of votes in this state, the yellow party succeeded in packing the
purple party’s votes, leaving them out to dry in many of the districts.
Another way to gerrymander in favor of the yellow party is to dilute the
voting power of the purple party by splitting their votes across many
different districts. This process is known as cracking.
122
Lesson 11 | Voter Math and Gerrymandering
Mathematically, packing and cracking can be detected by the efficiency
gap, which measures the extent to which packing or cracking occurs
when district boundaries are drawn.
Suppose there’s a state with 3 districts and a 2-party system with a
yellow party and a purple party. The vote count in each district is as
follows:
DISTRICT YELLOW PARTY PURPLE PARTY
1 70,000 20,000
2 30,000 60,000
3 40,000 70,000
In district 1, the yellow party won. Since there were 90,000 total votes,
the yellow party only needed a majority of votes to win, or 45,001 votes.
So, in essence, the other 24,999 (70,000 − 45,001) votes are wasted. If the
yellow party was in charge of district lines and wanted to gerrymander,
they could have maneuvered these extra votes to other districts instead.
Similarly, the purple party lost, so all the votes that went to the purple
party in this district were wasted—a total of 20,000 wasted votes.
In district 2, the purple party wasted 14,999 (60,000 − 45,001) votes, and
the yellow party wasted all 30,000 votes.
In district 3, there was a total of 110,000 votes, so 55,001 were needed to
win. The purple party wasted 14,999 (70,000 − 55,001) votes, whereas the
yellow party wasted all 40,000 votes.
123
Lesson 11 | Voter Math and Gerrymandering
In total, then, the yellow party wasted a total of 94,999 votes, whereas
the purple party wasted 49,998 votes.
YELLOW WASTED PURPLE WASTED
DISTRICT PARTY VOTES PARTY VOTES
1 70,000 24,999 20,000 20,000
2 30,000 30,000 60,000 14,999
3 40,000 40,000 70,000 14,999
TOTAL 94,999 49,998
These raw numbers of wasted votes depend heavily on the population
size, so to get a sense of the disparity in wasted votes, you look at the
difference between the number of wasted votes in each party (across all
districts) and divide by the total number of votes.
(94, 999 − 49,998)
= 0.155
290,000
This ratio is the efficiency gap. It was devised by Nicholas
Stephanopoulos and Eric McGhee in 2014.
Researchers widely accept that a 7% efficiency gap, or an efficiency gap
of 0.07 or higher, is evidence of partisan gerrymandering via packing
and cracking—which have happened throughout history.
So the efficiency gap picks up both phenomena of packing and
cracking. This tool is used throughout government in everyday life. For
instance, it is referenced in ongoing litigation proceedings that argue
for gerrymandering by certain states.
124
Lesson 11 | Voter Math and Gerrymandering
Reading
Brennan Center for Justice, “Current Partisan Gerrymandering Cases.”
Juckett and Feinberg, “The Redistricting Game.”
O’Neil, Weapons of Math Destruction.
125
12
TABLE OF
CONTENTS
Dividing a
Cake or an
Inheritance
Fairly
126
Lesson 12 | Dividing a Cake or an Inheritance Fairly
W
ether it’s sharing a cake with
h
friends and family or sharing
an inheritance with other heirs,
sharing is made difficult when
preferences collide among those
doing the sharing. How can it be
decided to equitably split assets
so that each party is as happy as
possible?
127
Lesson 12 | Dividing a Cake or an Inheritance Fairly
The Divide-and-Choose Method
Suppose Arjun and Jamilla are interested in sharing a cake. But the
cake is quite complex: Some parts have strawberry icing, some parts
have no icing, some parts have vanilla filling, and the edges on two
sides are coated with sprinkles. Each person has preferences for each of
these items, one of them liking some items more than the other person.
How can a procedure be devised for them to split the cake fairly?
What exactly does the word fairly mean here? One reasonable
interpretation of fairness is what game theorists refer to as envy-free.
In this sense, a cake-splitting is fair if no person is jealous of the other
person’s piece. Both people feel that their piece is at least as good as the
other person’s piece.
But how is this possible to implement?
One way to do this is known as the divide-and-choose process, where
one person divides the cake into two pieces and then the other person
chooses which of the two pieces they want.
Let’s say Jamilla divides. She can find a way to split the cake so that she
equally prefers each side of the cake—so that in her opinion, she would
be happy with either piece.
More generally, cake cutting is a metaphor for splitting a multifaceted
entity into pieces. Many real-life scenarios, such as splitting an
inheritance between two people, can be modeled as the decision of
splitting a cake decorated with many features between two individuals.
Once Jamilla has done the dividing and is equally happy with either
piece, Arjun can choose whichever of the two pieces he prefers. By
doing that, Arjun will not be envious of Jamilla because he can choose
the piece he thinks has greater value to him, so he won’t think Jamilla’s
128
Lesson 12 | Dividing a Cake or an Inheritance Fairly
piece has greater value. But this is also great for Jamilla: She won’t
be envious of Arjun because she thinks both pieces have equal value.
Thus, this strategy is envy-free.1
Splitting in a way that each person is happy
is usually referred to as fair division.
Sperner’s Lemma
Suppose Arjun, Jamilla, and Natchanon are feuding, adversarial
siblings who inherited a triplex that carried a mortgage on it. They
want to split the triplex among themselves, one person to each of the
three units.
The triplex consists of three units that have different specifications
from one another. For instance, one of them might have more natural
light than the other two, whereas another one might have a larger
master bedroom.
Arjun, Jamilla, and Natchanon are also facing different life
circumstances. For instance, one of them wants to move into one of the
units whereas another plans to use their unit for rental income.
Because of their life situations and personal preferences, each of them
values the different units in the triplex differently.
1 Envy-freeness is not the only measure of fairness that can be constructed.
There are many others, including proportional fairness.
129
Lesson 12 | Dividing a Cake or an Inheritance Fairly
How can they split the mortgage among the three units and assign the
siblings ownership to the units so that each sibling feels they are paying
for their chosen unit in a fair way?
Each person wants to end up in a situation where they do not perceive
that anyone else has more value on their share. Is there an envy-free
way to fairly divide this inheritance among these heirs?
There is a mathematical framework that does a great job in addressing
this question.
To start, consider a large equilateral triangle that’s split into many
smaller equilateral triangles. This is called a triangulation of the large
triangle. Each of the triangles has vertices, and they are labeled each
with one of three possible colors: red, blue, or yellow.
But there are some rules:
{ Along the left side of the large triangle,
all vertices must be red or blue.
{ Along the right side of the large
triangle, all vertices must be red
or yellow.
{ Along the bottom edge of the
large triangle, all vertices
must be colored blue or
yellow.
All other vertices can be any color.
A mathematical theorem called Sperner’s lemma says that there must
be a triangle in the triangulation whose vertices use all colors: red,
blue, and yellow.
130
Lesson 12 | Dividing a Cake or an Inheritance Fairly
Suppose the triplex that
Arjun, Jamilla, and
Natchanon want to split
has a monthly mortgage
of $3,000. Imagine
associating to each unit
of the triplex a color: Unit
1 is red, unit 2 is blue, and
unit 3 is yellow.
Now start with a large equilateral triangle. If you pick a point in an
equilateral triangle and look at its distance to the three sides, the sum of
these distances is always the same regardless of the point you choose.
You are going to represent each choice of mortgage by a point, P, inside
(or on the border) of this equilateral triangle. You will think of the point
P as representing a choice of mortgage allotments based on the relative
proportions of the distances from P to the sides of the equilateral triangle.
If you color the line segment from P to the leftmost edge of the
equilateral triangle red, the line segment from P to the rightmost
edge yellow, and to the bottom edge blue, then
the relative proportions of these lengths
will represent the assignments of unit
mortgage cost.
For example, if P is in the middle of
the equilateral triangle, then the
distances are all the same, so each
unit would be assigned $1,000 of
the mortgage.
131
Lesson 12 | Dividing a Cake or an Inheritance Fairly
If instead P was in the middle of the leftmost
edge, then the red unit would have a $0
mortgage, whereas both yellow and
blue units would have a $1,500
mortgage.
To decide on a fair mortgage
division, start by triangulating
the large equilateral triangle
into very small equilateral
triangles so that there are
many triangles and label the
vertices of the triangles with
letters so that
{ the corners of the big triangle are labeled A,
J, and N (one for each person) and
{ the vertices of every triangle in the
triangulation are labeled A, J, and N
(one for each person).
You can achieve the second point
by breaking up the big equilateral
triangle into small equilateral
triangles in a structured way
and labeling the vertices with
the letters A, J, and N in a
pattern.
Since each vertex in every triangle of this triangulation is a point in the
large equilateral triangle, it is associated with a mortgage allotment
based on its distance to the side of the equilateral triangle. The vertex is
also labeled by a person.
132
Lesson 12 | Dividing a Cake or an Inheritance Fairly
So you ask the person labeled by that vertex a question about the
mortgage allotment assigned to that vertex: Which unit would they
prefer with the mortgage allotment given by that point?
For example, in the vertex labeled J, you ask
Jamilla which unit she prefers with the
mortgage allotment assigned to that point.
Let’s say she chooses the red unit. Then,
you color that vertex red and move on.
You keep moving along every vertex,
asking the person labeled at that
vertex what their unit preference
is based on the mortgage
allotment at that vertex.
On the leftmost edge of the large equilateral triangle, every vertex
would correspond to a mortgage allotment that makes the red unit
mortgage $0.
It’s reasonable to assume, then, that people want this unit, so you can
assume that all of the vertices in the triangulation on the leftmost edge
of the large triangle, except possibly the vertices of the large equilateral
triangle, are all red.
Similarly, all the ones on the bottom edge
of the large triangle are blue (again,
except possibly the vertices of the large
equilateral triangle), and all the ones on
the rightmost edge are yellow (once
again, except possibly the vertices
of the large equilateral triangle).
133
Lesson 12 | Dividing a Cake or an Inheritance Fairly
This possibly causes some confusion with the colors at the vertices of
the large equilateral triangle.
But notice that you are very close to a Sperner labeling here: red and
another color on the left, blue and another color on the bottom, yellow
and another color on the right.
In the rare case that you don’t have a Sperner labeling, you’ll change
the colors of the vertices of the large equilateral triangle to make it so.
This amounts to adjusting unit choices when two of them happen to be
free, which is reasonable to do.
Now, no matter what, all the leftmost vertices use red and one other
color, the bottom use blue and another color, and the right use yellow
and another color.
So you have a labeling that is amenable to Sperner’s lemma! And this means
that there is a triangle somewhere whose vertices use all three colors.
That means that Jamilla, Natchanon, and Arjun are happy with the
mortgage assignment associated with these three vertices.
Since the triangles in the triangulation are small, these mortgage
assignments are very close to each other in dollar value, so it is likely
that each person is happy with any of them.
This is a fantastic procedure that allows you to divide costs in a way
that is fair to everyone. In the mathematical literature, this process is
usually referred to as rental harmony. 2
2 If you have a fair division problem, try playing with an online rental harmony
calculator, such as this one from The New York Times, to get a sense of what
divisions might work for you.
nytimes.com/interactive/2014/science/rent-division-calculator.html
134
Lesson 12 | Dividing a Cake or an Inheritance Fairly
Reading
Brams and Taylor, Fair Division.
Robertson and Webb, Cake-Cutting Algorithms.
Sun, “Divide Your Rent Fairly.”
135
Quiz
SOLUTIONS CAN BE FOUND ON PAGE 140.
1 A doctor gives you the option of choosing between two treatments
for a disease. Treatment 1 has three times as much chance of causing
long-term neurological damage as treatment 2. Treatment 1 costs $50
and treatment 2 costs $5,000. Which treatment would you choose?
2 A procedure for testing allergies to cats comes back positive for 80% of
patients who actually have the allergy. For patients who do not have a
cat allergy, roughly 10% test positive anyway.
a What is the percentage of false positives in this allergy test?
b It’s known that 1% of the population has a cat allergy.
Imagine a population of 1,000 random people—so 10 of them
have a cat allergy. How many of them would test positive?
How many of the people who don’t have a cat allergy
would test positive for cat allergies anyway? What is your
recommendation for administering this test, knowing these
numbers?
3 A round trip flight from LAX to JFK in economy class adds roughly
1,300 pounds of CO2e to your carbon footprint. Driving adds about 0.8
pounds of CO2e to your carbon footprint for every mile you drive. The
drive from LAX to JFK is roughly 2,800 miles. Which is more efficient
on the environment: three round trip flights from LAX to JFK or one
round trip by car?
136
Quiz
4 Suppose you invest $500 into an asset that appreciates at an average
of 6% per year over a time period of 30 years. If you instead waited
15 years and invested $1,000 appreciating at 6% for 15 years, which
investment would come out ahead?
5 Sia, a 40-year-old, has $200,000 in a retirement account. If she invests
$1,000 per month in the retirement account until she is 65 years old, at
an average annual interest rate of 4%, how much will she end up with
at the age of 65?
6 Suppose a house is for sale and a family wants to take out a $300,000
mortgage for it. The annual interest rate is 5%. How much more
would they spend if they took out a 30-year mortgage versus a 15-year
mortgage?
7 Susanna has $50,000 for a down payment on a home. She qualifies
for a 15-year mortgage at a 5% annual interest rate. If she wants a
monthly mortgage payment of no more than $2,000, how expensive of
a home can she buy?
8 A popular credit card allows you to earn 2% cash back on all
purchases. You do an audit of your finances and find that on months
that you solely use your debit card, you spend $4,000 per month,
but on months you solely use your credit card, you end up spending
$6,000. Is it a good idea to use this credit card for the cash back?
9 Suppose you want to choose one of two cars to drive with for your
commute to work. Car 1 averages 20 miles per gallon whereas car 2
averages 50 miles per gallon. Car 1 costs $25,000 whereas car 2 costs
$30,000. Also suppose that work is 20 miles away from home and that
gas averages $3.50 per gallon over the next 10 years (for simplicity,
don’t assume any inflation). How much would you save in commuting
costs in a 10-year period by driving car 2 as opposed to car 1 if you
commute an average of 220 days per year?
137
Quiz
10 Minkus earns a salary of $85,000. To make sure his family is OK
if something were to happen to him, he wants to ensure that they
receive money equivalent to his salary for 10 years. He is married and
has two healthy children, who are 13 and 14 years old. If he wants to
take out term life insurance, would you suggest a 10- or 20-year term?
And how much should the policy cover?
11 An individual has a total energy expenditure of 2,300 calories per
day. They want to lose 10 pounds over the next 20 weeks. How many
calories per day should they consume to reach this goal?
12 Two common sentiments in popular health are that restricting
carbohydrates is a requirement for weight loss and that eating six
small meals a day is a requirement for weight loss. Are these true?
13 How many dentists are there in the United States? Is the number on
the order of hundreds, thousands, tens of thousands, or more?
14 The height of a male in the United States is normally distributed,
with a mean of 70 inches and a standard deviation of 3 inches. What
percentage of males are at least 6 feet and 4 inches tall?
15 Explain why, if a data set has a Benford distribution, that the
probability that a leading digit is either a 2 or 3 is the same as the
probability that it is a 4, 5, 6, or 7.
16 Suppose you encounter a data set and notice that 90% of the entries
start with a 1. Can you conclude that there is fraudulence in the
data set?
138
Quiz
17 A state has three congressional districts and two major parties: the
purple party and the yellow party. Votes were cast for an election, and
the totals in the districts were as follows:
DISTRICT YELLOW PARTY PURPLE PARTY
1 10,000 500
2 4,000 4,500
3 5,000 5,200
Calculate the efficiency gap. Would you support a claim that the
districts are gerrymandered?
139
Solutions
1 You need to examine the overall chance that each treatment will cause
long-term damage. For example, if the chance of treatment 1 causing
long-term damage is 3 out of every 1,000,000 patients, then treatment
2 has a 1 out of 1,000,000 chance. Both of these are extremely low, so
treatment 1, the low-cost treatment, could be considered. However,
if treatment 1 has a 66% chance of causing long-term neurological
damage, whereas treatment 2 has a 22% chance, then one might
heavily consider treatment 2. Of course, if a situation like this comes
up in your life, consult a physician before making any decisions.
2 a Since 10% of those who do not have the allergy test positive,
the false positive percentage is 10.
b Since 80% of those who have the allergy test positive, 8 out
of the 10 would test positive. Since there is a total of 1,000
people, 990 must not have the allergy. Of those, 10%, or 99
people, test positive. So, a total of 99 + 8 = 107 people test
positive, but only 8 of them actually have an allergy. This
seems like an ineffective test.
3 Three round trips by plane would add up to 1,300 × 3 CO2e, or 3,900
CO2e. One trip by car from LAX to NYC and back would be a distance
of 5,600 miles. At 0.8 CO2e/mile, this would amount to 4,480 CO2e,
which is significantly greater than the greenhouse gas emissions from
flying three round trips.
140
Solutions
4 Using the compound interest calculator, the first investment would
end up with $500 × (1 + 0.06)30 = $2,871.75. The second investment
would end up with $1,000 × (1 + 0.06)20 = $2,396.55. So, investing double
the amount of money ($1,000 instead of $500) and compounding over
half the time (15 years instead of 30 years) still results in significantly
less money in the end. This is the power of compound interest!
5 First, her $200,000 will compound annually for 25 years. Using the
compound interest calculator, this will grow to $200,000(1 + 0.04)25 =
$533,167.27. Compounding for 25 years is equivalent to compounding
for 300 months. Her annual interest rate as a decimal is 0.04, so her
monthly interest rate is 0.04/12. Using the monthly contribution
formula, her monthly contributions will total
(1 + 0.04/12)300 − 1
$1000 · = $514,129.55,
0.04/12
giving a total of $533,167.27 + $514,129.55, or $1,047,296.81.
6 The monthly interest rate of the mortgage is r = 0.05/12. The 15-year
mortgage compounds over n = 15 × 12 = 180 months. Set the principal P
to equal $300,000. Using the monthly mortgage payment formula, this
gives a monthly payment of
(1 + 0.05/12)180
$300, 000 × (0.05/12) × ,
(1 + 0.05/12)180 − 1
or $2,372.38. Multiplying by the 180 months that is the life of the loan,
this is a total payment of $427,028.59.
141
Solutions
For the 30-year mortgage, n = 30 × 12 = 360 months. So the monthly
payment becomes
(1 + 0.05/12)360
$300, 000 × (0.05/12) × ,
(1 + 0.05/12)360 − 1
which is roughly $1,610,46. Multiplying by the 360 months that is the
life of the loan, this is a total payment of $579,767.35. So the family will
spend roughly $152,738.79 more if they take a 30-year mortgage versus
a 15-year one.
7 The monthly interest rate on the mortgage would be r = 0.05/12. The
number of months is n = 15 × 12 = 180. The goal monthly payment is
$2,000. If you let P be the principal on the mortgage, then the monthly
mortgage payment formula says that
(1 + 0.05/12)180
$2, 000 = P × (0.05/12) × ,
(1 + 0.05/12)180 − 1
which is $252,910.49. Adding her down payment of $50,000, Susanna
can afford a home of a little more than $300,000 to have a monthly
mortgage of $2,000.
8 If you use the credit card for all your expenses, you spend $6,000 and
get 2% cash back, which is $120. So your net expenditure is $5,880.
However, if you use your debit card only, you spend $4,000. So in this
scenario, it is not a good idea to use the credit card for the cash back.
9 You commute 220 days per year over 10 years for a total of 2,200 days
commuting. Each of these days, you commute a total of 40 miles
to and from work, so in total you commute 88,000 miles. To get the
number of gallons this amounts to, you divide by the number of miles
per gallon for each car, which results in 4,400 gallons for car 1 and
142
Solutions
1,760 gallons for car 2. That’s a difference of 2,640 miles. With a gas
cost of $3.50 per gallon, this is an additional cost of 2,640 miles × $3.50
gallons per mile, or $9,240. Given that car 2 is $5,000 more expensive
than car 1, choosing car 2 saves $4,240.
10 The term policy should cover $850,000. This covers Minkus’s salary
for 10 years. It could be safer to go with a $1,000,000 policy to cover
inflation. Since Minkus’s children are 13 and 14, they will be 23 and
24 in 10 years. This might give Minkus the freedom to assume that
his children can be self-sufficient if need be, given their health, so he
might opt for a 10-year term policy.
11 Recall that losing 1 pound requires a calorie deficit of 3,500 calories
per week. If the individual aims to lose 10 pounds over 20 weeks, they
can do so by losing 0.5 pounds every week. This means they need to
be in a 1,750-calorie deficit each week, or a deficit each day of 1,750/7
= 250 calories. Since the individual’s total energy expenditure is 2,300
calories per day, they should eat 2,300 − 250 = 2,050 calories per day to
achieve the goal.
12 Outside of individual health factors, the key to losing weight is being
in a calorie deficit. This can certainly be achieved while having
carbohydrates in your diet, so it’s not required for weight loss. Eating
six meals a day is also not a requirement; six small meals could
arguably help with satiation, but eating six small meals a day isn’t a
necessity.
13 It turns out that there are roughly 200,000 dentists in the United
States. Here’s one way you might have estimated this. Assume that a
dentist goes into the office about 200 days out of the year (accounting
for weekends, holidays, and vacations). Each day the dentist might see
approximately eight patients, one every hour. This means the dentist
sees roughly 1,600 patients in a year. Assuming that the average
individual who has dental care visits the dentist twice a year (some go
less often and others go more often), then a dentist sees 800 patients
143
Solutions
per year. This includes only people who have dental care. Assuming
that half of the country’s 330,000,000 people have dental care, this
means that 165,000,000 have dental care. If each of these people has
a different dentist, you get an estimate of 165,000,000/800, which is
roughly 206,000 dentists—the correct order of magnitude (hundreds of
thousands).
14 In total, 6 feet and 4 inches is (12 × 6) + 4 = 76 inches. Let μ = 70 inches
(the mean) and σ = 3 inches (the standard deviation). Then, 76 = μ + 2σ.
Since heights are normally distributed (i.e., bell-shaped), the
probability of the height being between μ − 2σ and μ + 2σ is 95%. So
the combined probability of being at most μ − 2σ and at least μ + 2σ is
5%. So, by the symmetry of the normal distribution, the probability of
being at least μ + 2σ is 2.5%. Thus, the percentage of males that are at
least 6 feet and 4 inches tall is 2.5%.
15 Let P(d) be the probability that the leading digit is d. P(d) = log(d + 1) −
log(d). Then,
P(2) + P(3) =
[log(3) − log(2)] + [log(4) − log(3)]
= log(4) − log(2)
= log(4/2) = log(2),
whereas
P(4) + P(5) + P(6) + P(7) =
[log(5) − log(4)] + [log(6) − log(5)]
+ [log(7) − log(6)] + [log(8) − log(7)]
= log(8) − log(4)
= log(8/4)
= log(2),
so these probabilities are the same.
144
Solutions
16 No. You’d need to know that the data set should be Benford-distributed
in order to make such a conclusion. For instance, if the data set
consisted of the ages of high school students, then almost all the
leading digits would be 1, and this would not likely spark suspicion of
fraudulence.
17 The total number of votes in district 1, district 2, and district 3 were
10,500, 8,500, and 10,200, respectively. So the total number of votes
needed to win in district 1, 2, and 3 were 5,251, 4,251, and 5,101,
respectively. From this, the total number of wasted yellow party
votes was
(10,000 − 5,251) + 4,000 + 5,000 = 13,749,
whereas the total number of wasted purple party votes was
500 + (4,500 − 4,251) + (5,200 − 5,101) = 848.
The difference in wasted votes was 13,749 − 848 = 12,901, and the total
number of votes cast altogether was 29,200, so the efficiency gap is a
whopping 0.442, or 44.2%—definitely suggestive of gerrymandering.
145
Bibliography
Brams, Steven J., and Alan D. Taylor. Fair Division: From Cake-Cutting to
Dispute Resolution. Cambridge University Press, 1996. This book dives
into a more detailed analysis of fair division, introducing many types of
algorithms for fairly allocating resources under personal constraints.
[Lesson 12]
Brander, Matthew. “Greenhouse Gases, CO2, CO2e, and Carbon: What
Do All These Terms Mean?” Ecometrica, August 2012. https://ecometrica.
com/assets/GHGs-CO2-CO2e-and-Carbon-What-Do-These-Mean-v2.1.pdf.
This document gives a synopsis of what CO2e is and provides the CO2e of
various greenhouse gases. [Lesson 6]
Brennan Center for Justice. “Current Partisan Gerrymandering Cases.”
https://www.brennancenter.org/our-work/court-cases/current-partisan-
gerrymandering-cases. This web page has up-to-date information on
current partisan gerrymandering court cases in the United States, with
details about the rule of the efficiency gap and other mathematical
markers of gerrymandering. [Lesson 11]
Charig, Clive R., David R. Webb, Stephen Richard Payne, and John E.
Wickham. “Comparison of Treatment of Renal Calculi by Open Surgery,
Percutaneous Nephrolithotomy, and Extracorporeal Shockwave
Lithotripsy.” British Medical Journal (Clinical Research Edition) 292, no.
6524 (1986): 879–882. This article details the adverse effects of Simpson’s
paradox on health predictions. [Lesson 5]
146
Bibliography
Collins, J. Carlton. “Data Mining Your General Ledger with Excel: Learn
How to Use Microsoft’s Ubiquitous Spreadsheet Application to Unlock the
Wealth of Information Stored in Your Company’s Financials.” Journal of
Accountancy 223, no. 1 (2017): 27. This article gives a detailed description of
how to use spreadsheets and Benford’s law to detect potential fraudulent
activity in a company’s financial statements. [Lesson 10]
Environmental Protection Agency. “Greenhouse Gas Emissions.” https://
www.epa.gov/ghgemissions. This is the premiere government hub for
information on greenhouse gas emissions in the United States. It provides
a plethora of data, from macroscopic to individual. It also contains
the greenhouse gas emissions of many fossil fuel–burning activities.
[Lesson 6]
GlobalChange.gov. “National Climate Assessment.” https://
nca2014.globalchange.gov/report/our-changing-climate/observed-
change#graphic-16679. This government website gives informed
predictions about observed changes in the climate. [Lesson 6]
HealthKnowledge.org.uk. “Standard Statistical Distributions (e.g. Normal,
Poisson, Binomial) and Their Uses.” https://www.healthknowledge.org.
uk/public-health-textbook/research-methods/1b-statistical-methods/
statistical-distributions. This web page gives further details on statistical
distributions beyond the normal distribution. It also details how these
distributions are used in practice. [Lesson 9]
Juckett, Emily, and Joseph Feinberg. “The Redistricting Game: Teaching
Congressional Gerrymandering through an Online Simulation Game.”
Social Education 74, no. 5 (2010): 274–278. This article discusses an online
simulation for redistricting that can be used as a tool for teaching
congressional gerrymandering concepts. It contains links to the
simulation for you to try. [Lesson 11]
147
Bibliography
Kaiser Family Foundation. “Average Monthly and Annual Premiums for
Covered Workers, by Plan Type and Firm Size, 2018.” https://www.kff.
org/report-section/2018-employer-health-benefits-survey-section-1-cost-
of-health-insurance/attachment/table-1-2/. This web page gives detailed
information and comparisons of annual average premiums for health
coverage under Kaiser. [Lesson 7]
Lawlor, Debbie A., George Davey Smith, and Shah Ebrahim. “Commentary:
The Hormone Replacement–Coronary Heart Disease Conundrum: Is
This the Death of Observational Epidemiology?” International Journal of
Epidemiology 33, no. 3 (2004): 464–467. This study illustrates the issue with
assuming that correlation implies causation—in this case, when applied to
medical practices. [Lesson 5]
McKenzie, Brian. “Who Drives to Work? Commuting by Automobile in
the United States: 2013.” Census.gov, August 2015. https://www.census.
gov/content/dam/Census/library/publications/2015/acs/acs-32.pdf. This
document provides a detailed account of commuting data in the United
States in 2013. [Lesson 4]
Mifflin, Mark D., Sachiko T. St. Jeor, Lisa A. Hill, Barbara J. Scott, Sandra
A. Daugherty, and Young O. Koh. “A New Predictive Equation for Resting
Energy Expenditure in Healthy Individuals.” The American Journal of
Clinical Nutrition 51, no. 2 (1990): 241–247. This article constructs the
updated formula for computing BMR based on weight and gender.
[Lesson 8]
Moneysmart.gov. https://moneysmart.gov.au/#tools-and-resources.
This is an online resource provided by the Australian government that
offers online calculators for life insurance, mortgages, and credit cards.
[Lesson 2, Lesson 3, Lesson 7]
148
Bibliography
Naylor, C. David, Erluo Chen, and Barry Strauss. “Measured Enthusiasm:
Does the Method of Reporting Trial Results Alter Perceptions of
Therapeutic Effectiveness?” Annals of Internal Medicine 117, no. 11 (1992):
916–921. This scientific study illustrates how claiming relative differences
in the probability of complications in a treatment plan isn’t as informative
as absolute probabilities. [Lesson 5]
NerdWallet.com. “Mortgage Calculator.” https://www.nerdwallet.com/
mortgages/mortgage-calculator/calculate-mortgage-payment. One
of many online calculators for monthly mortgage payments and total
mortgage payout, this resource also allows you to include mortgage
insurance and taxes in the payment. [Lesson 2]
Nigrini, Mark J. “An Assessment of the Change in the Incidence of
Earnings Management around the Enron‐Andersen Episode.” Review of
Accounting and Finance (2005). This article presents the details of the Enron
scandal, with the assessments informed by Benford’s law. [Lesson 10]
——— . Benford’s Law: Applications for Forensic Accounting, Auditing, and
Fraud Detection. vol. 586. John Wiley & Sons, 2012. This book builds on
Nigrini’s findings and lays out many applications of Benford’s law to
testing for fraudulence. It’s a great book if you’re interested in getting into
much more detail about the content covered in the Benford’s law lesson.
[Lesson 10]
O’Neil, Cathy. Weapons of Math Destruction: How Big Data Increases
Inequality and Threatens Democracy. Broadway Books, 2016. This fascinating
book looks at how data is not without bias and how its prevailing use for
predictions can exacerbate existing inequalities and threaten democracy.
[Lesson 9, Lesson 10, Lesson 11]
149
Bibliography
Robertson, Jack, and William Webb. Cake-Cutting Algorithms: Be Fair If
You Can. CRC Press, 1998. This is another resource to expand on the fair
division discussion in the final lesson of the course. [Lesson 12]
Stine, G. J. “AIDS Update 1999: An Annual Overview of Acquired Immune
Deficiency Syndrome.” Upper Saddle River, NJ: Prentice-Hall, 1999. This
body of work discusses the false positive rates of the ELISA test early on in
the AIDS epidemic. [Lesson 4]
Sun, Albert. “Divide Your Rent Fairly.” The New York Times. April 28,
2014. https://www.nytimes.com/interactive/2014/science/rent-division-
calculator.html. This web page provides a calculator that allows you
to fairly divide rent, given room preferences, among any number of
roommates. It implements Sperner’s lemma in multiple dimensions to do
this. [Lesson 12]
Thomsett, Michael C., and Michael M. Thomsett. The Mathematics of
Investing: A Complete Reference. John Wiley & Sons, 1989. This book provides
references for the mathematics of investing, including many tools for
calculating compound interest. [Lesson 1, Lesson 2, Lesson 3]
University of Maryland Physics Education Research Group. “University
of Maryland Fermi Problems Site.” https://www.physics.umd.edu/perg/
fermi/fermi.htm. This web page supplies a plethora of problems that can
be used to practice working on Fermi problems in estimation. [Lesson 9]
US Department of Agriculture. “Calculators and Counters.” https://www.
nal.usda.gov/fnic/calculators-and-counters. This online calculator offers
several options for keeping track of health markers. It includes a body
weight planner, BMI calculator, calorie calculators, and BMR calculators.
[Lesson 1, Lesson 8]
150
Bibliography
US Department of Energy. “Find and Compare Cars.” https://www.
fueleconomy.gov/feg/findacar.shtml. This web page allows you to compare
fuel economy between many different types of cars. It’s a very beneficial
tool in assisting with reducing the cost of commuting. [Lesson 6, Lesson 4]
US Department of Health and Human Services. “Current Dietary
Guidelines.” https://health.gov/our-work/food-nutrition/current-dietary-
guidelines. This web page presents the national dietary guidelines for
health for 2020 to 2025. It includes expanded information for various
demographics.
US Department of Transportation. “Ride Sharing Programs.” https://
www.transportation.gov/mission/health/ride-sharing-programs. This
web page offers a full discussion of the benefits of ride-sharing programs,
from impact on the environment, to health, to cost efficiency. [Lesson 5,
Lesson 6, Lesson 2]
US Securities and Exchange Commission. “Compound Interest Calculator.”
https://www.investor.gov/financial-tools-calculators/calculators/
compound-interest-calculator. This calculator can be used to calculate
compound interest. [Lesson 1, Lesson 2, Lesson 3]
Vigen, Tyler. “Spurious Correlations.” https://www.tylervigen.com/
spurious-correlations. This website does a fantastic job of showing how
correlation can be far from implying causation. It displays many data sets
that are correlated but would seem absurd to have a causal relationship.
[Lesson 5]
151
Copyright © The Teaching Company, 2021
Printed in the United States of America
This book is in copyright. All rights reserved.
Without limiting the rights under copyright reserved above,
no part of this publication may be reproduced, stored in or
introduced into a retrieval system, or transmitted, in any
form, or by any means (electronic, mechanical, photocopying,
recording, or otherwise), without the prior written permission of
The Teaching Company.
4840 Westfields Boulevard
Suite 500
Chantilly, VA 20151‑2299
USA
1-800-832-2412
www.thegreatcourses.com
You might also like
- BOB Bank: ACCOUNT NO: - 879 BALANCE: - 10301.57No ratings yetBOB Bank: ACCOUNT NO: - 879 BALANCE: - 10301.5743 pages
- Equity Note On Beximco Pharmaceuticals Limited (BXPHARMA)No ratings yetEquity Note On Beximco Pharmaceuticals Limited (BXPHARMA)7 pages
- GST On Sale of Old Cars Article Feb 2018No ratings yetGST On Sale of Old Cars Article Feb 20185 pages
- SAP Business Blueprint Document - FI-GL - Ver 1.0No ratings yetSAP Business Blueprint Document - FI-GL - Ver 1.032 pages
- The World Through Picture Books 3rd EditionNo ratings yetThe World Through Picture Books 3rd Edition292 pages
- 1.1 Master Vocabulary Workbook (More Exercises)No ratings yet1.1 Master Vocabulary Workbook (More Exercises)339 pages
- Great Minds of The Eastern Intellectual Tradition Grant Hardy100% (1)Great Minds of The Eastern Intellectual Tradition Grant Hardy168 pages
- How To Turn Your Passion Into Profit - Course Guidebook (The Great Courses Wondrium) (Tara McMullin) (Z-Library)No ratings yetHow To Turn Your Passion Into Profit - Course Guidebook (The Great Courses Wondrium) (Tara McMullin) (Z-Library)104 pages
- 100 Moments of Connection Tracker - Big Life JournalNo ratings yet100 Moments of Connection Tracker - Big Life Journal8 pages
- BSBFIM601 Manage Finances: ASSESSMENT 1 - Written Questions75% (4)BSBFIM601 Manage Finances: ASSESSMENT 1 - Written Questions41 pages
- 30-Second Biochemistry - The 50 Vital Processes in and Around Living Organisms, Each Explained100% (1)30-Second Biochemistry - The 50 Vital Processes in and Around Living Organisms, Each Explained162 pages
- Physics HL - Worked Solutions - Chris Hamper - Second Edition - Pearson 2014No ratings yetPhysics HL - Worked Solutions - Chris Hamper - Second Edition - Pearson 2014111 pages
- 4.3.A. Topic Research Activity For GC PDFNo ratings yet4.3.A. Topic Research Activity For GC PDF3 pages
- Near-Optimal Parameterization of The Intersection of Quadrics: I. The Generic Algorithm ?No ratings yetNear-Optimal Parameterization of The Intersection of Quadrics: I. The Generic Algorithm ?23 pages
- Important Technical Analysis Terms 1694528472No ratings yetImportant Technical Analysis Terms 169452847212 pages
- Study Guide: Logic and Contemporary RhetoricNo ratings yetStudy Guide: Logic and Contemporary Rhetoric114 pages
- Written Communications - Being H - Allison FriederichsNo ratings yetWritten Communications - Being H - Allison Friederichs122 pages
- Eyewitness Companions Philosophy Stephen Law PDFNo ratings yetEyewitness Companions Philosophy Stephen Law PDF2 pages
- Explicit Direct Instruction: The Power of The Well-Crafted, Well-Taught Lesson100% (1)Explicit Direct Instruction: The Power of The Well-Crafted, Well-Taught Lesson1 page
- Speak The Culture Spain Be Fluent in Spanish Life and CultureNo ratings yetSpeak The Culture Spain Be Fluent in Spanish Life and Culture356 pages
- The Black Death - New Lessons FR - Dorsey Armstrong100% (1)The Black Death - New Lessons FR - Dorsey Armstrong64 pages
- FRIA by Atty Zarah Villanueva-Castro-2020 Dec 11100% (1)FRIA by Atty Zarah Villanueva-Castro-2020 Dec 115 pages
- Varmora - Carlyle Acquires Significant Minority Stake in Ahmedabad-Based Varmora Granito - The Economic TimesNo ratings yetVarmora - Carlyle Acquires Significant Minority Stake in Ahmedabad-Based Varmora Granito - The Economic Times1 page
- SEM - 6 Entrepreneurship and Small Scale BusinessNo ratings yetSEM - 6 Entrepreneurship and Small Scale Business13 pages
- Delegate From Co, Joshua Joy The Kingdom of Sweden Position Paper For United Nations Security Council I. Urgently Addressing The Syrian Civil War CrisisNo ratings yetDelegate From Co, Joshua Joy The Kingdom of Sweden Position Paper For United Nations Security Council I. Urgently Addressing The Syrian Civil War Crisis3 pages
- (TTC) Goldman, Science in The Twentieth Century Guidebook 1 PDFNo ratings yet(TTC) Goldman, Science in The Twentieth Century Guidebook 1 PDF37 pages
- Explorations in Nanophotonics-6: The Left Is Sometimes Right!No ratings yetExplorations in Nanophotonics-6: The Left Is Sometimes Right!2 pages
- (The Great Courses) Barbara J. King-Biological Anthropology An Evolutionary Perspective-The Teaching Company (2002) PDFNo ratings yet(The Great Courses) Barbara J. King-Biological Anthropology An Evolutionary Perspective-The Teaching Company (2002) PDF26 pages
- Four Steps To Writing A Position Paper You Can Be Proud ofNo ratings yetFour Steps To Writing A Position Paper You Can Be Proud of6 pages
- BANTUREETHI KOLU Hamsanaadam Adi Thyagaraja50% (2)BANTUREETHI KOLU Hamsanaadam Adi Thyagaraja3 pages
- ICSE - Economics Sample Paper-1-Class 10 Question Paper100% (2)ICSE - Economics Sample Paper-1-Class 10 Question Paper2 pages
