Lazy Student Integrals

Author(s): Gregory Galperin and Gregory Ronsse

Source: Mathematics Magazine , Apr., 2008, Vol. 81, No. 2 (Apr., 2008), pp. 152-154
Published by: Taylor & Francis, Ltd. on behalf of the Mathematical Association of
America

Stable URL: https://www.jstor.org/stable/27643097

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Taylor & Francis, Ltd. and Mathematical Association of America are collaborating with JSTOR
to digitize, preserve and extend access to Mathematics Magazine

This content downloaded from

192.195.76.32 on Thu, 03 Oct 2024 07:38:28 UTC
All use subject to https://about.jstor.org/terms
 152 MATHEMATICS MAGAZINE

x, y. We refer to the coefficients in the Maclaurin series as the loss sequence. The table
below gives some sample results. Taking (x, y) = (5, 1), the table shows that L5, L7,
and L9 have 1, 4, and 13 elements, respectively. If we take (x, y) = (4, 2), the table
shows that L4, L6, and L8 have 1, 4, and 13 elements, respectively.

Loss Function g(u) Loss Sequence?first ten terms

1/(1-") 1, 1, 1, 1, 1, 1, 1, 1, 1, 1

(1 -u)/(l -2u) 1, 1,2,22,23,24,25,26,27,28

(1 -2w)/(l -3u+u2) 1,1,2,5,13,34,89,233,610,1597
1/(1 -4u + 3u2) 1, 4, 13, 40, 121, 364, 1093, 3280, 9841, 29524

1/(1 -4w + 3w2) 1,4, 13, 40, 121, 364, 1093, 3280, 9841, 29524
1/(1 -3u) 1,3,32,33,34,35,36,37,38,39

We have been able to show that the radii of convergence r of the Maclaurin series of all
loss functions obey 1/4 < r < 1. The case (jc, y) = (1, 1) is an exceptional case, since
in this case, we do not get an infinite sequence since the game then ends in one flip. For
n > 1, the pairs (jc, y) = (n, 1) and (x, y) = (n ? 1, 2) yield the same loss function.
Also, for any particular loss function, there can be only finitely many pairs (jc, y) that
have this same loss function. The loss sequence for (jc, y) = (1,4) yields the odd
indexed Fibonacci numbers (plug Fibonacci numbers or Fibonacci polynomials into
an internet search).

Acknowledgements. The authors would like to thank the referees, the editor, and Emeric Deutsch for valuable
suggestions.

REFERENCES
1. W. Feller, An Introduction to Probability Theory and Its Applications, Vol. 1, 3rd ed., John Wiley, 1968.
2. J. Bak, The Recreational Gambler: Paying the Price for More Time at the Table, this Magazine 80 (2007)
183-194.
3. J. Harper and K. Ross, Stopping Strategies and Gambler's Ruin, this Magazine 78 (2005) 255-269.
4. E. Parzen, Modern Probability Theory and Its Applications, Wiley Classics, 1992.
5. R. I. Jewett and K. A. Ross, Random Walks on Z, College Math. J. 19 (1988) 330-342.
6. S. Ross, A First Course in Probability, 5th Edition, Prentice Hall, 1997.

Lazy Student Integrals

GREGORY GALPERIN
GREGORY RONSSE
Eastern Illinois University
Charleston, IL 61920
[email protected]
[email protected]

A challenging integral

Let a e R and consider the problem of evaluating

Jo (1 + jc?)(1 + x2)

This content downloaded from

192.195.76.32 on Thu, 03 Oct 2024 07:38:28 UTC
All use subject to https://about.jstor.org/terms
VOL.81, NO. 2, APRIL 2008 153
as a function of a. As a first attempt, we might substitute x = tan# to obtain the
integral

r12 cos" 0 d0
J0 cos? 0 + sin" 0 ' (1)
This seems no better than the original. As a second attempt, split the integral over
intervals [0, 1] and [1, oo). On the second interval substitute u = l/x to obtain

dx fl uadu
1(a) =Jo[ (1+jc?)(1+jc2) +io (l+w?9(l+?
Replacing the dummy variable of integration w with x a
we obtain

fl dx it
which is independent of a.
Many people, at first glance, think we should obtain a decreasing function of a
and find it surprising that 1(a) is constant. It is a trick of the mind. Upon looking at
an integral of a bounded function over [0, oo), we tend to think of the behavior of
the function for large values of x and ignore the behavior for 0 < x < 1. The natural
symmetry of inversion between [0, 1] and [1, oo) reveals that the integral over each of
these two intervals must sum to a constant.
Now, of course, the integral in equation (1) is the constant 7r/4, independent of a.
Is there symmetry here?

Tale of the lazy student

In first semester calculus class, students learn to evaluate definite integrals by finding
anti derivatives. In order to remind them that an integral is a limit of Riemann sums, it
is wise for an instructor to ask them to evaluate an integral similar to

/33 Vlsinx3dx
+xA + cos(2*) '
The answer is, of course, 0. We are integrating an odd function over an interval which
is symmetric about 0. The area above the x-axis is equal to the area below the x-axis.
The lazy student, upon seeing such complicated integrals, has become conditioned
to write down 0 immediately and get the right answer. He has noticed that such prob
lems always seem to have positive and negative portions that cancel each other. The
instructor must grudgingly admire this valid insight, but he seeks to enforce more care
ful analysis by altering the problem. So he adds a constant to the integrand, but keeps
it mysterious by combining the constant with the fraction to keep the previous denomi
nator but alter the numerator. Our lazy, but perceptive, student now notices a new rule.
"Complicated integrals" can be evaluated by evaluating the integrand at 0 and then
multiplying by the length of the interval. The exasperated instructor throws in another
gimmick by translating the integral along the x-axis by translation to an interval [a, b]
in order to disguise the symmetry. The "good students" are completely baffled and an
gry. However, our lazy, but ingenious, student rises to the occasion. He evaluates the
integrand at the interval's midpoint and multiplying by the length of the interval.

This content downloaded from

192.195.76.32 on Thu, 03 Oct 2024 07:38:28 UTC
All use subject to https://about.jstor.org/terms
 154 MATHEMATICS MAGAZINE

The lazy student's method will not, of course, work for all "complicated integrals".
Nevertheless, the lazy student would consider the value of the integral in equation (1)
as obvious (provided this student was not too lazy in algebra and trig class). Another
formula that is obvious to the lazy student is

(3 + ^),/* =5_
Jo 6 + Jt^ + (10-jt)^
Let us endeavor to make these formulas obvious to the non lazy, by generalization
and abstraction.

Lazy student formulas

Let / : [0, a] > R be any continuous function. Substitute u a x to obtain

pa pa
I f(x)dx = I f{
Jo Jo
Geometrically, the substitution simply reflects the graph
which obviously leaves the area under the curve invarian
Now suppose that / satisfies the following symmetr

f(x) + f(a-x) = l.
Integrating, we obtain the lazy student formula

ta a
\ f(x)dx = -.

This is fine, but how do we obtain functions / that satisfy this restrictive symmetry
condition? Set

f(x) =
g(x) + g(a - x)
where g is any continuous function on [0, a] such that the denominator of the above
does not vanish. It is easily verified that / satisfies the symmetry condition. Further
more, any function / satisfying the symmetry condition has this form, just take g = f.
Equation (2) is now obvious by taking g(x) = 3 + x^ and a = 10.
As a special case, note that sin(x) = cos(7i/2 x), to obtain the lazy student for
mula:

?7Z/2 f(cosx)dx IT
Jo /(sinjc) + /(cosjc) 4
where / is any continuous function defined on [0, 1], such that the denominator in the
above integrand does not vanish. This includes the integral in (1) as a special case.

Reference
James Stewart, Calculus, 5th ed., Brooks/Cole-Thomson Learning, Belmont, CA, 2003.

This content downloaded from

192.195.76.32 on Thu, 03 Oct 2024 07:38:28 UTC
All use subject to https://about.jstor.org/terms