PHYSICS FOR YOU Physics of animation and statistical Physics 1

Physics of Animation
Introduction:
Animation is a method of photographing successive drawings or models to create an illusion of
movement in a sequence. Since our eyes can retain an image for nearly 0.1 s, when multiple images
appear fast, the brain blends them into a single moving image.
In initial days animation was done by drawing or painting pictures on transparent celluloid sheets
and then photographed but today most of the animation work is done with computer-generated
imagery or CGI.

Animation is inherently fake. It’s not real actors on real sets—it is entirely manipulated from a
computer. Even though animators sometime break the laws of physics animation requires an
understanding of physics and an animator should have a basic understanding of mechanics and
bio-mechanics. Principles of physics are universal, they can be applied to cartoon-style drawings
as well as CGI and make audiences to escape reality and enter a fantasy world!

The Taxonomy of Physics-Based Animation Methods (classifications)

Since animation is an illusion of movement in sequence one need to consider the physics of motion.
There are two categories in this. They are

 Kinematics
The study of motion of bodies without considering the actual cause for the motion is known as
kinematics. Here mass and forces are not considered.

 Dynamics
The study of motion of bodies by considering the actual cause for the motion is known as dynamics.
Here mass and forces are taken into consideration

Joint angles
End effector

In both of these we have two subgroups. They are

 Inverse: It is the study of motion when both the starting and ending points are known. Here
one typically knows where to go, but needs to figure out how to do it. It is also known as
backward. Ex, in robots the motion of its joint angles in robots is inverse or backward

DR. NAGARAJA D 1
 PHYSICS FOR YOU Physics of animation and statistical Physics 2

 Forward: It is the study of motion when only the starting point is given. Here the goal is to
predict the final destination. Ex, end effector movement in the robots is forward motion. It
only moves towards the destiny

Frames and Frames per second (FPS)

In animation successive drawings or pictures are made. Each picture is treated as frame. To create
the appearance of motion from these images, frames shall be displayed.
Definition: The number of frames displayed in one second for smooth movement effect is known
as frame rate, or frames per second (FPS)
The human brain can process only about 10 to 12 FPS, ie, individuals can distinguish separate still
pictures in a series with a frame rate under 12 FPS. 12 frames do produce the motional effect but
may look choppy (recall very old black and white movies!!-may be Charlee Chaplin movies).
Hence the frame rate has to be increased to produce smooth moving effects. In cinema, a frame
rate of 24 frames per second is often used (Different video formats have different FPS). For TV
the frame rate is 30 FPS

24 FPS

12 FPS
Scale and Size

Animation of large objects (cricket field, railway stations, dams etc) is not done in their actual size.
Their size must be reduced several times. This is nothing but scaling. Scaling is not only done for
larger dimension objects. Even smaller dimensional objects (insects, body organs etc) also need
scaling.

Definition: The ratio of animation size to the actual size of the object is known as scale or scale
of animation. It is also defined as the ratio of pixels of the animation to the physical units of
length.

Scaling is not only done for altering the sizes of the characters but also done in other aspects such
as movement, energy etc. This is essential to distinguish between other parameters such as weight
and strength, younger and older etc.

DR. NAGARAJA D 2
Improper scaling Proper scaling
 PHYSICS FOR YOU Physics of animation and statistical Physics 3

For ex: If only the size of the cat and its kitten is scaled down then the kitten will look like small
adult cat only. Actually kitten will move slower than cat and has lesser strength. Hence the scaling
is needed to these factors also for better effects

Proportion and Scale (weight and strength)

When object is scaled its volume and area does not change in equal proportionate. Change in the
volume is more compare to area. Volume increases by cube times and area increases by square
times. Hence proportionate scaling is very important. Consider the following example.

The weight of a man depends on his body volume and the muscle strength depends on cross
sectional area. If we want to double the muscle strength the width must be increased 2 times. But
the weight is not scaled proportionately. Hence care has to be taken while creating larger or smaller
objects by means of scaling.

Motion and Timing in Animations

Motion is an essential part of animation. Several types of motion need to be considered while
animating a scene. Common types of motion are
 Linear motion
 Parabolic motion
 Circular motion
 Wave motion or oscillatory motion

Line of action and path of action

All these types are familiar to us. In connection with animation of motion, we define two
parameters namely line of action and path of action.

Line of action is an imaginary line that describes the direction and motion of a character’s body.
A good line of action improves character’s poses, makes the character look more dynamic,
energetic and alive. There are 3 types of lines of action, the C, reverse C and S curves. Look at the
following fig. It is clear that the second pose gives better impression than first one

DR. NAGARAJA D 3
 PHYSICS FOR YOU Physics of animation and statistical Physics 4

The path along which the object or character moves is known as path of action or arc. For ex, path
of bouncing ball, trajectory of projectile, jumping of ballet dancer, path of moving arm etc.

Arcs are used extensively in animation because they create motion that is more expressive and less
stiff than action along a straight path.

Timing
Timing refers to the time it takes for an action to complete from the starting point to the end.
Timing is the amount of frames it takes for an action to take place.

Timing can have a huge effect on how we perceive a character. If a character moves with less
frames rate it will appear to be heavy and likely very big. If a character has high frames rate it will
appear to be light and probably small. It is explained in the following example

Consider three blocks of same size. 1st block is given 6 frames rate, 2nd block is given 12 frames
rate and 3rd one is given 24 frames rate. As a result 1st block falls fast giving the illusion of being
heavy

Timing of action consists of placing objects or characters in particular locations at specific frames
to give the illusion of motion. If images are placed at same location in all the frames it won’t give
motion effect (fig-A). Hence we work with very small intervals of time (fig-B).

Spacing
Spacing is the distance an object moves within a defined time for that action (it is also the distance
an object moves for every frame of that action)

DR. NAGARAJA D 4
 PHYSICS FOR YOU Physics of animation and statistical Physics 5

The main difference between timing and spacing is that timing refers to everything that happens
over time whereas spacing refers to how much fast something moves. It is illustrated as follows

Consider two balls moving same horizontal distance in same time. Assume the top ball will be
moving slowly to begin with, then speed up and then slow down again at the end. The bottom ball
will just begin and keep a constant speed until the end. Both will be in the same position at the
beginning, middle and end of the animation. To achieve this we are going to change the spacing.
If we have more drawings near the starting pose, one or two in the middle, and more drawings near
the next pose lesser drawings make the action faster and more drawings make the action slower.

Linear motion, Uniform motion and timing

A body moving in a straight line is said to be in linear motion. It always move along same direction.
Ex: a ball rolling on the inclined plane, a ball moving on a horizontal plane, a stone falling under
gravity etc

In linear motion if the body travels without acceleration (or deceleration) then it is known as
uniform motion. Here the net force acting on the object is zero and distance travelled between
consecutive frames is same. Hence timing is very easy. If the speed is less the spacing between the
frames is less. If the speed is more the spacing is also more

DR. NAGARAJA D 5
 PHYSICS FOR YOU Physics of animation and statistical Physics 6

When the net force acting on the object is not equal to zero then according to Newton’s second
law, there will be acceleration (even deceleration also). As a result the object will get faster and
faster until the force stops acting on it. Now comes one of the Principles of Animation. It is called
"slow-in and slow-out".

Slow-in and slow-out

All types of movements start with acceleration and end with deceleration. Even in animation also
this principle is followed. An object or a character in animated video starts its movements more
slowly, then picks up speed and finishes with deceleration. As a result the beginning and end of
the movement are softened. This is called ‘slow-in’ (or ease in) &’slow-out’(or ease out)

.
Slow-in (or ease in) is process in which the body is preparing for stopping and Slow-out (or ease
out) is the process in which the body is speeding up from a still position.

Slow-in and slow-out is achieved by adjusting the spacing (as explained in the above section). To
create a slow-out, we need to place the frames close together and then move them apart gradually.
To create the slow-in we will need to gradually place the frames closer together at the end of the
animation.

Consider the animation of oscillating pendulum. There are more circles at slow-in and slow-out
positions and less at middle positions.

Constant force, acceleration and timing

If the applied force is not varying with time then it is called constant force. Some of the examples
for constant force are gravitational force, frictional force, force on a charge in uniform electric
field, force on a charge in uniform magnetic field etc.

DR. NAGARAJA D 6
 PHYSICS FOR YOU Physics of animation and statistical Physics 7

If the force is varying with respect to time then it is variable force. Ex, force on a charge in variable
magnetic and electric field, spring force etc. Timing of the action is very difficult for motions due
to variable forces.

Timing for variable force

Normally all the forces exerted by us in everyday life are variable. Force applied during walking,
running, jumping, skating, dancing etc is a variable one. If you break the motion into smaller
segments then force in each segment is treated as constant and timing becomes easier.

In the below fig, jumping (by girl) is divided into 3 segments. Force is nearly constant over each
segment. The timing is very short for each of these segments

Timing for constant force and acceleration

If the applied force is constant then the body will move with constant acceleration. The acceleration
depends on the direction of the applied force. There are 3 possibilities

i. If a constant force is applied on a body at rest then it accelerates

ii. If a constant force is applied in the direction of motion then the body is accelerated
iii. If a constant force is applied in the opposite direction to motion then it is decelerated.

For such type of motions timing is done with odd rule

Odd rule

The distance travelled by the object between two successive frames is calculated by odd rule.
According to this rule

“The distance travelled between two successive frames is obtained by multiplying the base
distance by odd numbers 1,3,5,7 etc during accelerated motion and by multiplying the base
distance by odd numbers 7,5,3,1 etc during decelerated motion”

DR. NAGARAJA D 7
 PHYSICS FOR YOU Physics of animation and statistical Physics 8

Base distance: Base distance is the smallest distance between the two frames. For a slow-out,
this is the distance between the first two frames; for a slow-in, it’s the distance between the last
two frames. It is given by
total distance
Base distance =
(last frame number − 1)2
Explanation:
This rule is applicable to both vertical and horizontal motion. It is explained
for a body falling under gravity. Consider a ball falling under gravity with zero
initial velocity. Distance travelled
1 1 1
after 1 sec 𝑆1 = 𝑔𝑡 2 = 𝑔 = 1 × ( 𝑔)
2 2 2
1 2 1
after 2 sec 𝑆2 = 𝑔𝑡 = 4 × ( 𝑔)
2 2
1 2 1
after 3 sec 𝑆3 = 𝑔𝑡 = 9 × ( 𝑔)
2 2
1 2 1
after 4 sec 𝑆4 = 𝑔𝑡 = 16 × ( 𝑔)
2 2
and so on

Here the difference between any two successive distances is an odd number.
For ex, 4-1 = 3; 9-4 =5; 16-9 = 7 and so on. Hence the rule is named as odd rule

Odd rule multipliers

The process of calculating the distance from the first frame to the current frame and use these
distances to place the object on specific frames is known as odd rule multiplier. For slow-out, it
is as follows

i. Find the distance between first two frames. It is also known as “base distance”
ii. Multiply it by 1 to get the distance between frame 1 and 2. The total distance travelled
is 0+1=1
iii. Multiply by 3 to get the distance between frame 2 and 3. The total distance travelled is
0+1+3=4
iv. Multiply by 5 to get the distance between frame 3 and 4. The total distance travelled is
0+1+3+5=9
And so on…It is summarized as follows

Frame No. Distance between Distance between 1st

(A) successive frames (B) to present frames (C)
1 -- 0
2 1 0+1=1=12
3 3 0+1+3=4=22
4 5 0+1+3+5=9=32
5 7 0+1+3+5+7=16=42
6 9 0+1+3+5+7+9=25=52

DR. NAGARAJA D 8
 PHYSICS FOR YOU Physics of animation and statistical Physics 9

The same is represented in the following fig. From fig, it may be observed that after the first
increment, all the others are longer by the same amount (amount of 2- Red arrow mark).
Ie, 3 – 1 =2; 5 – 3 = 2; 7 – 5 = 2 and so on

It is possible to write the relation between A, B and C as follows

𝐵 = 2(𝐴 − 1) − 1
𝐶 = (𝐴 − 1)2
Ex:
If A = 5 then B = 2(5-1)-1 = 7

C = (5 – 1)2 = 16

This is true even for a horizontal motion also

The whole process is applicable for ‘slow-in’ process also (decelerated motion). Only difference
is that we have to multiply in the reverse manner ie, 7, 5, 3, 1.

DR. NAGARAJA D 9
 PHYSICS FOR YOU Physics of animation and statistical Physics 10

Odd Rule Scenarios

Here we summarize the odd rule for 4 different cases.

1. Slow-out process

First of all, find the distance between first two successive frames. It is also known as base distance.
Then multiply it by odd numbers 1,3,5,7 etc to get the distances between consecutive frames.
Finally use squares to multiply the base distance to get the total distance traveled on each frame
(See the above table)
If the total distance is known then the base distance is calculated using the formula

total distance
Base distance =
(last frame number − 1)2

For ex: In the above fig, the total distance is 2m and number of frames is 5 then
2
Base distance = = 0.125𝑚
(5 − 1)2

Distance between frame 1 and 2 = 1 × 0.125m = 0.125 m

Distance between frame 2 and 3 = 3 × 0.125m = 0.375 m
Distance between frame 3 and 4 = 5 × 0.125m = 0.625 m
Distance between frame 4 and 5 = 7 × 0.125m = 0.875 m

Total distance = 0.125+0.375+0.625+0.875 = 2m

2. Slow-in process

First, find the distance between the last two frames. Then multiply it by odd numbers in reverse
order, ie 7,5,3,1
If the total distance is known then the base distance is calculated using the formula

total distance
Base distance =
(last frame number − 1)2

DR. NAGARAJA D 10
 PHYSICS FOR YOU Physics of animation and statistical Physics 11

In the above fig, the total distance is 0.9 m and number of frames is 4 then
0.9
Base distance = = 0.1𝑚
(4 − 1)2

Distance between frame 4 and 3 = 5 × 0.1m = 0.5 m

Distance between frame 3 and 2 = 3 × 0.1m = 0.3 m
Distance between frame 2 and 1 = 1 × 0.1m = 0.1 m

Total distance = 0.5 + 0.3 + 0.1 = 0.9 m

3. Jumping up process

Consider a girl jumping up. The whole process is divided into 3 small intervals (or 3 frames). Let
the total height is 0.3 m, then

0.3
Base distance = = 0.075𝑚
(3 − 1)2

Distance between frame1 and 2 = 1 × 0.075m = 0.075 m

Distance between frame2 and 3 = 3 × 0.075m = 0.0225 m

Total distance = 0.075 + 0.225 = 0.3 m

4. Jumping down process

Consider a girl jumping down through a distance 0.54 m. The whole process is divided into 4
frames, then
0.54
Base distance = = 0.06𝑚
(4 − 1)2

Distance between frame 4 and 3 = 5 × 0.06m = 0.3 m

Distance between frame 3 and 2 = 3 × 0.06m = 0.18 m
Distance between frame 2 and 1 = 1 × 0.06m = 0.06 m

Total distance = 0.3 + 0.18 + 0.06 = 0.54 m

DR. NAGARAJA D 11
 PHYSICS FOR YOU Physics of animation and statistical Physics 12

Motion Graphs
A graph of distance versus time is known as motion graph. It is also called as position-time graph
or x-t graph. It is very important for animation because the nature of the motion is ascertained from
the graph. Some of the motion graphs are as follows

Fig. (i) represents constant velocity and constant slope. There is no acceleration
Fig. (ii) represents positively increasing slope and velocity. Hence acceleration is positive. More
the acceleration more will be the curvature
Fig.(iii) represents negatively increasing slope. It represents decreasing velocity. Hence it is
deceleration

Examples of Character Animation

Motion of an object is an integral part of animation. By observing and studying the real life
movements one can make the animation more attractive and lively. Jumping and walking are two
real life actions where the entire body is in motion. Now we shall take up their character animation
by applying the knowledge of physics in animation

1. Jumping

Following are the different stages of jumping process.

 Crouch—It is the bending pose taken as preparation for jumping.

 Takeoff —Character pushes up fast and straightens legs with feet still on the ground. The
distance from the character’s center of gravity (CG) in the crouch to the CG in takeoff
position is called the push height. The amount of time (or number of frames) needed for
the push is called the push time.

 In the air—Both the feet are off the ground, and the character’s center of gravity (CG)
moves in a parabolic arc. First it reaches a maximum height (apex), and then falls back to
the ground. The maximum height or jump height, is measured from the CG at takeoff to
the CG at the apex of the jump. The amount of time the character is in the air from takeoff
DR. NAGARAJA D 12
 PHYSICS FOR YOU Physics of animation and statistical Physics 13

to apex is called the jump time. If the takeoff pose and the landing pose are similar, then
jump time is same for going up and coming down (it is exactly similar to time of ascent
and time of descent in projectile motion).

 Landing—Character touches the ground and bends knees to return to a crouch. The
distance from the character’s CG when her feet hit to the ground to the point where the
character stops crouching is called the stop height. The stop height is not always exactly
the same as the push height

 Path of action—It is the trajectory along which the character jumps

Calculations
i. Calculation of jump timing
Let the jump height be 1.2 m and acceleration due to gravity is 9.8 ms-2. The jump time is
2ℎ 2 × 1.2
𝑡2 = = = 0.25
𝑔 9.8
∴ 𝑡 = 0.5 𝑠
This is the time taken to go from crouch position to apex position. As we know that the frame rate
is 24 (FPS), the total number of frames for this action should be 24 × 0.5 = 12 frames
ii. Jump magnification (JM)
The JM is the ratio of the jump height to the push height.
jump height
𝐽𝑀 = − − − (1)
push height
It is also defined as the ratio of jump time and push time. It is used to calculate the push timing
and stop timing.

DR. NAGARAJA D 13
 PHYSICS FOR YOU Physics of animation and statistical Physics 14

jump time
𝐽𝑀 = − − − (2)
push time
In terms of acceleration it is given by
push acceleration
𝐽𝑀 = − − − (3)
𝑔
[To prove Eq (3): (Optional – not in syllabus)
Jumping motion is equivalent to projectile motion. While going from crouch to take off position the girl will
have some acceleration called push acceleration. After wards her body moves upward and experiences
acceleration due to gravity. During downward motion she experiences gravity only.
Jump velocity = push acceleration × push time (because v = at)
Landing velocity = landing acceleration × jump time
But the jump velocity shall be equal to landing velocity. Hence
push acceleration × push time = landing acceleration × jump time
Since landing acceleration is same as g we can write
push acceleration × push time = g × jump time
jump time push acceleration
=
push time 𝑔
push acceleration
𝐽𝑀 =
𝑔

Hence the proof ]

Ex, if there are 15 frames and JM is 3 then
According to Eq (2)
jump time 15
𝐽𝑀 = =
push time push time
15
push time = = 5 frames
3
iii. Jump height
The distance between CG at crouch and take off position is called push height. Assume that it is
0.4 and JM is 3 then from eq (1) we get
jump height jump height
𝐽𝑀 = =
push height 0.4
jump height = JM × 0.4 = 3 × 0.4 = 1.2 m
iv. Stop time
Since timing of the push and stop are same we can write

DR. NAGARAJA D 14
 PHYSICS FOR YOU Physics of animation and statistical Physics 15

push height stop height

=
push time stop time
push time
stop time = stop height ×
push height
It may be noted here that the stop height is little bit longer than push height (because the girl may
finally lay down on the ground after landing). For ex, if push height is 0.4, stop height is 0.5, push
time 5 frames then

5
stop time = 0.5 × ≈ 6 𝑓𝑟𝑎𝑚𝑒𝑠
0.4
2. Walking
Walking is another event where a lot of physics is involved and the animation of walking needs
the complete knowledge of mechanics. Walking is nothing but a step-by-step movement. Each
step is divided into 4 poses namely passing, step, contact, and lift. Also we need the concept of
Strides and steps as well as walking time
Strides and Steps
A step means one step with one foot. A stride means two steps, one with each foot. Step length is
the distance between two successive steps and Stride length is the distance between two parts of
the same step. Step and stride length indicate lengthwise spacing for the feet during a walk.

Walk timing
While walking each foot is in contact with ground for 60% of time (ie, single support) and both
the feet will be in contact for 20 % of time (double support). In this connection we define a quantity
called Gait. It is the timing of the motion for each foot, including how long each foot is on the
ground or in the air. For normal walk it is about ½ second
To walk faster, obviously one has to increase stride length and decrease the time of double support.
A fast walk has a stride rate of about 4 feet/sec. If it becomes 6-7 then walking turns into running.

DR. NAGARAJA D 15
 PHYSICS FOR YOU Physics of animation and statistical Physics 16

Statistical Physics
Introduction

Statistics is the area of applied mathematics that deals with the collection, organization, analysis,
interpretation, and presentation of data.

For ex, in the study of tiger population in a particular forest region, one has to use the statistical
methods such as collecting all possible data (like number of tigers, deer, monkeys, area of the
grass field etc) over definite period of time, arrange and compare it with the previous year’s data,
identify the differences and analyze, find out the reasons for the population variation, interpret
the result and present it in a convincing manner

Statistical Physics

Statistics is equally important in Physics and it is called as statistical physics. It is basically

evolved from statistical mechanics

Statistical Physics is mainly intended to explain the properties of matter in total, in terms of
physical laws governing atomic motion. It also explain numerous physical phenomena, such as
superconductivity, magnetism, changes in states of matter, energy state distribution, electron
distribution in metals, radioactivity, noise in electronic devices etc,

[Kinetic theory of gases is one of the best example in which macroscopic parameters such as
kinetic energy, pressure and temperature are explained in terms of microscopic quantities-number
of molecules and their rms velocity]

Descriptive statistics and inferential statistics

Statistical techniques can be categorized as descriptive statistics and inferential statistics

The process of organizing and describing the known data using charts, bar graphs, etc., is known
as descriptive statistics. It mainly focuses on describing the visible characteristics of a database.
It enables researchers to present data in a more meaningful way such that easy interpretations can
be made.

The process of making predictions or inferences (conclusions) and generalizations about the data
is known as inferential statistics. When the data is very large it becomes difficult to use it. In such
cases, certain samples are taken as representative of the entire population. Inferential statistics
draws conclusions using these samples.

DR. NAGARAJA D 16
 PHYSICS FOR YOU Physics of animation and statistical Physics 17

Both descriptive and inferential statistics need to be used hand in hand so as to analyze the data in
the best possible way.

Ex:
Assume that you have the marks scored by 100 students (of a specific school where the total
strength is 5000).
By using descriptive statistics, the spread of the marks can be done and can obtain a clear idea
regarding the performance of each student.
Inferential statistics will now make use of this sample data of 100 students and make
generalizations about the population (ie all 5000 students).

Poisson’s distribution

Introduction
Look at these examples

i. Count the number of accidents occurring in a metropolitan city in one year

ii. Count the number of deaths occurred in frequently flooded area
iii. Count the number of people visit a shopping mall a day
iv. Count the number of radioactive nuclei disintegrate in a sample
v. Count the number of collisions between electrons and lattice points in a sample of
given volume
vi. Count the number of website visitors per month

In all these examples we are counting some of the events (green color) occurring within a given
interval of time or space or volume. Events are occurring randomly and independently. By
knowing ‘when one event happens’ gives absolutely no information about ‘when another event
will occur’. For ex, if one U-235 nuclei disintegrates now, it is impossible to predict when the
other one will disintegrate (same thing is true in all the above examples). In such cases the
probability of a given number of events occurring in a fixed interval of time or space is explained
using Poisson’s distribution function

Definition and explanation

Poisson’s distribution is a discrete distribution and describes probabilities for counts of events
occurring in a specific interval of time or space. Counts may be random but over certain interval
of time it will be a ‘definite average’. The discrete outcome is a number represented by k. It is
always non negative whole number.

The Poisson distribution mass function is given by

𝜆𝑘 𝑒 −𝜆
𝑃(𝑋 = 𝑘) =
𝑘!
DR. NAGARAJA D 17
 PHYSICS FOR YOU Physics of animation and statistical Physics 18

Where
X is a random variable following a Poisson distribution
k is the number of times an event occurs in given interval of time or space
P is the probability that an event will occur k times
 is the average number of times an event occurs

Ex:
One of the first applications of the Poisson distribution was by statistician Ladislaus Bortkiewicz.
In the 1870s he investigated accidental deaths by horse kick of soldiers in the Prussian army.
He analyzed 20 years of data for 10 army corps (batch), (equivalent to 200 years of observations
of one corps)
He found that a mean of 0.61 soldiers per corps died from horse kicks each year. However,
most years, no soldiers died from horse kicks. On the other end of the spectrum, one tragic year
there were four soldiers in the same corps who died from horse kicks.
Using modern terminology:
 A death by horse kick is an “event.”
 The time interval is one year.
 The mean number of events per time interval, λ, is 0.61.
 The number of deaths by horse kick in a specific year is k (=4).

The Poisson distribution may be useful to model events such as the radioactive decays, number of
laser photons hitting a detector in a particular time interval, thermionic emission, scattering of
particles etc

A graph of P v/s k is known as Poisson distribution curve. It is as follows. When λ is low, the
distribution is much longer on the right side of its peak than its left. As λ increases, the distribution
looks more and more similar to a normal distribution

Ex.1. On a particular river, overflow floods occur once every 100 years on average. Calculate
the probability of k = 0, 1, 2, 3, 4, 5, or 6 overflow floods in a 100 year interval, assuming the
Poisson model is appropriate.
Ans:
 = 1 (ie the average flood over 100 years occurs only once)
DR. NAGARAJA D 18
 PHYSICS FOR YOU Physics of animation and statistical Physics 19

𝜆𝑘 𝑒 −𝜆
𝑃(𝑋 = 𝑘) =
𝑘!
Let k = 0
10 𝑒 −1
𝑃(𝑋 = 0) = = 0.368
0!
Let k =1
11 𝑒 −1
𝑃(𝑋 = 1) = = 0.368
1!
Let k =2
12 𝑒 −1
𝑃(𝑋 = 2) = = 0.184
2!

Ex.2. Careful measurements have established that a sample of radioactive thorium emits
alpha particles at a rate of 1.5 per min. what is the expected average result in 2 min interval?
What is the probability of getting this number? What is the probability for observing k
particles for k = 0, 1,2,3,4 & 5? (Here k is the number of counts or number of particles decay
in any 2 min interval)

Here  = 1.5 × 2 = 3
𝜆𝑘 𝑒 −𝜆
𝑃(𝑋 = 𝑘) = 25
𝑘!
20
15
Let k = 0 (ie, no decay in 2 min interval) and  = 3
P
10
30 𝑒 −3 5
𝑃(𝑋 = 0) = = 0.049 = 4.9%
0! 0
Let k =1 (ie, 1 decay in 2 min interval) and  = 3 1 2 3 4 5 6
31 𝑒 −3 k
𝑃(𝑋 = 1) = = 0.149 = 14.9%
1!
Let k =2 (ie, 2 decay in 2 min interval) and  = 3
32 𝑒 −3
𝑃(𝑋 = 2) = = 0.224 = 22.4%
2!
Let k =3 (ie, 3 decay in 2 min interval) and  = 3
33 𝑒 −3
𝑃(𝑋 = 3) = = 0.224 = 22.4%
3!
Let k =4 (ie, 4 decay in 2 min interval) and  = 3
34 𝑒 −3
𝑃(𝑋 = 4) = = 0.168 = 16.8%
4!
Let k =5 (ie,5 decay in 2 min interval) and  = 3
35 𝑒 −3
𝑃(𝑋 = 5) = = 0.100 = 10%
5!
These results are plotted as shown

DR. NAGARAJA D 19
 PHYSICS FOR YOU Physics of animation and statistical Physics 20

Ex:3 On average each of the 18 hens lays 1 egg per day. If you check the hens once an hour
and remove any eggs that have been laid what is the average number  of eggs you find on
hourly visit?. Use Poisson distribution to calculate probability that you find k eggs for k = 0,
1 , 2, 3 and 4

Here 18 hens lay one egg each in 24 hours. Hence the average count per hour is 18/24 = 0.75
ie  = 0.75
𝜆𝑘 𝑒 −𝜆
𝑃(𝑋 = 𝑘) =
𝑘!
50
40
Let k = 0 (ie, no egg in 1hr interval) and  = 0.75 30

P
0 −0.75
0.75 𝑒 20
𝑃(𝑋 = 0) = = 0.472 = 47.2%
0! 10
0
Let k =1 (ie, 1 egg in 1hr interval) and  = 0.75
1 2 3 4 5
1 −0.75
0.75 𝑒 k
𝑃(𝑋 = 1) = = 0.354 = 35.4%
1!
Let k =2 (ie, 2 egg in 1hr interval) and  = 0.75
0.752 𝑒 −0.75
𝑃(𝑋 = 2) = = 0.133 = 13.3%
2!
Let k =3 (ie, 3 egg in 1hr interval) and  = 0.75
0.753 𝑒 −0.75
𝑃(𝑋 = 3) = = 0.033 = 3.3%
3!
Let k =4 (ie, 4 egg in 1hr interval) and  = 0.75
0.754 𝑒 −0.75
𝑃(𝑋 = 4) = = 0.006 = 0.6%
4!

Modeling the Probability for Proton Decay

Background:
In Physics there are some theoretical conservation laws which forbid certain specific reactions or
decays to occur in nature. Many Experiments are intended to check the validity of these laws by
searching for the presence (or absence) of such reactions or decays. If one or more events are
observed in time T interval then the theoretical law is disproven. If no events are observed, the
converse cannot be said to be true. Instead, a limit on the life-time of the reaction or decay is set.
(This is a sort of Null Experiment aimed at setting confidence limits when no counts are observed).
One such example is life time of proton or proton decay

DR. NAGARAJA D 20
 PHYSICS FOR YOU Physics of animation and statistical Physics 21

Proton decay

Proton decay is a theoretical form of particle decay. Here proton disintegrates into two particles
namely pion and a positron with mean life of 1033 years which is very much greater than the age
of the universe!. (Age of the universe is about 1 × 1010 years). If you have only one proton then
you will have to wait for 1033 years to see the decay. But we know the decay is a statistical process.
If there are huge number of protons then some of them may decay (if at all the decay is possible!)

Actually conservation laws forbid the proton decay. Despite many serious experiments and
significant effort across the globe, even a single proton decay has never been observed so far. One
such experiment is explained here

In Super-K detector (JAPAN) 50,000 tons of ultra-pure water containing 7×1033 protons is used.
The 40% of the experimental area is covered by detector tubes to count the ‘event’-decay

Consider the decay law

𝑁 = 𝑁0 𝑒 −𝜆𝑡
Where
N0 = initial number of protons at t = 0 = 7×1033
N = number of protons remain un decayed after t years
(N0 – N) = number of protons decayed in t years
 = decay constant = 1/ (not to be confused with  in Poisson’s formula) = 1/1033 = 10-33

Since  is very small we can approximate 𝑒 −𝜆𝑡 = 1 − 𝜆𝑡

𝑁 = 𝑁0 (1 − 𝜆𝑡) = 𝑁0 − 𝑁0 𝜆𝑡
(𝑁0 − 𝑁) = 𝑁0 𝜆𝑡
Put N0 = 7×1033, t = 1 year and  = 10-33 we get
(𝑁0 − 𝑁) = 7 × 1033 × 10−33 = 7
So, 7 decays (events) shall occur in every one year. Since there are 40% detectors, the average
detected count must be 40% of 7 ie 2.8 or nearly 3. Hence the mean number of events =  = 3.
Actually Super-K has started measurement since 1996 but not observed any evidence of proton
decay yet. Hence k = 0.
Substitute  = 3 and k = 0 in Poisson’s formula we get
𝜆𝑘 𝑒 −𝜆
𝑃(𝑋 = 𝑘) =
𝑘!
30 𝑒 −3
𝑃(𝑋 = 0) = = 0.05
0!

DR. NAGARAJA D 21
 PHYSICS FOR YOU Physics of animation and statistical Physics 22

From this low probability for null result (i.e. no decay) Super-K scientists have come to a
conclusion that the mean life of proton should be at least 2 × 1034 years! Does it mean that protons
live longer than estimated time? or, doesn’t decay? Nobody knows the answer! Many countries
have taken up this as a challenge and conducting deep underground neutrino experiment (Including
India- long back such kind of experiment was initiated near KGF)

Normal distribution (Gaussian distribution) and Bell curve

The Normal distribution is one in which the data tends to be distributed equally around a central
value with no bias left or right. It is the most universal distribution and plays a central role in all
of statistics. The distribution curve assumes bell shape and hence the curve is called “Bell curve”.
The peak of the curve represents the mean or median

Examples for normal distribution are

 heights of people
 size of things produced by machines
 errors in measurements
 blood pressure
 marks on a test

The Poisson distribution is applied to experiments where the data is strictly bounded on one side;
the curve of the graph is highly asymmetric. It is approximated to normal distribution when P, the
probability, and , the mean number of events are large. The Poisson distribution becomes more
symmetric and assumes bell curve shape (as shown in fig)

Monte Carlo Simulation

Introduction
In certain experiments the possible outcome is random and cannot be easily predicted. The
outcome takes different values in each experimental trial. Hence the outcomes are called random

DR. NAGARAJA D 22
 PHYSICS FOR YOU Physics of animation and statistical Physics 23

variables. Random variables are always real numbers. The likelihood that any of these possible
values would occur is known as probability distribution.

The concept of random variable and probability distribution are better understood through these
examples

1. Consider the example of tossing of two coins. The probability of getting HEADS (or even TAILS)
is a random event and it is not likely to happen equally in all the trials. If we take Y as the random
variable (ie, getting H) it is either 0 (ie, no head), 1 (ie, one head) or 2 (ie two heads). Since the
coins can land in 4 possible ways namely TT, TH, HT and HH then probability distribution of

i. P(Y = 0) = 1/4 ----- ie both coins land with TT

ii. P(Y = 1) = 1/2 ------ ie, either TH or HT
iii. P(Y = 2) = 1/4 ------ ie, both coins with HH

When you add all the probabilities we get 1, ie 1/4 + 1/2 + 1/4 =1

2. Consider the rolling of a dice. Let’s say that the random variable, Y, is the number on the top
face of a die when it is rolled once. The possible values for Y will thus be 1, 2, 3, 4, 5, and 6. The
probability (P) of each of these values is 1/6 as they are all equally likely to be the value of Y

3. Consider the rolling of three dice. Let’s say that the random variable Y may be the sum of the
resulting numbers after three dice are rolled. Y could be 3 (1 + 1+ 1) or18 (6 + 6 + 6) or somewhere
between 3 and 18 since the highest number of a die is 6 and the lowest number is 1.

In such types of uncertain processes the possible outcome is estimated using Monte Carlo method
(also called as Monte Carlo simulation)
Monte Carlo method
Monte Carlo method is a mathematical technique used to estimate the possible outcomes of an
uncertain events through repeated random sampling. It finds its application in many fields
including business, physics, artificial intelligence, astronomy and engineering.
NOTE: The technique was initially developed by Stanislaw Ulam, a mathematician who worked on the
Manhattan Project, the secret project on nuclear bomb during II world war

How does it works?

In this method, the repeated random sampling and recalculating the results again and again is very
important. However each time use different set of random numbers between the minimum and

DR. NAGARAJA D 23
 PHYSICS FOR YOU Physics of animation and statistical Physics 24

maximum values. This exercise can be repeated thousands of times (or even more) to obtain a
large number of likely outcomes. Monte Carlo techniques involves four basic steps

1. Set up the predictive model (domain identification), identifying both the dependent variable
and the independent variables (also known as the input, risk or predictor variables)
2. Specify probability distributions (generate unbiased inputs)of the independent variables.
3. Run simulations repeatedly (repeated random sampling), generating random values of the
independent variables.
4. Do this until enough results are gathered to make up a representative sample of the near
infinite number of possible combinations.

For ex, you want to find the average height of humans in India. The most accurate method is to
find the height of country’s entire population and take the average. Any way it is impossible.
What is the next alternate? Find the height of few people randomly (say height of 100 members)
and then take the average. If you do this only, say, in Karnataka the output (answer) is not a
reliable one. Then what to do?? Follow these steps

i. Identify few states covering North, East, South and West part of the country(Identify
the domains of possible inputs)
ii. Since height has inherent uncertainty, one has to measure the heights of selected
people within the selected state (generate the inputs)
iii. The sample selection must remain unbiased. If surveyors collect samples of tall or short
people, it will not give accurate results. Hence, the correct data would be obtained only
through fair sample selection using a probability distribution.(generate the inputs
randomly from a probability distribution over the domain)
iv. Take the average of the measured values.
v. Perform this procedure over and over, each time using a different set of random
numbers between the minimum and maximum values (Repeated random sampling))
vi. As samples would represent many people, data collectors need to use more and more
people as random samples. The greater the number of people, the better and more
accurate range of results will be.(repetition of the exercise thousands of times to
produce a better outcomes)

To determine the value of 

One of the basic examples of getting started with the Monte Carlo method is the estimation of 
(pi). It is as follows

DR. NAGARAJA D 24
 PHYSICS FOR YOU Physics of animation and statistical Physics 25

1. Consider a square of side length l = 2m and enclosing a circle of radius r = 1m. The center
of both circle and square coincides at (0,0)
2. Generate a large number of uniformly distributed random points. These points can be in
any position within the square i.e. between (0,0) and (1,1).
3. Keep track of the total number of points (NT), and the number of points that are inside the
circle (Ni)
4. We can show that the ratio of Ni and NT is equal to the ratio of area of the circle and the
square

𝑁𝑖 𝑎𝑟𝑒𝑎 𝑜𝑓 𝑡ℎ𝑒 𝑐𝑖𝑟𝑐𝑙𝑒 𝜋𝑟 2

= = 2
𝑁𝑇 𝑎𝑟𝑒𝑎 𝑜𝑓 𝑡ℎ𝑒 𝑠𝑞𝑢𝑎𝑟𝑒 2
𝑁𝑖 𝜋
=
𝑁𝑇 4
𝑁𝑖
𝜋 = 4×
𝑁𝑇

5. By substituting the values of Ni and NT in the above equation we get the value of 
6. When we only have a small number of points, the estimation is not very accurate, but when
we have hundreds of thousands of points, we get much closer to the actual value - to within
around 2 decimal places of accuracy. It can be verified using following examples

i. If Ni = 169, & NT = 200 then  = 3.38

ii. If Ni =468, & NT = 600 then  = 3.12
iii. If Ni = 7067, & NT = 9000 then  = 3.14

Additional information:

DR. NAGARAJA D 25
 PHYSICS FOR YOU Physics of animation and statistical Physics 26

Image of Super-K plant

Some of the salient features of Super-K experiment in Japan

1. A cylindrical tank is buried 1km underground

2. The tank is 40 meter tall and 130ft radius
3. This tank contains 50,000 tons of ultra-pure water
4. 40% of the tank is covered by the detectors
5. In 1996 the scientists of super-K turned their detector ON and started looking for proton
decay
6. Even today (after 27 years) not a single decay is detected!

DR. NAGARAJA D 26