Dr. Prapun Suksompong [email protected]
Introduction
2
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Course Organization
4
Course Website: http://www2.siit.tu.ac.th/prapun/ecs315/
Lectures:
Tuesday 13:00-14:20 BKD 3206
Thursday 10:40-12:00 BKD 3206
Tutorial/make-up sessions: Monday 10:40-12:00 BKD 3206
Textbook:
Probability and stochastic processes: a friendly introduction for electrical and computer engineers By Roy D. Yates and David J. Goodman
2nd Edition
ISBN 978-0-471-27214-4
Library Call No. QA273 Y384 2005
Student Companion Site: http://bcs.wiley.com/he-bcs/Books?action=index&itemId=0471272140&bcsId=1991
10
Probability
11
“The most important questions of
life are, for the most part, really only
problems of probability.”
Pierre Simon Laplace (1749 - 1827)
“Les questions les plus importantes de la vie ne sont en effet, pour la plupart, que des problèmes
de probabilité.”
12
“One sees, from this Essay, that the theory of probabilities
is basically just common sense reduced
to calculus; it enables us to appreciate with exactness that which
accurate minds feel with a sort of instinct, often without being able to account for it.”
Pierre Simon Laplace (1749 - 1827)
“On voit, par cet Essai, que la théorie des probabilités n'est, au fond, que le bon sens réduit au
calcul; elle fait apprécier avec exactitude ce que les esprits justes sentent par une sorte
d'instinct, sans qu'ils puissent souvent s'en rendre compte.”
Levels of Study in Probability Theory
13
Probability theory is the branch of mathematics
devoted to analyzing problems of chance.
Art of Guessing
1. High School: classical
2. Undergraduate: calculus
3. Graduate: measure-theoretic
We are here!
More references
14
Use ones that say probability
and random (or stochastic)
processes
If it has the word “statistics”
in the title, it may not be
rigorous enough for this
class
If it has the word “measure”
or “ergodic” in there, it is
probably too advanced.
The Drunkard's Walk
15
The Drunkard's Walk: How
Randomness Rules Our Lives
By Leonard Mlodinow
Deals with randomness and people's
inability to take it into account in
their daily lives.
A bestseller, and a “NY Times
notable book of the year”
Named “one of the 10 best science
books of 2008” on Amazon.com.
[Thai Translation: ชีวิตนี้ ฟ้าลิขิต: การสุ่มเลือก ควบคุมบญัชา ทุกเร่ืองราวในชีวิตของเรา]
Leonard Mlodinow
16
Euclid’s Window: the Story of Geometry from Parallel Lines to Hyperspace
Feynman’s Rainbow: a Search for Beauty in Physics and in Life
A Briefer History of Time
with Stephen Hawking
an international best-seller that has appeared in 25 languages.
The Drunkard's Walk: How Randomness Rules our Lives
Apart from books on popular science, he also has been a screenwriter for television series, including Star Trek: The Next Generation and MacGyver.
Watch Mlodinow’s talk
17
Delivered to Google employees
About his book (“The Drunkard's Walk”)
http://www.youtube.com/watch?v=F0sLuRsu1Do
18
Examples Prelude to the Theory of Probability
19
Game 1:
Seven Card Hustle
The Seven Card Hustle
20
Take five red cards and two black cards from a pack.
Ask your friend to shuffle them and then, without looking at the faces, lay them out in a row.
Bet that they can’t turn over three red cards.
Explain how the bet is in their favor.
The first draw is 5 to 2 (five red cards and two black cards) in their favor.
The second draw is 4 to 2 (or 2 to 1 if you like) because there will be four red cards and two black cards left.
The last draw is still in their favor by 3 to 2 (three reds and two blacks).
The game seems heavily in their favor, but YOU, are willing to offer them even money that they can’t do it!
The Seven Card Hustle
21
Take five red cards and two black cards from a pack.
Ask your friend to shuffle them and then, without looking at the faces, lay them out in a row.
Bet that they can’t turn over three red cards.
Explain how the bet is in their favor.
The game seems heavily in their favor, but YOU, are willing to offer them even money that they can’t do it!
[Lovell, 2006]
Even odds or
even money
means 1-to-1
odds.
The Seven Card Hustle: Sol
22
5 4 3
7 6 5
2
7
5
3 5!
7 3!
3
3!
2!
4!
7!
15 4 3
7 6 5
2
7
[Lovell, 2006]
The correct probability that they can do it is
Alternatively,
Do not worry too much about the math here.
Some of you may be able to calculate the probability
using knowledge from your high school years.
We will review all of this later.
23
Game 2:
Monty Hall Problem
Monty Hall Problem (MHP): Origin
24
Problem, paradox, illusion
Loosely based on the American television game show
Let’s Make a Deal. (Thai CH7 version: ประตูดวง.)
The name comes from the show’s original host, Monty
Hall.
One of the most interesting mathematical brain teasers of
recent times.
Monty Hall Problem: Math Version
25
Originally posed in a letter by Steve Selvin to the American
Statistician in 1975.
A well-known statement of the problem was published in
Marilyn vos Savant’s “Ask Marilyn” column in Parade
magazine in 1990: “Suppose you're on a game show, and
you're given the choice of three doors:
Behind one door is a car; behind the others,
goats. You pick a door, say No. 1, and the
host, who knows what's behind the doors,
opens another door, say No. 3, which has a
goat. He then says to you, "Do you want to
pick door No. 2?" Is it to your advantage to
switch your choice?”
Marilyn vos Savant
26
Vos Savant was listed in each edition of the Guinness Book
of World Records from 1986 to 1989 as having the “Highest
IQ.”
Since 1986 she has written “Ask Marilyn”
Sunday column in Parade magazine
Solve puzzles and answer questions from readers
[ http://www.marilynvossavant.com ]
MHP: Step 0
27
There are three closed doors.
They look identical.
MHP: Step 0
28
Behind one of the doors is the star prize - a car.
The car is initially equally likely to be behind each door.
Behind each of the other two doors is just a goat.
MHP: Step 1
29
Obviously we want to win the car, but do not
know which door conceals the car.
We are asked to choose a door.
That door remains closed for the time being.
“Pick one of
these doors”
MHP: Step 2
30
The host of the show (Monty Hall), who knows what is behind
the doors, now opens a door different from our initial choice.
He carefully picks the door that conceals a goat. We stipulate that if Monty has a choice of doors to open, then he chooses randomly from among his options.
MHP: Step 3
31
Monty now gives us the options of either
1. sticking with our original choice or
2. switching to the one other unopened door.
After making our decision, we win whatever is behind our door.
“Do you want
to switch
doors?”
Monty Hall Problem
32
Will you do better by sticking with your first choice, or
by switching to the other remaining door?
Make no difference?
Assuming that our goal is to maximize
our chances of winning the car, what
decision should we make?
33
Let’s play!
Interactive Monty Hall
34
http://montyhallgame.shawnolson.net/
http://www.shodor.org/interactivate/activities/SimpleMontyHall/
http://www.math.uah.edu/stat/applets/MontyHallGame.xhtml
http://scratch.mit.edu/projects/nadja/484178
http://www.math.ucsd.edu/~crypto/Monty/monty.html
Interactive Monty Hall
35
http://www.nytimes.com/2008/04/08/science/08monty.html
The New York Times’s Version
Need More Examples or Practice?
50
Textbook in the library: Schaum’s
outline of theory and problems of
probability, random variables, and
random processes / Hwei P. Hsu. Call
No. QA273.25 H78 1997
Free pdf textbook:
Introduction to Probability by
Grinstead and Snell http://www.dartmouth.edu/~chance
/teaching_aids/books_articles/proba
bility_book/book.html
Monty Hall Problem: a first revisit
51
Will you do better by sticking with your first choice, or
by switching to the other remaining door?
Make no difference?
Assuming that our goal is to maximize
our chances of winning the car, what
decision should we make?
Monty Hall Problem: vos Savant’s
Answer
52
“You double your chances of
winning by switching doors.”
Monty Hall Problem: Controversy
53
Approximately 10,000 readers,
including nearly 1,000 with PhDs
(many of them math professors),
wrote to the magazine
claiming the published solution was wrong.
“You blew it,” wrote a mathematician from George Mason
University.
From Dickinson State University came this: “I am in shock
that after being corrected by at least three mathematicians,
you still do not see your mistake.”
[Mlodinow, 2008, p 42-45]
Controversy (2)
54
From Georgetown: "How many irate mathematicians are
needed to change your mind?"
And someone from the U.S. Army Research Institute
remarked, "If all those Ph.D.s are wrong the country would
be in serious trouble."
When told of this, Paul Erdős, one of the leading
mathematicians of the 20th century, said, "That's
impossible."
Then, when presented with a formal mathematical proof of the
correct answer, he still didn't believe it and grew angry.
55
Let’s learn some concepts
so that we can analyze
interesting examples!
Dr. Prapun Suksompong [email protected]
1 Probability and You
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
2
Life is random
3
Life is random
In 2005, this fact
showed up all over the world…
4
Life is random
Applications of Probability Theory
5
The subject of probability can be traced back to the 17th century when it arose out of the study of gambling games.
The range of applications extends beyond games into business
decisions, insurance, law, medical tests, and the social sciences.
The stock market, “the largest casino in the world,” cannot do without it.
The telephone network, call centers, and airline companies with their randomly fluctuating loads could not have been economically designed without probability theory.
“The Perfect Thing”
6
“The Perfect Thing”
7
Perfect?!...
8
What about the shuffle function?
9
http://ipod.about.com/od/advanceditunesuse/a/itunes-random.htm
http://electronics.howstuffworks.com/ipod-shuffle2.htm
http://www.cnet.com.au/itunes-just-how-random-is-random-339274094.htm
How to Interpret a Probability
10
Many think that probabilities do not exist in real life.
Nevertheless, a given or a computed value of the probability
of some event A can be used in order to make conscious
decisions.
Long-run frequency interpretation.
If the probability of an event A in some actual physical
experiment is p, then we believe that if the experiment is
repeated independently over and over again, then in the long
run the event A will happen 100p% of the time.
Frequency Interpretation and LLN
11
These assumptions are motivated by the frequency
interpretation of probability.
If we repeat an experiment a large number of times then the
fraction of times the event A occurs will be close to P(A).
If we let N(A, n) be the number of times A occurs in the first n
trials then
Later on, this result will be a theorem called the law of
large numbers (LLN).
,
limn
N A nP A
n
USA Currency Coins
12
Penny = 1 cent
(Abraham Lincoln)
Nickel = 5 cents
(Thomas Jefferson)
Dime = 10 cents
(Franklin D. Roosevelt)
Quarter = 25 cents
(George Washington)
Coin Tossing: Relative Frequency
13
2 4 6 8 10
0
0.2
0.4
0.6
0.8
1
200 400 600 800 10000
0.2
0.4
0.6
0.8
1
20 40 60 80 1000
0.2
0.4
0.6
0.8
1
2 4 6 8 10
x 105
0
0.2
0.4
0.6
0.8
1
,N A n
n
1,2, ,10n 1,2, ,100n
1,2, ,1000n 61,2, ,10n
If a fair coin is
flipped a large
number of times,
the proportion of
heads will tend to
get closer to 1/2 as
the number of
tosses increases.
Intersting behavior
14
The difference between #H and #T will not be close to 0.
1 2 3 4 5 6 7 8 9 10
x 105
-500
0
500
1000
1500
2000
2500
Another trial
15
2 4 6 8 10
0
0.2
0.4
0.6
0.8
1
200 400 600 800 10000
0.2
0.4
0.6
0.8
1
20 40 60 80 1000
0.2
0.4
0.6
0.8
1
2 4 6 8 10
x 105
0
0.2
0.4
0.6
0.8
1
1 2 3 4 5 6 7 8 9 10
x 105
-600
-400
-200
0
200
400
600
800
1000
Another trial
16
2 4 6 8 10
0
0.2
0.4
0.6
0.8
1
200 400 600 800 10000
0.2
0.4
0.6
0.8
1
20 40 60 80 1000
0.2
0.4
0.6
0.8
1
2 4 6 8 10
x 105
0
0.2
0.4
0.6
0.8
1
1 2 3 4 5 6 7 8 9 10
x 105
-600
-500
-400
-300
-200
-100
0
100
200
300
Dr. Prapun Suksompong [email protected]
2 Review of Set Theory
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Venn diagram
2
Venn diagram: Examples
3
Partitions
4
Dr. Prapun Suksompong [email protected]
3 Classical Probability
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Real coins are biased
2
From a group of Stanford researchers
http://gajitz.com/up-in-the-air-coin-tosses-not-as-neutral-as-you-think/
http://www.codingthewheel.com/archives/the-coin-flip-a-fundamentally-unfair-proposition
http://www-stat.stanford.edu/~susan/papers/headswithJ.pdf
Example
3
In drawing a card from a deck, there are 52 equally likely
outcomes, 13 of which are diamonds. This leads to a
probability of 13/52 or 1/4.
The word “dice”
4
Historically, dice is the plural of die.
In modern standard English, dice is used as both the
singular and the plural.
Example of 19th Century bone dice
Dice Simulator
6
http://www.dicesimulator.com/
Support up to 6 dice and also has some background
information on dice and random numbers.
Two Dice
7
Two-Dice Statistics
8
Two-Dice Statistics
9
Two Dice
10
A pair of dice
Double six
Two dice: Simulation
11
[ http://www2.whidbey.net/ohmsmath/webwork/javascript/dice2rol.htm ]
Two dice
12
Assume that the two dice are fair and independent.
P[sum of the two dice = 5] = 4/36
Two dice
13
Assume that the two dice are fair and independent.
Dr. Prapun Suksompong [email protected]
4 Combinatorics
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Heads, Bodies and Legs flip-book
2
Heads, Bodies and Legs flip-book (2)
3
One Hundred Thousand Billion Poems
4
Cent mille milliards de poèmes
One Hundred Thousand Billion Poems
(2)
5
Example: Sock It Two Me
6
Jack is so busy that he's always throwing his socks into his top drawer without pairing them. One morning Jack oversleeps. In his haste to get ready for school, (and still a bit sleepy), he reaches into his drawer and pulls out 2 socks.
Jack knows that 4 blue socks, 3 green socks, and 2 tan socks are in his drawer.
1. What are Jack's chances that he pulls out 2 blue socks to match his blue slacks?
2. What are the chances that he pulls out a pair of matching socks?
[Greenes, 1977]
“Origin” of Probability Theory
7
Probability theory was originally inspired by gambling problems.
In 1654, Chevalier de Mere invented a gambling system which bet even money on case B on the previous slide.
When he began losing money, he asked his mathematician friend Blaise Pascal to analyze his gambling system.
Pascal discovered that the Chevalier's system would lose about 51 percent of the time.
Pascal became so interested in probability and together with another famous mathematician, Pierre de Fermat, they laid the foundation of probability theory.
best known for Fermat's Last Theorem
[http://www.youtube.com/watch?v=MrVD4q1m1
Vo]
Example: The Seven Card Hustle
8
Take five red cards and two black cards from a pack.
Ask your friend to shuffle them and then, without looking at the faces, lay them out in a row.
Bet that them can’t turn over three red cards.
The probability that they CAN do it is
[Lovell, 2006]
3
3
5 5
7
4 3
7 6 5
2
7
5
3 5!
7 3!
3
3!
2!
4! 1 25 4 3
7! 7 6 5 7
Finger-Smudge on Touch-Screen
Devices
9
Fingers’ oily smear on the
screen
Different apps gives different
finger-smudges.
Latent smudges may be usable
to infer recently and frequently
touched areas of the screen--a
form of information
leakage.
[http://www.ijsmblog.com/2011/02/ipad-finger-smudge-art.html]
Lockscreen PIN / Passcode
10
[http://lifehacker.com/5813533/why-you-should-repeat-one-digit-in-your-phones-4+digit-lockscreen-pin]
Smudge Attack
11
Touchscreen smudge may give away your password/passcode
Four distinct fingerprints reveals the four numbers used for
passcode lock.
[http://www.engadget.com/2010/08/16/shocker-touchscreen-smudge-may-give-away-your-android-password/2]
Suggestion: Repeat One Digit
12
Unknown numbers:
The number of 4-digit different passcodes = 104
Exactly four different numbers:
The number of 4-digit different passcodes = 4! = 24
Exactly three different numbers:
The number of 4-digit different passcodes =
2
3 4 36
Choose the
number that
will be
repeated
Choose the
locations of
the two non-
repeated
numbers.
News: Most Common Lockscreen PINs
13
Passcodes of users of Big Brother Camera Security iPhone
app
15% of all passcode sets were represented by only 10
different passcodes
[http://amitay.us/blog/files/most_common_iphone_passcodes.php (2011)]
out of 204,508 recorded passcodes
Even easier in Splinter Cell
14
Decipher the keypad's code by the heat left on the buttons.
Here's the keypad viewed with your thermal
goggles. (Numbers added for emphasis.)
Again, the stronger the signature, the more
recent the keypress.
The code is 1456.
Actual Research
15
University of California San Diego
The researchers have shown that codes can be easily discerned from quite a distance (at least seven metres away) and image-analysis software can automatically find the correct code in more than half of cases even one minute after the code has been entered.
This figure rose to more than eighty percent if the thermal camera was used immediately after the code was entered.
K. Mowery, S. Meiklejohn, and S. Savage. 2011. “Heat of the Moment:
Characterizing the Efficacy of Thermal-Camera Based Attacks”. Proceed-
ings of WOOT 2011.
http://cseweb.ucsd.edu/~kmowery/papers/thermal.pdf
http://wordpress.mrreid.org/2011/08/27/hacking-pin-pads-using-
thermal-vision/
The Birthday Problem (paradox)
16
How many people do you need to assemble before the
probability is greater than 1/2 that some two of them have
the same birthday (month and day)?
Birthdays consist of a month and a day with no year attached.
Ignore February 29 which only comes in leap years
Assume that every day is as likely as any other to be someone’s
birthday
In a group of r people, what is the probability that two or
more people have the same birthday?
Probability of birthday coincidence
17
Probability that there is at least two people who have the
same birthday in a group of r persons
terms
if
365
365 1365 3641 · · · , if 0 365
365 365 3 5
1,
6r
r
rr
Probability of birthday coincidence
18
The Birthday Problem (con’t)
19
With 88 people, the probability is greater than 1/2 of having
three people with the same birthday.
187 people gives a probability greater than1/2 of four people
having the same birthday
Birthday Coincidence: 2nd Version
20
How many people do you need to assemble before the
probability is greater than 1/2 that at least one of them have
the same birthday (month and day) as you?
In a group of r people, what is the probability that at least one
of them have the same birthday (month and day) as you?
Binomial Theorem
21
1 1 2 2( ) ( )x y x y
1 2 1 2 1 2 1 2x x x y y x y y
1 1 2 2 3 3( ) ( ) ( )x y x y x y
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3x x x x x y x y x x y y y x x y x y y y x y y y
( ) ( )x y x y
( ) ( ) ( )x y x y x y
2 323 33
xyy yxy yyxx yx
xy
y xyx y yy
y
xxx
x
xx
x y
2 22yyxx xxy yx xy y
1 2 3
1 2 3
x x x
y yy
x
y
Distinct Passcodes (revisit)
22
Unknown numbers:
The number of 4-digit different passcodes = 104
Exactly four different numbers:
The number of 4-digit different passcodes = 4! = 24
Exactly three different numbers:
The number of 4-digit different passcodes =
Exactly two different numbers:
The number of 4-digit different passcodes =
Exactly one number:
The number of 4-digit different passcodes = 1
Check:
104
⋅ 24 +103
⋅ 36 +102
⋅ 14 +101
⋅ 1 = 10,000
2
3 4 36
43
+42
+41
= 14
Ex: Poker Probability
23
[ http://en.wikipedia.org/wiki/Poker_probability ] Need more practice?
Success Runs (1/4)
24
Suppose that two people are separately asked to toss a fair
coin 120 times and take note of the results. Heads is noted as
a “one” and tails as a “zero”.
Results: Two lists of compiled zeros and ones:
[Tijms, 2007, p 192]
Success Runs (2/4)
25
Which list is more likely?
[Tijms, 2007, p 192]
Success Runs (3/4)
26
Fact: One of the two individuals has cheated and has
fabricated a list of numbers without having tossed the coin.
Which list is more likely be the fabricated list?
[Tijms, 2007, p 192]
Success Runs (4/4)
27
Fact: In 120 tosses of a fair coin, there is a very large probability that at some point during the tossing process, a sequence of five or more heads or five or more tails will naturally occur. The probability of this is approximately 0.9865.
In contrast to the second list, the first list shows no such sequence of five heads in a row or five tails in a row. In the first list, the longest sequence of either heads or tails consists of three in a row.
In 120 tosses of a fair coin, the probability of the longest sequence consisting of three or less in a row is equal to 0.000053 which is extremely small .
Thus, the first list is almost certainly a fake.
Most people tend to avoid noting long sequences of consecutive heads or tails. Truly random sequences do not share this human tendency!
[Tijms, 2007, p 192]
Fun Books…
28
Dr. Prapun Suksompong [email protected]
Events-Based Probability Theory
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Dr. Prapun Suksompong [email protected]
5 Foundation of Probability Theory
2
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Axioms of probability theory
3
Abstractly, a probability measure is a function that
assigns numbers to events, which satisfies the following
assumptions:
1. Nonnegativity: For any event A,
2. Unit normalization:
3. If A1, A2, . . . , is an infinite sequence of (pairwise)
disjoint events, then
11
( )i i
ii
P A P A
( ) 1P
0P A
Kolmogorov
4
Andrey Nikolaevich Kolmogorov
Soviet Russian mathematician
Advanced various scientific fields probability theory topology classical mechanics computational complexity.
1922: Constructed a Fourier series that diverges almost everywhere, gaining international recognition.
1933: Published the book, Foundations of the Theory of Probability, laying the modern axiomatic foundations of probability theory and establishing his reputation as the world's leading living expert in this field.
I learn probability theory from
5
Rick Durrett
Eugene Dynkin Philip Protter Gennady Samorodnitsky
Terrence Fine Xing Guo Toby Berger
Not too far from Kolmogorov
6
You can be
the 4th-generation
probability theorists
Dr. Prapun Suksompong [email protected]
Event-Based Properties
7
Probability and Random Processes ECS 315
Daniel Kahneman
8
Daniel Kahneman
Israeli-American psychologist
2002 Nobel laureate
In Economics
Hebrew University, Jerusalem, Israel.
Professor emeritus of psychology and public affairs at Princeton University's Woodrow Wilson School.
With Amos Tversky, Kahneman studied and clarified the kinds of misperceptions of randomness that fuel many of the common fallacies.
K&T: Q1
9
K&T presented this description to a group of 88 subjects and
asked them to rank the eight statements (shown on the next
slide) on a scale of 1 to 8 according to their probability, with
1 representing the most probable and 8 the least.
[Daniel Kahneman, Paul Slovic, and Amos Tversky, eds., Judgment under
Uncertainty: Heuristics and Biases (Cambridge: Cambridge University Press,
1982), pp. 90–98.]
Imagine a woman named Linda, 31 years old,
single, outspoken, and very bright. In college
she majored in philosophy. While a student she was
deeply concerned with discrimination and
social justice and participated in antinuclear
demonstrations.
[outspoken = given to expressing yourself freely or insistently]
K&T: Q1 - Results
10
Here are the results - from most to least probable
[feminist = of or relating to or advocating equal rights for women]
K&T: Q1 – Results (2)
11
At first glance there may appear to be nothing unusual in
these results: the description was in fact designed to be
representative of an active feminist and
unrepresentative of a bank teller or an insurance salesperson.
Most probable
Least likely
K&T: Q1 – Results (3)
12
Let’s focus on just three of the possibilities and their average
ranks.
This is the order in which 85 percent of the respondents
ranked the three possibilities:
If nothing about this looks strange, then K&T have fooled you
K&T: Q1 - Contradiction
13
The probability that two events will both
occur can never be greater than the
probability that each will occur individually!
K&T: Q2
14
K&T were not surprised by the result because they had given
their subjects a large number of possibilities, and the
connections among the three scenarios could easily have
gotten lost in the shuffle.
So they presented the description of Linda to another group,
but this time they presented only three possibilities:
Linda is active in the feminist movement.
Linda is a bank teller and is active in the feminist movement.
Linda is a bank teller.
K&T: Q2 - Results
15
To their surprise, 87 percent of the subjects in this trial also incorrectly ranked the probability that “Linda is a bank teller and is active in the feminist movement” higher than the probability that “Linda is a bank teller”.
If the details we are given fit our mental picture of something, then the more details in a scenario, the more real it seems and hence the more probable we consider it to be
even though any act of adding less-than-certain details to a conjecture makes the conjecture less probable.
Even highly trained doctors make this error when analyzing symptoms.
91 percent of the doctors fall prey to the same bias.
[Amos Tversky and Daniel Kahneman, “Extensional versus Intuitive Reasoning:
The Conjunction Fallacy in Probability Judgment,” Psychological Review
90, no. 4 (October 1983): 293–315.]
Related Topic
16
Page 34-37
Tversky and Shafir @
Princeton University
K&T: Q3
17
Which is greater:
the number of six-letter English words having “n” as their fifth letter or
the number of six-letter English words ending in “-ing”?
Most people choose the group of words ending in “ing”. Why? Because words ending in “-ing” are easier to think of than generic six letter words having “n” as their fifth letter.
The group of six-letter words having “n” as their fifth letter words includes all six-letter words ending in “-ing”.
Psychologists call this type of mistake the availability bias
In reconstructing the past, we give unwarranted importance to memories that are most vivid and hence most available for retrieval.
[Amos Tversky and Daniel Kahneman, “Availability: A Heuristic for Judging
Frequency and Probability,” Cognitive Psychology 5 (1973): 207–32.]
Misuse of probability in law
18
It is not uncommon for experts in DNA analysis to testify at a criminal trial that a DNA sample taken from a crime scene matches that taken from a suspect.
How certain are such matches?
When DNA evidence was first introduced, a number of experts testified that false positives are impossible in DNA testing.
Today DNA experts regularly testify that the odds of a random
person’s matching the crime sample are less than 1 in 1 million or 1 in 1 billion.
In Oklahoma a court sentenced a man named Timothy Durham to more than 3,100 years in prison even though eleven witnesses had placed him in another state at the time of the crime.
[Mlodinow, 2008, p 36-37]
Lab/Human Error
19
There is another statistic that is often not presented to the
jury, one having to do with the fact that labs make errors, for instance, in collecting or handling a sample, by accidentally mixing or swapping samples, or by misinterpreting or incorrectly reporting results.
Each of these errors is rare but not nearly as rare as a random match.
The Philadelphia City Crime Laboratory admitted that it had swapped the reference sample of the defendant and the victim in a rape case
A testing firm called Cellmark Diagnostics admitted a similar error.
[Mlodinow, 2008, p 36-37]
Timothy Durham’s case
20
It turned out that in the initial analysis the lab had failed to
completely separate the DNA of the rapist and that of the
victim in the fluid they tested, and the combination of the
victim’s and the rapist’s DNA produced a positive result
when compared with Durham’s.
A later retest turned up the error, and Durham was released
after spending nearly four years in prison.
[Mlodinow, 2008, p 36-37]
DNA-Match Error + Lab Error
21
Estimates of the error rate due to human causes vary, but
many experts put it at around 1 percent.
Most jurors assume that given the two types of error—the 1
in 1 billion accidental match and the 1 in 100 lab-error
match—the overall error rate must be somewhere in
between, say 1 in 500 million, which is still for most jurors
beyond a reasonable doubt.
[Mlodinow, 2008, p 36-37]
Wait!…
22
Even if the DNA match error was extremely accurate + Lab
error is very small,
there is also another probability concept that should be taken
into account.
More about this later.
Right now, back to notes for more properties of probability
measure.
Dr. Prapun Suksompong [email protected]
6.1 Conditional Probability
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Example
2
Roll a fair dice
Sneak peek:
3
Disease Testing
4
Suppose we have a diagnostic test for a particular disease
which is 99% accurate.
A person is picked at random and tested for the disease.
The test gives a positive result.
Q1: What is the probability that the person actually has the
disease?
Natural answer: 99% because the test gets it right 99% of the
times.
99% accurate test?
5
If you use this test on many persons with the disease, the
test will indicate correctly that those persons have disease
99% of the time.
False negative rate = 1% = 0.01
If you use this test on many persons without the disease, the
test will indicate correctly that those persons do not have
disease 99% of the time.
False positive rate = 1% = 0.01
Disease Testing
6
Suppose we have a diagnostic test for a particular disease
which is 99% accurate.
A person is picked at random and tested for the disease.
The test gives a positive result.
Q1: What is the probability that the person actually has the
disease?
Natural answer: 99% because the test gets it right 99% of the
times.
Q2: Can the answer be 1% or 2%?
Q3: Can the answer be 50%?
A1:
7
Q1: What is the probability that the person actually has the
disease?
The answer actually depends on how
common or how rare the disease is!
Why?
8
Let’s assume rare disease.
The disease affects about 1 person in 10,000.
Try an experiment with 106 people.
Approximately 100 people will have the disease.
What would the (99%-accurate) test say?
Test 106 people
Results of the test
9
100 people w/ disease
999,900 people w/o disease
99 of them will test positive
1 of them will test negative
989,901 of them will test negative
9,999 of them will test positive
approximately
Results of the test
10
100 people w/ disease
999,900 people w/o disease
99 of them will test positive
1 of them will test negative
989,901 of them will test negative
9,999 of them will test positive
Of those who test positive, only 991%
99 9,999
actually have the disease!
Bayes’ Theorem
11
Using the concept of conditional probability and Bayes’
Theorem, you can show that
the probability that a person will have the disease given
that the test is positive
is given by
where, in our example,
pD = 10-4
pTE = 1 – 0.99 = 0.01
(1 )
(1 ) (1 )
TE D
TE D TE D
p p
p p p p
Bayes’ Theorem
12
Using the concept of conditional probability and Bayes’
Theorem, you can show that
the probability that a person will have the disease given
that the test is positive
is given by
When different value of pD is assumed,
(1 )
(1 ) (1 )
TE D
TE D TE D
p p
p p p p
1
1
pD
In log scale…
13
10-6
10-5
10-4
10-3
10-2
10-1
100
10-5
10-4
10-3
10-2
10-1
100
dpD
Effect of pTE
14
pTE = 1 – 0.99 = 0.01
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
pTE = 1 – 0.9 = 0.1
pTE = 1 – 0.5 = 0.5
Wrap-up
15
Q1: What is the probability that the person actually has the
disease?
A1: The answer actually depends on how common or how
rare the disease is! (The answer depends on the value of d.)
Q2: Can the answer be 1% or 2%?
A2: Yes.
Q3: Can the answer be 50%?
A3: Yes.
Example: A Revisit
16
Roll a fair dice
Sneak peek:
Prosecutor’s fallacy
17
O. J. Simpson
At the time a well-known celebrity famous
both as a TV actor and as a retired
professional football star.
Defense lawyer: Alan Dershowitz
Renowned attorney and Harvard Law
School professor
[Mlodinow, 2008, p. 119-121],[Tijms, 1007, Ex 8.7]
Murder case
“one of the biggest media events of 1994–95”
“the most publicized criminal trial in American history”
The murder of Nicole
18
Nicole Brown was murdered at her home
in Los Angeles on the night of June 12,
1994.
So was her friend Ronald Goldman.
The prime suspect was her (ex-)
husband O.J. Simpson.
(They were divorced in 1992.)
Prosecutors’ argument
19
Prosecutors* spent the first ten days of the trial entering
evidence of Simpson’s history of physically abusing her
and claimed that this alone was a good reason to suspect him
of her murder.
As they put it,
“a slap is a prelude to homicide.”
Prosecutor* = a government official who conducts criminal prosecutions on behalf of the state
Counterargument
20
The defense attorneys argued that the prosecution* had spent two weeks trying to mislead the jury and that the evidence that O. J. had battered Nicole on previous occasions meant nothing.
Dershowitz’s reasoning:
4 million women are battered annually by husbands and boyfriends in the US.
In 1992, a total of 1,432, or 1 in 2,500, were killed by their (ex)husbands or boyfriends.
Therefore, few men who slap or beat their domestic partners go on to murder them.
True? Yes.
Convincing? Yes.
prosecution = the lawyers acting for the state to put the case against the defendant
batter = strike violently and repeatedly
The verdict:
21
Not guilty for the two murders!
The verdict was seen live on TV by more than half of the U.S.
population, making it one of the most watched events in
American TV history.
Another number…
22
It is important to make use of the crucial fact that Nicole Brown was murdered.
The relevant number is not the probability that a man who batters his wife will go on to kill her (1 in 2,500) but rather the probability that a battered wife who was murdered was murdered by her abuser.
According to the Uniform Crime Reports for the United States and Its Possessions in 1993, the probability Dershowitz (or the prosecution) should have reported was this one: of all the battered women murdered in the United States in 1993, some 90 percent were killed by their abuser.
That statistic was not mentioned at the trial.
A Simplified Diagram
23
Physically abused by
husband
Murdered by
husband
Murdered
Probability Comparison
24
Physically abused by
husband
Murdered by
husband
Murdered
Physically abused by
husband
Murdered by
husband
Murdered
1 in 2,500
90%
The Whole Truth …
25
Dershowitz may have felt justified in misleading the jury
because, in his words, “the courtroom oath—‘to tell the
truth, the whole truth and nothing but the truth’—is
applicable only to witnesses.
Defense attorneys, prosecutors, and judges don’t take this
oath . . . indeed, it is fair to say the American justice system is
built on a foundation of not telling the whole truth.”
Ex. Fair results from a biased coin
26
A biased coin can still be used for fair results by changing the game slightly.
John von Neumann gave the following procedure: 1. Toss the coin twice.
2. If the results match, start over, forgetting both results.
3. If the results differ, use the first result, forgetting the second.
Key idea: The probability of getting heads and then tails must be the same as the
probability of getting tails and then heads, Assumptions: the coin is not changing its bias between flips and the two flips are
independent.
By excluding the events of two heads and two tails by repeating the procedure, the coin flipper is left with the only two remaining outcomes having equivalent probability.
This procedure only works if the tosses are paired properly; if part of a pair is reused in another pair, the fairness may be ruined.
Dr. Prapun Suksompong [email protected]
6.2 Independence
1
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
Example: Club & Black
2
spades
clubs
hearts
diamonds
Example: Black & King
3
spades
clubs
hearts
diamonds
Sally Clark
4
[http://www.sallyclark.org.uk/]
[http://en.wikipedia.org/wiki/Sally_Clark]
[http://www.timesonline.co.uk/tol/comment/obituaries/article1533755.ece]
Sally Clark
5
Falsely accused of the murder of her two sons.
Clark's first son died suddenly within a few weeks of his birth in 1996.
After her second son died in a similar manner, she was arrested in 1998 and tried for the murder of both sons.
The case went to appeal, but the convictions and sentences were confirmed in 2000.
Released in 2003 by Court of Appeal
Wrongfully imprisoned for more than 3 years
Never fully recovered from the effects of this appalling miscarriage of justice.
Misuse of statistics in the courts
6
Her prosecution was controversial due to statistical
evidence
This evidence was presented by a
medical expert witness
Professor Sir Roy Meadow,
Meadow testified that the frequency of sudden infant death
syndrome (SIDS, or “cot death”) in families having some of
the characteristics of the defendant’s family is 1 in 8500.
He went on to square this figure to obtain a value of 1 in
73 million for the frequency of two cases of SIDS in such a
family.
2
8110
8500
Royal Statistical Society
7
“This approach is, in general, statistically invalid.”
“It would only be valid if SIDS cases arose independently
within families, an assumption that would need to be justified
empirically. “
“There are very strong a priori reasons for supposing that the
assumption will be false.”
“There may well be unknown genetic or environmental
factors that predispose families to SIDS, so that a second
case within the family becomes much more likely.”
[http://www.rss.org.uk]
Aftermath
8
Clark's release in January 2003 prompted the Attorney
General to order a review of hundreds of other cases.
Two other women convicted of murdering their children
had their convictions overturned and were released from
prison.
Trupti Patel, who was also accused of murdering her three
children, was acquitted in June 2003.
In each case, Roy Meadow had testified about the
unlikelihood of multiple cot deaths in a single family.
How Juries Are Fooled by Statistics
9
By Peter Donnelly
http://www.youtube.com/watch?v=kLmzxmRcUTo
http://www.stats.ox.ac.uk/people/academic_staff/peter_donnelly
@ 11:15-13:50 Disease Testing
@ 13:50-18:30 Sally Clark
Professor of Statistical
Science (Dept Statistics) at
University of Oxford
Prosecutor’s Fallacy
10
Aside from its invalidity, figures such as the 1 in 73 million are very easily misinterpreted.
Some press reports at the time stated that this was the chance that the deaths of Sally Clark's two children were accidental.
This (mis-)interpretation is a serious error of logic known as the Prosecutor's Fallacy.
The jury needs to weigh up two competing explanations for the babies' deaths: 1) SIDS or 2) murder.
Two deaths by SIDS or two murders are each quite unlikely, but one has apparently happened in this case.
What matters is the relative likelihood of the deaths under each explanation, not just how unlikely they are under one explanation (in this case SIDS, according to the evidence as presented).
Independence among three events
11
Can be checked via 23-3-1 = 4 conditions:
𝑃 𝐴 ∩ 𝐵 = 𝑃 𝐴 𝑃 𝐵𝑃 𝐴 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐶𝑃 𝐵 ∩ 𝐶 = 𝑃 𝐵 𝑃 𝐶
𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃 𝐶
Remarks: Pairwise independence among the three events is
defined by the first three conditions
Pairwise independence Independence
Independence among four events
12
Can be checked via 24-4-1 = 11 conditions:
𝑃 𝐴 ∩ 𝐵 = 𝑃 𝐴 𝑃 𝐵𝑃 𝐴 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐶𝑃 𝐴 ∩ 𝐷 = 𝑃 𝐴 𝑃 𝐷𝑃 𝐵 ∩ 𝐶 = 𝑃 𝐵 𝑃 𝐶𝑃 𝐵 ∩ 𝐷 = 𝑃 𝐵 𝑃 𝐷𝑃 𝐶 ∩ 𝐷 = 𝑃 𝐶 𝑃 𝐷
𝑃 𝐵 ∩ 𝐶 ∩ 𝐷 = 𝑃 𝐵 𝑃 𝐶 𝑃 𝐷𝑃 𝐴 ∩ 𝐶 ∩ 𝐷 = 𝑃 𝐴 𝑃 𝐶 𝑃 𝐷𝑃 𝐴 ∩ 𝐵 ∩ 𝐷 = 𝑃 𝐴 𝑃 𝐵 𝑃 𝐷𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 = 𝑃 𝐴 𝑃 𝐵 𝑃 𝐶
𝑃 𝐴 ∩ 𝐵 ∩ 𝐶 ∩ 𝐷 = 𝑃 𝐴 𝑃 𝐵 𝑃 𝐶 𝑃 𝐷
Pairwise independence requires
only these six conditions
Dr. Prapun Suksompong [email protected]
6.3 Bernoulli Trials
13
Probability and Random Processes ECS 315
Office Hours:
BKD 3601-7
Monday 14:40-16:00
Friday 14:00-16:00
0 5 10 15 20 25 30 35 40 45 500
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
n
n Bernoulli trials
16
Assume success probability = 1/n
#successes 1P
#successes 1P
#successes 0P #successes 2P
#successes 3P
10.3679
e
11 0.6321
e
10.1839
2e
Error Control Coding
14
Repetition Code at Tx: Repeat the bit n times.
Channel: Binary Symmetric Channel (BSC) with bit error probability p.
Majority Vote at Rx
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.50
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
n = 15
n = 5
n = 1
n = 25
p
P