Professional Documents
Culture Documents
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Outline
Tossing coins
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Outline
Tossing coins
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Tossing coins
Suppose we toss a million fair coins. How many heads will we get? About half a million, yes, but how close to that? Will we be o by 10 or 1000 or 100,000? How can we describe the error? Lets try this out.
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Tossing coins
Toss n coins. What is probability to see k heads? n Answer: 2k k . Lets plot this for a few values of n. Seems to look like its converging to a curve. If we replace fair coin with p coin, whats probability to see k heads. n Answer: p k (1 p )nk k . Lets plot this for p = 2/3 and some values of n. What does limit shape seem to be?
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Outline
Tossing coins
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Say X is a (standard) normal random variable if 2 e x /2 . fX (x ) = f (x ) = 1 2 Clearly f is always non-negative for real values of x , but how do we show that f (x )dx = 1? Looks kind of tricky. 2 Happens to be a nice trick. Write I = e x /2 dx . Then try to compute I 2 as a two dimensional integral. That is, write 2 2 2 x 2 /2 y 2 /2 I = e dx e dy = e x /2 dxe y /2 dy .
re r
2 /2
dr = 2 e r
2 /2
,
0
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Say X is a (standard) normal random variable if 2 e x /2 . f (x ) = 1 2 Question: what are mean and variance of X ? E [X ] = xf (x )dx . Can see by symmetry that this zero. Or can compute directly: 1 1 2 2 e x /2 xdx = e x /2 E [X ] = = 0. 2 2 How would we compute Var[X ] = f (x )x 2 dx =
2 1 e x /2 x 2 dx ? 2 2
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Again, X is a (standard) normal random variable if 2 e x /2 . f (x ) = 1 2 What about Y = X + ? Can we stretch out and translate the normal distribution (as we did last lecture for the uniform distribution)? Say Y is normal with parameters and 2 if 2 2 1 e (x ) /2 . f (x ) = 2 What are the mean and variance of Y ? E [Y ] = E [X ] + = and Var[Y ] = 2 Var[X ] = 2 .
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Again, X is a standard normal random variable if 2 f (x ) = 1 e x /2 . 2 What is the cumulative distribution function? a 2 Write this as FX (a) = P {X a} = 1 e x /2 dx . 2 How can we compute this integral explicitly? Cant. Lets just give it a name. Write 2 a 1 (a) = 2 e x /2 dx . Values: (3) .0013, (2) .023 and (1) .159. Rough rule of thumb: two thirds of time within one SD of mean, 95 percent of time within 2 SDs of mean.
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Outline
Tossing coins
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Let Sn be number of heads in n tosses of a p coin. Whats the standard deviation of Sn ? Answer: npq (where q = 1 p ).
n np The special quantity S npq describes the number of standard deviations that Sn is above or below its mean. Whats the mean and variance of this special quantity? Is it roughly normal? DeMoivre-Laplace limit theorem (special case of central limit theorem):
18.440 Lecture 19
Tossing coins Normal random variables Special case of central limit theorem
Problems
Toss a million fair coins. Approximate the probability that I get more than 501, 000 heads. Answer: well, npq = 106 .5 .5 = 500. So were asking for probability to be over two SDs above mean. This is approximately 1 (2) = (2) .159. Roll 60000 dice. Expect to see 10000 sixes. Whats the probability to see more than 9800? 5 Here npq = 60000 1 6 6 91.28. And 200/91.28 2.19. Answer is about 1 (2.19).
18.440 Lecture 19
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.