Expected length of longest winning streak in a 162-game baseball season

[ACG]

Hi!

The Red Sox just finished an 11-game winning streak, which made me wonder. Do most teams experience winning streaks this long over the course of a season? What is the expected value of the length of the longest winning streak for a .500 team over the course of a season?

In general: if you flip a coin N times with a probability of getting a head of P, what is the expected length of the longest series of consecutive heads embedded in the results of the N flips?

ACG

 

[CRGreathouse]

To a good first approximation, there are (n - k + 1) independent chances with probability 1/2^k each to get a winning streak of length k in n games. (Of course they're not independent at all, but it's close enough.)

So the chance of getting an 11-game winning streak in 162 games is about 7%.

 

[ACG]

Correct. However, as you know they're far from independent. If game 11 is a loss (50% chance) then there's no way for the set of games ranging from 2 to 11 to be the start of an 11-game streak. That doesn't sound like it will work.

ACG

 

[CRGreathouse]

Correct. However, as you know they're far from independent. If game 11 is a loss (50% chance) then there's no way for the set of games ranging from 2 to 11 to be the start of an 11-game streak. That doesn't sound like it will work.

So in that case you overstate the probability of #2 to #11 by 1/2^11 (1/2^11 rather than 0). But if you know that the 11th game is a win, then you're understating the probability of #2 to #11 by 1/2^11 (1/2^11 rather than 1/2^10). So I don't think it's actually a problem!