World Series Baseball

June 19, 2004 | Baseball, My Projects | By: Mark VandeWettering

I was mulling over the mathematics of the World Series in one of my more bored moments, and decided to write some simple Python code to test an idea. The question underlying these graphs is essentially this: given that the probability of Team #1 winning any individual game is a probability p, what is the odds that the team will win the Series? And what is the average duration of the Series?

A few minutes of Python hacking gives the results in nice graphical form (courtesy of gdchart):

Probability of Winning
Length of Series

For some reason, the title along the Y axis of the second graph is screwy, I haven’t been able to figure out why. I’ll play with it later.

There are few surprises here: the probability graph is a sigmoid which begins at 0, goes through 0.5 at x = 0.5, and ramps off at 1.0. Similarly, the duration peaks when teams are evenly matched, somewhere around 5.8 games.

What was this worth? Not much. I just liked to make the graphs.