THE BOOK--Playing The Percentages In Baseball

Monday, September 04, 2006

Quirks in the Win Expectancy Tables

A reader found a problem with the WE tables here, http://www.sportsmogul.com/vbulletin2/showthread.php?t=119234 .  I responded:

When I create the WE tables, I use a basic run frequency table, like this one:
http://www.tangotiger.net/RE9902score.html

I know how often to expect a certain number of runs to end of inning, from each of the 24 base/out states. If we focus on the 1b/3b, 0 out lines and the 2b/3b 0 out lines, we see the chance of a scoreless inning is only 12% with the guy on 1B, but 14% with the guy on 2b. However, the chance of a multi-run inning is way higher with the guy on 2B.

So, in this case, on average, small-ball is played much more often with the guy on 1B and 3b, than 2b and 3b. Again, on average.

However, when I run my numbers, I assume this applies to every inning. Clearly, this can’t be true in the 9th inning of a tie game, where teams play completely differently.

Therefore, the “human element” is missing from my basic WE charts that I have published. It would be definitely doable to incorporate, so you have something that is more logical.