THE BOOK--Playing The Percentages In Baseball

If I understand correctly what Josh did he used a Monte Carlo simulation to model the probability of outcomes using a range of differen OPS eg, .600, .700, .800 etc. and league average base running for all base out situations.

I believe that Josh then superimposes OPS by count data that B-R publishes to drive to WPA by count.

Any chance someone wants to design a search function for this, like the one at walkoffbalk.com?

Not sure why you are assuming the “range of OPS”.  It seems to me that he simply took league average of everything, and ran a Markov on that.

Just recounting Josh’s own words on the matter ...

http://www.baseball.bornbybits.com/blog/2007/10/intentional-walk-tool.html

In that case, he had to do that, since he’s trying to find the break-even point for the IBB.

In the current case, there’s really not much point in using anything other than average all-around.

Yes Tom you are exactly right.  I took my MC set to league average when I made the table.  I wouldn’t really call it a Markov chain though as I only needed to run one batter because the next batter starts with a 0-0 count and I just use the WE from Dave and Jon’s chart.  Lastly, I just want to clarify league average.  I used the 2007 OPS by count from B-R so if the count was 2-1 I used the league average OPS from the “after a 2-1 count” row for the OPS.  As for the WPA vs. WE thing I am planning on fixing all those errors this weekend when I update the site.

Oh and a search function would be a great idea James.  Sadly, I am swamped right now and there are about three things I need to add before that.  Maybe in a week or two I will make a search tool that will look like the IBB tool.