A few days ago in the thread about the CBS story on the brewers I commented that the author must not be watching a lot of brewer games if he thought our baserunning had been solid. Jacob repsonded to my negativity saying that overall the successful first to thirds and such cancel out some of the blunders. He ended with "There's been some horrible base running... but there has been just as much, if not more excellent base running. As a spectator, i'll take it, it is way more exciting. :)".
So that got me wondering if his positive view was right or my negative view. I looked around the internet to try to find the answer but none of the stats sites had too much on the topic. Speed score was the closest thing I could find and that didn't quite answer what I wanted.
So today I decided to write a quick C++ program to figure out the answer. It turned out to be more complicated than I thought but I have a partial answer.
What I did was take the game logs from retrosheet and parse them looking for baserunning successes and failures. Every time a play was made I calcuated the effect that play had on the expected runs scored using BP's Run expectancy matrix.
So for example, runner on first with one outs. The batter hits a double and the runner tries to score from first. If he makes it then the expectancy of runs from that inning is 1 + 0.684 (the one run that he scored plus the expectancy from a man on second one out situation). If the runner was tagged out at home then the expectance is 0.333 (two outs man on second situation).
Anyway, I tallied all the times a brewer player did the following: stole a base, got caught stealing, got picked off, went first to third on a single (or made an out at 3rd), went second to home on a single (or out at home), or went first to home on a double (or out at home). I applied the change in expectancy for each instance and then tallied them up.
The final total is 15.7729 runs. That represents the increase in number of expectant runs from the world where the brewer's players never did any of the previous actions. This is actually a pretty remarkable number since their total on stolen bases, caught stealings, and pick offs was -4.45.
So it is good that this is a positive number. The baserunning has, in general, helped them score more runs. The problem with this is there is nothing to compare it to since I only downloaded the brewer box scores. I could pretty easily run a number for brewer opponents but to run numbers for the whole league would take some more time. If people are interested in that though I could probably whip it up.
Anyway, I have to run but I will check in tonight. And because I used some retrosheet data I am adding:
The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at "www.retrosheet.org".
EDIT: Ok I hacked some terrible code to get these numbers. I am going to have to go back and rewrite if this stat becomes interesting if I want to rerun over a larger set. In any case here are the numbers for the nation league:
team New Fangled Stat
Ick. Not so good for the crew. Only Arizona has gotten less value on the basepaths than the brewers. Still no sac flies in here. I am having trouble getting the 2nd to 3rd on the sac fly working. If/when I get this running for individual players it will be interesting to see how much guys like Reyes are skewing the data. The average is the simple average of the numbers. S.D. is the standard deviation for people who like to see things like that.