« Paper Trail | Main | Cats Vs. Dogs »
January 01, 2005
Moneyball Part II
I was thinking about the calculations of the value of on base percentages the other day and I think I discovered a slight flaw in the measurement of the value of on base percentage. In Moneyball and many other theories of valuing baseball statistics a walk is considered equal to a single. They both get the batter to first base. But in many cases that is not all they do. If there is a runner on base a hit often advances them farther than a walk. In some cases (when there is no one on first and men on second and/or third or when there is a man on third but no one on second) a walk will not advance these runners at all. In these same cases a hit will most often advance these runners. The other benefit of a hit is that it will sometimes advance a runner more than one base, while a walk at most advances a base runner one base.
How does one measure this? I am not aware of any statistic that measures a hitters ability to advance runners. RBI kind of measures it, but a hit can still be beneficial that moves a runner from first to third and this is not an RBI. The biggest problem is that walks are only determined by the batter and pitcher, while base advancement is determined by hit placement, runner speed, and fielding. However it is determined the runner advancement statistic would need to be figured in somehow along with on base percentage and slugging percentage to get the true worth of a batter.
Posted by Pete at January 1, 2005 02:38 PM
Trackback Pings
TrackBack URL for this entry:
http://www.petetheelder.com/mt-tb.cgi/257
Comments
I can't recall who it was but last year a blogger who was huge on stats came up with a stat that measured both the total bases a player produced plus the number of bases the hitter advanced the baserunners. He found that in about 90% of the games - the team with the higher number for this stat won the game.
The stat also took into account bases advanced on sacrifices.
Posted by: chris at January 3, 2005 09:15 AM
Slugging percentage pretty effectively measures how far a player advances other baserunners ahead of him. Combining both components to scoring runs (getting on base, and moving runners around) is why many folks add OBP and slugging to get a quick measure of a player's offensive worth (OPS).
The biggest problem with RBI is how context dependent it is; you can only get RBI if there are runners on base ahead of you, and that circumstance is entirely dependent on your teammates. If you look at slugging, then you have a measure that more effectively isolates an individual's run prodcuing impact.
Posted by: Lennox at January 6, 2005 05:40 PM
Lennox, I thought that slugging percentage might be a very close approximation and in Moneyball at one point the author mentions a formula (I think it is Beene's idea and what follows is from memory) of six times OBP plus SLG since by the math he did slugging was valuable in measuring run production, but not nearly as valuable as OBP. One of the other people in the book thought it should be more like four times OBP plus SLG. I have not looked at the math enough to know the right formula.
Part of the problem with SLG as a measure of base advancement is that even with a high slugging percentage runner advancment partly depends on the skill of the other base runners and the whole point of the original exercise was to see how many runs the individual batter producers independent of the rest of the team and circumstances. This makes SLG a good stat to measure individual performance because SLG varies so much by player, it effects run production, and (unlike RBI) it is soley a measure of the batter, not the batters circumstance.
Posted by: Pete at January 6, 2005 11:51 PM