Miscellanity on AstroTurf and my recent obsession with DIPS
July 22, 2007 3 Comments
A few minor musings that I’ve been tinkering with lately. For years I’ve heard that AstroTurf is a “faster” surface, so a ball hit on the ground rolls faster, and presumably has more of a chance to go through for a base hit. I guess rolling a ball across fuzzy cement is easier than rolling it across actual grass, but does that really translate into more grounders that get through the infield? To check, I isolated all ground balls hit from 2003-2006 (Retrosheet, I love you.) If the ball was fielded by an infielder (even if it went for an infield hit), I coded it as not going through the infield. The fielder got there. It’s not the grass or turf’s fault that he couldn’t make the throw. If it was fielded by an outfielder, the ball clearly went through the infield.
I calculated a park effect of sorts for this particular event. I compared how teams’ defenses did on the road at cutting balls off vs. those numbers at home, and then how their offenses did at punching the ball through at home and on the road. To get my park effect, I took home % / road % for offense and defense separately and averaged the two results. Like regular park effects, the results were actually pretty variable. ICC was around .18, meaning that there was very little year-to-year consistency. Not only that, but of the three stadia that still employ artificial turf (Tropicana Field, SkyDome… er, the Rogers Center, and the Metrodome) were generally in the middle of the pack. Veterans Stadium also used turf in 2003, as did Stade Olympique in Montreal in 2003 and 2004, although the Expos were too busy playing in Puerto Rico half the time that year.
Artificial turf may actually make for a faster rolling ball, but teams probably solve that by playing back a bit and the strategy doesn’t seem to have any ill effects for balls rocketing through the infield on the ground. In 1988, Bill James wrote, “Our idea of what makes a team good on artificial turf is not supported by any research.” Seems like finding a bunch of ground ball hitting speed demons to take advantage of some sort of property of the turf isn’t such a good idea. That property isn’t there.
Also, I’ve been a bit obsessed with DIPS lately and a thought occurred to me. I’m now fairly convinced that there is a small amount of skill (small, but present) in a pitcher influencing what happens to the ball when it comes off the bat. Perhaps the pitcher can place the ball in a particular spot to induce a ground ball that will go toward the third baseman. Again, it might not be a great amount of influence, but maybe we can figure out some of what drives that skill. What if we controlled for… well, control? A pitcher who has good control doesn’t walk many batters. I regressed BABIP on walk rate and saved the residuals. (My data set was pitchers from 2003-2006 with a minimum of 50 IP.) The ICC for the residuals is .209, which is a touch higher than I’d found previously. But then again, I looked at the ICC for BABIP proper in this sample and it was .216. So, controlling walk rate actually made things worse. I tried with strikeout rate and that didn’t work either.
Back to the drawing board, I suppose.