Boyd's World-> Breadcrumbs Back to Omaha-> Updated Park Factors

About the author, Boyd Nation

The last couple of years my kids have spent a good bit of time on a PC game called Backyard Baseball. It's a really cool game, all in all. Done with big oversized graphics, you control a team of kids playing a sandlot game. Included are kid-size representations of some major leaguers -- the lanky-but-still-ten-year-old looking Randy Johnson is hilarious -- but most of the kids are just average kids and a few of them are fairly cool; there's at least one kid in a wheelchair, for example, who still manages to hit for a decent average most games due to his speed. It's intended more as a gameplay game than as a simulation, although you can run whole seasons with teams that you create.

The part that relates to my topic for today, though (ah, there it is, and only in the second paragraph [which is considerably quicker than, for example, Melville introducing the whale in Chapter 123 {but less self-referential, I suppose}]) is that there are different fields that you can play on, and some of them are marvelously creative and really affect the play. There's a typical abandoned lot, bad hops and all. There's a beach field, where stolen bases are obviously non-existent. There's a blacktop field. There's the typical back yard with the short left field fence leading to mean Mr. McGarrity's yard where you never get your ball back, so those are outs instead of home runs. All told, you have to construct your team to be able to play on a real field, instead of on some theoretical construct, just like real college teams have to do when they, for example, build their stadium on top of a mountain.

If you've been following this column from the beginning (c'mon, at least pretend you've read the archives), you've gotten to follow the development of the best park factors I can manage, from the initial shrugging "there's just no way to do that with an unbalanced schedule" to an in-conference factor to a factor-against-home-and-home opponents to the current version that does a network analysis to try to put every park into a national perspective. My thinking for now is that this last one is the most accurate park factor we're going to get for the colleges, even if it's not perfect, but I'd love to be proven wrong.

For this year, I've re-run the park factors using 2000-2003 data, and I've done something else in the way of verification work for the method I'm using. I took the 1999-2002 numbers and checked to see how much they differed from the 2000-2003 numbers. It turns out that the correlation between the two sets of numbers is .79. What that means is that the method works, in general (a perfect method might get a correlation of .90 or so due to actual changes in parks and weather from year to year), but that it's not perfectly precise, which fits pretty well into what I've said all along. There just aren't enough data points to get a perfect measure, but whether New Mexico's PF is actually 211 or 185 (and none of the others change nearly that much), it's still the highest-offense park in the nation.

The whole list is over in The Filing Cabinet, but here are the top and bottom 10:

PF  TPF Team

189 162 New Mexico
156 137 New Mexico State
155 145 Air Force
153 130 Nevada
145 112 Washington State
138 119 Texas Southern
137 130 Hawaii
136 114 Western Carolina
136 109 New York Tech
135 113 Ball State

 59  87 IUPU-Fort Wayne
 62  85 Savannah State
 72  87 UC Irvine
 73  85 Arkansas State
 74  80 Mississippi
 75  82 Tulane
 75  90 Northeastern
 76  80 South Carolina
 76  82 Clemson
 76  87 Boston College

The second column, TPF for Team Park Factor, is another measure I've added this year and hope to make good use of. In the majors, at least historically, it's fairly easy to adjust statistics if you have an accurate park factor. You just assume that the team is playing half their games at home and half on the road, assume the road averages out to a neutral park, and split the difference. (Note that with the current unbalanced schedule, this may not be true, but it's the general way it's been done historically.) This doesn't work at all for college, since few teams play half their games at home, and the road parks they play in seldom average out to a neutral park. Therefore, the TPF number is a weighted average of all the parks they play in, their own included, which should be useful for stats adjustment. In other words, not only is New Mexico's park small and high enough to boost scoring, most of their other games are in places less offense-friendly but still well above average, so their offensive stats must be discounted by more than just allowing for their home park would require.

If you're interested in reprinting this or any other Boyd's World material for your publication or Web site, please read the reprint policy and contact me

Boyd's World-> Breadcrumbs Back to Omaha-> Updated Park Factors

About the author, Boyd Nation