Trip, Trip, Triple Your Fun

Jason Michael Barker

Sunday I wrote a piece detailing various correlations between team runs scored and other batting statistics -- walks, hits and home runs. To recap, home runs are a slightly better predictor of runs scored than walks, and hits are better than either of them. To close out the column I wrote:

Finally, in case you're interested, I ran the data for correlation between doubles, triples, and runs scored. Doubles scored .61, slightly less than walks. Triples checked in at -.24, a very weak, however negative, relationship. Could it be that the more triples you hit, the less runs you score? I can't imagine why that would be the case, and the relationship is a very weak one at that. It's probably safe to chalk it up to a one-year aberration, although on the whole it would appear triples have very little affect on scoring.

I didn't really think much more of it until I got an e-mail from Tom Maccarone:

I read your recent article about correlations between various offensive stats and runs, and you said you couldn't think of a reason why triples would be anti-correlated with run scoring. I think I can. Both the parks and types of players who get a lot of triples tend to have fewer homers than average. Example -- NL players last year with at least 8 triples were Dellucci, Larkin, Wilton Guerrero, Neifi Perez, DeShields, Karim Garcia, and Andruw Jones. Only Andruw hit at least 20 homers. AL list was topped by two Royals (Kauffman has always been a great triples park and a bad home run park), although it did include a few power hitters in Garciaparra and O'Leary. I bet the anti-correlation between triples and homers or triple and doubles would be even stronger than the one between triples and runs, and since the actual number of triples is barely significant to run scoring, the fact that triples hitters don't hit homers may dominate the correlation between triples and runs.

Tom

Good points all, Tom. One thing I'd like to point out, which perhaps wasn't clear enough in the original article, is that correlation is not the same thing as causation. As a psychology major, I have this fact repeated to me on nearly a daily basis. In other words, just because two things occur together does not mean that one is causing the other. It may be that A is causing B, B is causing A, or a third factor is influencing both.

This is what Tom was getting at in his email. It's not that triples themselves actually decrease run scoring -- that wouldn't make any sense. However, players and teams who tend to hit a lot of triples don't tend to hit as many home runs, which is detrimental to run scoring.

To take his points in order, park effects are certainly part of what's going on here. Fortunately, these numbers are available in the back of the STATS Major League Handbook (a great resource, by the way). For the purposes of ballpark effects, STATS uses a scale based on 100 being neutral for a particular statistic. If a park favors home runs, for example, then the number will be greater than 100. Last season Coors Field in Colorado received a 138 for home runs; over the last three years its scored 140.

A couple of parks are of note for the purposes of this discussion. The first was Edison International Field, home of the Anaheim Angels. "The Ed" rated a 145 for triples last season, but a 92 for home runs and 95 for runs. More triples, less homers, less scoring. Same goes for Turner Field in Atlanta -- 136 for triples, 94 for homers, and 98 for runs scored. Ditto New York's Shea stadium.

Three year (1996 to 1998) numbers reveal similar patterns for each of the following "triples friendly" parks: Comiskey (Chicago), Cinergy (Cincinnati), Pro Player (Florida), Olympic (Montreal), Three Rivers (Pittsburgh), and the SkyDome (Toronto).

Not all parks which favor triples decrease home runs, however, and those which favor both (predictably) also favor increased run scoring. Kauffman Stadium, which ranked a robust 173 in favoring triples over the past three seasons, slightly favored run scoring over that period (rating a 103), probably thanks to a 107 rating for home runs.

However, it appears that the pitcher-friendly parks which favor triples are just enough to make a difference in the correlation between triples and runs scored. In fact, one of the correlations Tom suggested does exist, that being between home runs and triples. Sure enough, triples and home runs scored a -.26 correlation last season. It's not much, but something.

Finally, I decided to look at the this relationship over a longer time period than just last season. In looking back three years (to include the last three full, 162-game, seasons), the correlation between triples and runs scored is even weaker than it was in 1998, at -.21.

Like I said before -- as a team statistic, triples have very little effect on run scoring. Just don't tell Willie Wilson, Tim Raines, or Lance Johnson.

about the author

Jason Michael Barker once hit a triple in an intramural softball game, but he tried to stretch it into a home run when his team waved him around third. He was thrown out by 20 feet.

Custom Search