How again does Jeff Kunkel figure into this?

huh huh huh huh huh huh huh huh... he said... he said...de-spite........huhuhuhuhuhuhuhuhuhhuhuhuh

Rich Rifkin is among the living. Praise Jeebus!

wrong thread...

I don't understand why he shows us the home/road splits of three players as a suggestion that there might be a "hangover" effect for Texas players. First of all, there are going to be all kinds of player to player fluctuations in home/road splits. That tells us nothing. Secondly, showing us home/road splits, even extreme ones, simply tells us something about the difference between the home park and the road ones. It gives us no suggestion whatsoever of a "hangover" effect. Thirdly, it is unlikely that there is a hangover affect for Texas players. Texas is a hitter's park because of the dimensions of the field and the hot, humid weather. That is unlikely to cause a hangover effect. The reason there is one for Rockies' players (we think) is that the ball (when pitched) does "funny things" in Denver as opposed to in other parks.

The only way to tell, and it is simple enough, if there is a hangover affect for any stadium is to compare its players' road stats from when they played with that team to when they played for another team. If you have a large enough sample, those two should be roughly the same. If they are not, and the difference is statisticially significant given the sample size, then you have evidence of a hangover effect.

As I said, you can question whether other extreme hitters' parks (or even pitchers' parks) have hangover effects like Coors (although as I said I doubt it unless they are at a high altitude as well - perhaps there is a small one with ARI), but a large home/road split for all the players, or for any players in particular, is not evidence of a hangover effect.

This article reminds me of one of MGL's important rules (I have not mentioned them in a long time):

Even if the whole world acknowledges the existence and an understanding of park effects, hitters in hitters parks will always be over-rated and pitchers under-rated, and vice versa in pitchers parks.

There are many correlaries to this rule. They are all of the same vein. For example, even if the whole world acknowledges and understands how basic projections are established (a weighted average of a player's career stats), players with recent banner years, high or low, will always be over- or under-rated, whichever the case may be.

I have to say that I wouldn't worry about the Ballpark too much. I mean, what can you do about it, exactly? You just have to find players who can hit better than their opponents, period. Mark Teixeira is capable of pretty good hitting no matter where (he hit 20 road home runs in 2004), and if he's a truly awesome hitter in Arlington, tant mieux. Mike Young hit .331 at home last year and .330 on the road; he earned his batting title. Dellucci hit quite a bit better on the road last year. Barajas hit twice as many HR on the road as at home in 2005. Soriano is wisely traded. And with Blalock, I would be much more concerned that he hit .236 in the second half of the season that about where he was hitting or wasn't.

Is a study of the Coors hangover effect available anywhere?

I've always been skeptical. Comparing road stats with the Rockies and for other teams has a few problems:

1) The Rockies play a large share of their road games in LA, SD, and SF - extreme pitchers parks.

2) The Rockies play zero road games in Coors field. Every other national league team plays 5-10% of their road games in Coors.

(I've never experienced the Coors hangover personally, I prefer imported beers.)

7: A link with links and stuff.

Come on, Anaheim Monkey, this is MGL! Don't you think I have controlled for that? Not only is it real, but it is substantial, and it is not really a "hangover" effect at all. I don't know if I have ever published the results, but they are easy enough to duplicate using retrosheet data. Just make sure you don't include road games in Coors Field (of course), and match up the stats park for park, or park adjust all the stats (park by park), if you want to control for the unequal schedule as well. The results will be imediately evident...

All the talk you hear and read about players' home/road splits, how they will do or not do in another park, how a particular park will or does affect a particular player, ad nauseam, is all nonsense and a waste of energy (as is 80% of all discourse for that matter). Take a player's stats, and park adjust his home stats (using L/R park factors if you want to be a tad more accurate), and that is a player's value (as compared to any other player), plain and simple, no matter where he plays...

Take a player's stats, and park adjust his home stats (using L/R park factors if you want to be a tad more accurate), and that is a player's value (as compared to any other player), plain and simple, no matter where he plays...

How do you account for a hitter that doesn't hit many fly balls in a home run park, or a pitcher that doesn't give up many fly balls in the same park?

Willie Harris won't necessarily hit significantly more home runs in Colorado than he would in San Francisco, because he rarely hits the ball hard into the outfield. Because the Park Factors for singles tend to be pretty similar across the board, I submit that his numbers will look pretty similar no matter where he's playing.

I understand what you're saying, that a player who hits 40 homers in Colorado isn't necessarily more valuable than one who hits 25 in Seattle (and if you reversed their location, they might put up each other's batting line), but I think you need to look closely at a player before applying park effect adjustments.

3 Points:
1) While I think you can break out stats in such a way that does or does not support a hangover effect (for Coors), it would be difficult to define a sample that accurately estimates the population effect (assuming there is one). I suspect that there would be too much variation, player-to-player. Say Bonds played there. He would take pretty much the same approach to each AB, regardless of where he played. A less disciplined player would shift more, home to road, but it could last 2 games for one player, a week for another.
2) That said, I think the so-called "hangover effect" is do in large part to the "ball doing funny" things as MGL says. My sense of the problem is that when you're hitting in Coors, you can swing at just about anything and you're rewarded for it. Plus, the ball simply isn't moving as much coming in. Unless you're a very disciplined hitter, it will be hard to adjust immediately to a more active ball where you have to be selective. Sort of like the difference between hitting against a machine or a Single A pitcher w/ a 95 MPH fastball v. a big-leaguer who can move the ball around.
3) MGL's point in #10 is a good one. While it's great to see the media more aware of park effects, there is this misperception that parks selectively help/hurt certain players, or more importantly, change them into someone else. A great example of this is the speculation of how much Soriano will be hurt moving from Texas to Washington. Sure, his stats will go down, but it doesn't mean he's a worse player, as has been some of the speculation elsewhere.

Sorry, 4th point. The reverse of #3 has been one of Dan O'Dowd's fatal flaws. While he's understood that marginal players will put up better stats playing half their games in Coors, I really think he believes this makes them better players.

MGL, how substantial is the Coors hangover, in terms of run production?

I know the standard home - road factor for Coors is somewhere around +30%.

Jerry, it's not that simple. Parks affect players' stats in strange, wonderous, and non-linear ways. It is not nearly so obvious which batters will be more or less affected by small parks, big parks, etc. (other than L/R). Coors field has all kinds of affects on batters (increases GDP's, many more ground balls go through the IF, many more short fly balls drop in in front of the OF'ers, etc.), other than fly balls going further. It is folly to think that you know that Willie Harris or any other batters would or would not be benefited more or less than any other player by Coors. Ditto for all other fields. For example, it is a myth that high HR pitchers, like Milton, are more adversely affected in high HR parks line Cynergy. Trust me, I have done a lot of research on how different parks and classes of parks affect different hitters. So have other people. I have basically come up with squat (a few things which I won't discuss). The biggest problem is the non-linearity of effects. A high HR park will not increase HR's for players and pitchers X% across the board. It might be 20% for low HR players and 10% for high HR players.

As far as the "hangover" effect, we know how it affects all players as a group and it is substantial. Whether or not it affects players differently is another story. I don't know. And if it does, whether we can find it out, I don't know either. Probably not, as there is not enough data to work with to be able to tease out the noise. Does the fact that there might be a different effect for different players minimize the importance of acknowldeging and adjusting for the affect as a whole? Definitely not.

Basically by not considering the substantial adverse affect that playng for the Rockies has on a player's road stats, we end up substantially undervaluing Rockie players after doing to standard park adjustments on their home stats. In fact, I don't know of anyone besides myself that not only does park adjustments on player stats, but does "hangover" adjustments as well. It is ironic, as non-sabermetric types will tend to overrate Rockie players and sabermetric types will tend to underrate them...

Ana, it is pretty big. While playing with the Rox (compared to playing for another team, adjusted for all road parks):

BB are reduced 8%
SO are increased 9%
Singles decreased 7% (IF singles are actually increased 4%, probably from more "mishits.")
Doubles decreased 5%
ROE decreased by 15%
Ironically, HR's increased 7% (probably becauxse players are HR "happy")

That is about an 8 run decrease per season, or around a 10% decrease.

The effect is similar for pitchers (in revrese of course)...

Trying to context-neutralize Rockie player stats is actually a nightmare. Don't forget that Coors Field also affects the home players differently from the visiting players (which is inextricable from Coors Field unique HFA, which is also inextricable from the hangover affect). Trying to come up with a park factor for home players, one for road players, and then adjust for the hangover affect is a nightmare. Basically don't trust all that much (withing reason of course) any park neutral stats for Rockie players. I can't even begin to explain how I do adjustments for Rockie players...