from Amazon
Tuesday, July 22, 2008
ERA+
This is really SABR 101 stuff, so most of you will not care at all about this, but in case some newbies are around, I respond to Geoff Baker who says:
Right now, his ERA+ is down to 84. That’s well off his 100 from last year. But part of that is a function of the league ERA as a whole, which is way lower than in 2007.
To which I replied:
Geoff, the idea behind ERA+ is that if you have a 4.50 ERA and the league is 4.50, and then next year you have a 4.00 ERA and the league is 4.00, you have not improved. Whatever it is that causes the league to go from 4.50 to 4.00 in one year (say larger strike zone, or deader ball or whatnot) affects everyone. Therefore, you have to treat ERA+ as a “normalized” ERA, similar to treating your money in “real dollars” as inflation-adjusted.
Your point is only valid if you believe that there’s been an influx of good pitchers who came in all of a sudden in one year. The presumption is that the overall quality of pitchers and batters does not change much (at least is not noticeable) in any two consecutive years. Certainly over a 10 or 20 year period, it may change, but over 2 years, no one will be able to tell.
Insofar as your statement in the article is made, it doesn’t hold.
He replied:
I appreciate what you are saying. But if American League ERA has fluctuated between 4.35 and 4.63 for the past five years, is it not unreasonable to expect this year’s 4.14 ERA to “regress to the mean” once a full sample size of one season is played? I mean, I haven’t checked this out, but I’d imagine the effect of interleague play and throwing to weaker NL lineups has taken the AL ERA to better heights, which will regress once more games against AL squads offset that? AL ERA hasn’t been this low, at least not by season’s end since 1993. That’s why I find it hard to believe this year’s ERA will remain so low come the end of September.
Yes, you could make the same argument that Washburn’s ERA will rise accordingly as well. But then again, not necessarily to the same terrible level he was at in May, when squads were obliterating him. I don’t expect his ERA to regress at quite the rate the league’s will. Even if he regresses slightly from where he’s been the past eight starts as we move forward and he faces more AL squads, his overall ERA could still improve, just because of how poorly he did early on. At that rate, his ERA+ would also improve, so long as AL ERA regressed to its usual standard of the past 15 years. Am I wrong?
And me:
Geoff, the expectation is that what has happened to the league has also happened to Jarrod. You are suggesting that it’s possible that we’ve got a sampling issue, in that what has happened to the league is not necessarily real, and therefore, we shouldn’t, at this point, be too crazy about comparing Jarrod to the league. It has some validity to it, if this were early in the season. At this point, I don’t think it’s true. However, it is a fair point, and one that could be explored further.
He definitely makes a fair point. In fact, I think that the reason the AL run scoring (and ERA of course) is so low this year is a combination of worse hitters (older and less juiced), better pitchers (influx of better pitching and/or bad ones leaving), and luck. The only thing that would affect Washburn or any other individual pitcher is worse hitters, so what he says is more than reasonable.
The only thing that would make you want to compare a pitcher to the league in any given year even if the entire league changes is if the ball, the umpire/strike zone, weather, or batters change.
And just like we always assume that when a batter hits above the mean, he necessarily got a little lucky (and he is good of course), regardless of the sample size, we have to assume that since AL ERA’s have historically and most recently been higher, that this year HAS to be at least partly a statistical aberration no matter where in the season we are.
So his logic makes a lot of sense. You definitely want to be careful about using ERA+ or any league normalized stat for any one player when the league mean is different than it has been in recent years, and we don’t know why. In fact, I would rather use the last 3 year’s league ERA to do a ERA+ than just the current year, I think.