Regression fallacy

In a recent explosion of offense, the Angels have cruised to two lopsided victories over the San Diego Padres while scoring 20 runs and hitting 7 balls over the fence (4 of them by Torii Hunter). And this comes on the heels of a humiliating 11-1 drubbing at the hands of the Tampa Bay Rays that finally deflated whatever air was left in the party balloons at the end of a disappointing 4-5 road trip. This just goes to show you, there is no such thing as momentum in baseball. And thank God for that.

We've already heard a lot about Scioscia's "chewing out" session after Thursday's game. If the Angels continue to win, we'll probably hear a lot more about how it "turned things around" and "inspired" the team to step up and "elevate" their game. Mark Gubicza said as much in a mindless rant during the 5th inning of Thursday's game. I'm paraphrasing, but he accused the entire offense (with the exception of Torii Hunter who had an RBI hit in the inning) of "not taking responsibility" for their play and "stepping up" to get that clutch hit. What the hell does that even mean? If baseball players had the ability to do that by sheer force of will alone, don't you think they would do it, like, all the time, maybe?

I'm not going to be the pushy stat geek and bring up the R² value of the most recent study on clutch hitting (or the nonexistence thereof) or compute the kurtosis of team batting-average. You can get that lots of other places. But there are a few simple facts about statistics that border on obviousness. For example, you can't win every game is an expression of the statistical fact that, well, you can't win every game, just like you can't flip 500 coins and get 500 heads. The probability isn't actually zero, but it's so close to zero that it might as well be zero. Similarly, you'll often hear commentators (ahem, Rex) talking about players "going good," like the Force is with them. Players get hot and cold, and the industry secret to success is knowing how to deal with being cold. Sometimes when players slump they try to consciously change what they've spent their whole life learning how to do unconsciously: when to swing, when not to, when to release a slider, how to go back on a fly ball, etc. This approach rarely ends well at the big league level. Just ask Howie Kendrick.

This is called a regression fallacy. Highly random things, like coin tosses and batting averages, fluctuate naturally. Everyone knows this in baseball as streaking and slumping. But it is incorrect to attribute a cause to a natural fluctuation. This works both ways. Just like Howie didn't necessarily go from hot to cold because he did something wrong, the Angels didn't necessarily go from cold to hot because they did something right, in other words, there's no "stepping up." Of course baseball is ultimately played by humans who genuinely feel things like confidence and frustration, but that's not what makes a good (or bad) baseball team.

Oh great, another stat geek trying to tell everyone that everything they know about baseball is wrong, their favorite team sucks, all their hopes and dreams will come to naught, etc., etc. Well yes, maybe, but that's not my point here. What I mean to say is that, this season, the Angels are, have been, and will continue to be a good baseball team. A frustrating road trip doesn't mean the team has gone soft, and an exhilarating home stand isn't evidence that they've gained an edge. They're trying to win every day. I still think that, barring catastrophe, at the end of the year they'll have somewhere in the vicinity of 85-90 wins, which should be enough to edge the Rangers by a few games.

We've become so used to beating up on the Angels offense in recent years that we haven't noticed that it's actually become competent. It's not great, but it's at least average. Scoring 4 to 5 runs a night, as the Angels have on average, would give your team a decent chance to win a ball game if the pitching came through, and that's where the Angels have won their bread in recent years. Except for this year, where, while the starting pitching has been pretty good, the bullpen has been stupendously bad, and for no discernible reason. I still claim that this team's ultimate problem is in the pen, not in the batter's box. Without a healthy and effective Scot Shields, the Angels have lacked a late-inning fireman to come in the game and throw strikes. Pitching behind in the count has killed them with walks and hittable pitches. What is annoying is that this should be a reliable group of guys, although I still don't think Bulger is any good and Arredondo is just a kid. Get well soon, Scot. If you could take 6 of those bullpen losses and turn them into wins, the Angels would be 37-23, 12 games over .500 and 2.5 games up on Texas. That's made all the difference in our perceptions.

So in summary, Mr. Gubicza, don't blame "not taking responsibility" when things go bad. Blame something you can actually point to, like relievers not throwing strikes. You don't have to plot it or run a linear regression analysis or whatever. But it should be something you can actually tell someone how to fix, like, for example: throw more strikes. Otherwise, despite how many fancy-sounding words you use, in the end, you aren't saying anything more intelligent than "you suck!". I can give viewers that kind of commentary for much less than what Fox is paying you, but unfortunately, I can't do it with your greasy Superman coif. That costs extra.