Hot Streak vs. Permanent Skill Growth

• By: Chris Liss
• On: 5/31/2009 1:59:00 PM

Related: Chris Liss Minnesota Twins MLB 



• View Comments : 4

When a player's performing well, it's important to be able to distinguish between short-term positive results, i.e., a hot streak, and the emergence of a new baseline, i.e., permanent skill growth. In the former case, one should expect regression to the player's career mean. In the latter, a new mean is established, and the player should be valued accordingly going forward.

This week Erickson and I discussed Joe Mauer's power surge in Charging the Mound and much of the commentary both on RotoWire and Yahoo was that Mauer's streak was no different than Lance Berkman's from last year or Ian Kinsler's in April.

But while the small sample size argument is often valid, it's misplaced here. Mauer had 16 homers in 2007 and 2008 combined. For him to hit 11 in one month cannot simply be due to the normal fluctuation in balls leaving the park or pitches finding more of the plate. If Mark Buehrle, for example, pitched 50 innings and allowed just two runs with a 6 K/9, I wouldn't change my opinion of him going forward. But if Buehrle suddenly was clocked at 96 mph and struck out 60 batters in a 50-inning span, I would have to. The sample is exactly the same size, but in the latter case Buerhle would be displaying a heretofore unseen skill. And as our colleague at BaseballHQ says: "Once a player displays a skill, he owns it."

I'd argue that Berkman and Kinsler were like the successful Buehrle with the 6K/9 (or let's even say 6.5 K/9). Improved skills, but nothing radically different from one's normal skill set. Perhaps the slight boost in Ks moves the baseline ever so slightly, but you're not going to value Buehrle (or Berkman/Kinsler) significantly differently that you did at the start of the year. (Some people did with Kinsler this year, but in my opinion that's precisely the sample-size error that many think I'm making with Mauer.).

By contrast, the 96 mph/10 K/9 Buehrle (even if his ERA was 3.00), is a different pitcher. He has displayed a new skill. You cannot any longer simply assume regression to the mean, because the mean was established with an 89 mph fastball and 6K/9. There is a new mean, and you must value him accordingly.

Now it's possible that Mauer has just been extremely lucky, and he'll go back to being an 8-HR/year guy after hitting 11 HR in month. Possible, but unlikely because an 8 HR/year guy is just not capable of hitting 11 HR in a month. I'm not sure it's happened before in the history of MLB.

It's also possible a Kinsler type at age 27 does create a new baseline, too. But unless a player is in a massive growth phase, e.g. Justin Upton, the odds are against Kinsler, a good, but not elite prospect, suddenly taking his game to a far higher level than he did last year. Kinsler 2008, with the high BABIP, was likely his peak level, and he'd have to show sustained skills in 2009 beyond that level before I'd stop expecting a regression to the mean. One month is not enough.

But Mauer's performance is so uncharacteristic of his prior skill set, that it's almost impossible to think he hasn't shown substantive growth that's sustainable, health permitting.