Still Filthy

In the 2007 regular season, Josh Beckett went 20-7 with a 3.27 ERA, and finished second to C.C. Sabathia in the Cy Young voting. So when he dominated in October--4-0, 1.20 ERA, 35:2 K:BB ratio--we were spared some (not all, but some) of the "clutch" talk, since the Cy Young runner-up is expected to win in the playoffs.

This year could be different. Beckett is just 12-10, with an ERA of 4.03. If he pitches well in the postseason, we will inevitably be told the he struggled all year, but picked it up in October when the clutchiness kicked in.

Don't believe that for a second. In terms of peripherals, Beckett's '07 and '08 seasons have been remarkably similar:

The only thing that's really changed is his ability to get ground balls (which in turn effects the HR%). His K:BB ratio is actually higher this year, as his K rate has improved slightly. The difference in his actual ability amounts to less than five runs over 35 starts. Just something to keep in mind when Joe Buck is describing how Beckett has had a "down" year.

Run Environments and DIPS

In 2007, American League teams allowed, on average, 4.82 runs per game. This year, that number is down to 4.49. The table below sheds some light on what factors have contributed to this discrepancy.

The only two changes that seem particularly significant here are H/9 and HR/9. But the increase in homers doesn't match up with the GB%, which has barely changed. Here are the more detailed batted ball stats, from FanGraphs:

Fewer fly balls overall, and more infield flies, have allowed for a 5.8% decrease in outfield fly balls, a category which I just made up. The increase in line drives is somewhat strange, since the average DER was .692 last year, and is up to .703 in 2008. Incredibly, about 1/3 of that increase can be explained by Tampa. Some of it is a zero-sum game (trading Harris to the Twins for Barlett), but not all of it (Upton to center, Aki to second, calling Longoria up).

Anyway, the point is that these changes, regardless of whether they're temporary or not, have caused issues with some DIPS stats. I'm going to stop isolating the AL now, since I'm not sure that there's any reason to do that other than for illustrative purposes, and turn to the majors as a whole. The numbers below are for guys who have thrown at least 90 innings thus far in 2008.

So if you see that a guy's QERA is a bit higher than his actual ERA, don't think too much of that, since that's true for the whole league. The same is true for FIP, although to a lesser extent.

Finally, I'm not sure that this would have much predictive value, but if you run a regression on the QERA variables (K%, BB%, GB%) using these 114 guys, you get this:

(In both cases, you then have to square the result to get the expect ERA. QERA is also explained/defined here.)