Have baseball's mathematicians built a better crystal ball to predict future pitcher performance?

In trying to answer this question from a 2004 Dodger perspective, I'm gonna try to keep this as simple as possible - for my sake as well as yours.

The theory: Major League pitchers essentially only control walks, strikeouts and home runs. On balls hit in the playing field, defense and chance play more of a role than the pitcher's skill.

As Jay Jaffe writes in his detailed examination on The Futility Infielder:

The Defense Independent Pitching Statistic (DIPS) system was invented by Voros McCracken. His studies of pitching statistics suggest that major league pitchers do not differ greatly on their ability to prevent hits on balls in play. The rate at which a pitcher allows hits on balls in play has more to do with defense and luck than to his own skill, and can vary greatly from year to year.

The upshot: By predicting what earned-run averages would be based only on innings pitched, strikeouts, walks allowed and home runs allowed (using a method of run estimation, a time-honored sabermetric concept, according to Jaffe), one can eliminate the chance elements from the equation and arrive at an defense-independent ERA, or DERA, that more accuarately reflects a pitcher's performance.

That number is more useful in predicting a pitcher's performance for the following season than his actual ERA is (obvious mitigating factors like injuries and age notwithstanding).

The mitigating factors: Subsequent studies on pitchers with long careers have shown that they do have some ability to prevent hits on balls in play, though this is still less than their impact on walks and strikeouts.

Let's see how DIPS works with Dodger pitchers from 2002 and 2003 as we look to 2004:

Comparison of Dodger ERAs with DERAs, 2002-2003

Pitchers whose 2002 DERA was higher than their 2002 actual ERA
(and therefore should have expected a poorer 2003):
Wilson Alvarez, Andy Ashby, Omar Daal, Kaz Ishii, Hideo Nomo, Odalis Perez, Paul Quantrill, Jeff Weaver.
Correct? Ashby, Daal, Perez, Weaver
Incorrect? Alvarez, Ishii, Nomo, Quantrill
Comment: Alvarez's injury recovery was probably a factor. Ishii and Nomo, both high-walk pitchers, continue to defy gravity in their ERAs. Note that while Daal had a 2.44 jump in actual ERA, his DERA remained almost the same.

Pitchers whose 2002 DERA was lower than their 2002 actual ERA
(and therefore should have expected an improved 2003):
Kevin Brown, Eric Gagne, Guillermo Mota, Paul Shuey (negligible)
Correct? Brown, Gagne, Mota, Shuey
Incorrect? None
Comment: One starting pitcher recovering from injury and three relievers.

As we consider the key guys on the 2004 Dodger staff, the system has had trouble predicting Nomo and Ishii - which may signal a flaw in the system, or may mean that those two are particularly due for a decline. Alvarez's ERA should normalize.

On the other hand, Weaver and Perez would be due for improvement, as would Dreifort if he is healthy. Even Gagne underperformed his 2003 DERA and could be considered a candidate to improve (!) this year. (Meanwhile, as many predict, Brown may be headed for a decline pitching in 2004 in front of the Yankee defense.)

I asked Jay Jaffe to comment further and he wrote, "Ishii and Nomo don't look so hot because of their high walk rates, but Nomo's K rate is good - he's not going to be your problem. Perez doesn't look so hot because his K rate is down, but he'll be OK. Call him and Nomo 'somewhat better than average but not spectacular.' Gagne ought to remain ungodly - it's not results on balls in play that makes him dominant, it's keeping balls from getting in play. Mota looks pretty good (though not as great as his ERA would lead you to believe), Dreifort too. I don't see any serious red-flags here, even with Nomo and Ishii's presence in the Higher dERA than ERA board."

Below is the full chart for the 2003 Dodgers, courtesy of Jaffe. Please go to The Futility Infielder to read much, much more about DIPS.

(BFP is actual batters faced by the pitcher. BIP is actual batting average against the pitcher on balls in play. ERAR is how many fewer earned runs the pitcher will have allowed over the number of innings pitched, compared to a replacement-level pitcher - that is to say, not an average player, but your average last man on the roster.)

Dodger DIPS Chart - 2003

I continue to think that Steve Colyer is a good candidate for the Opening Day roster.