February 23, 2007

Projecting The Top 3 Pitchers

Michael Salfino at SNY, profiles the top 3 pitchers in the Mets rotation:

Bill James, in the great Bill James Handbook that includes a bevy of enlightening statistics from our friends at Baseball Info Solutions, pegs Glavine to log 185 innings and finish 11-9 with a 3.89 ERA (101 Ks/62 BBs). Baseball Prospectus's famed PECOTA system, developed by Nate Silver, is markedly less kind to Glavine, giving him just 168 innings, a 10-9 record and a 4.03 ERA and a similar decline in K/BB rate. PECOTA determines which past players were most similar to current players. The current players are then predicted to perform as those past players actually did.

Let's give Glavine 200 innings of 4-ish ERA, 135 Ks, 70 BBs, and a 15-9 record given that his offensive support is likely to be in line with 2006 (5.91 runs per start, the seventh-highest total in the NL last year). Some interesting statistics on Glavine courtesy of the Bill James Handbook: his average fastball was clocked at 85.3 mph, not as slow as Greg Maddux, Livan Hernandez, Mark Redman or Kenny Rogers, and almost as fast as Barry Zito's (85.8 mph). Close Mets watchers will not be surprised to learn that Glavine threw the highest percentage of changeups in baseball by a mile (37.5 percent of all pitches).

PECOTA gives Hernandez just 21 starts and 10 relief appearances. I can't see Hernandez pitching in relief at all unless he struggles mightily. And the PECOTA projection doesn't anticipate bad pitching: 137 1/3 innings of 4.13 ERA with 122 Ks against 51 walks, but only an 8-8 record. Does PECOTA think the Mets are scoring 100 fewer runs this year than in '07? James has Hernandez at 9-9 with a 4.14 ERA over 163 innings (136 Ks, 58 walks).
The Mets would take the 163 innings James projects right now. Again, even with a 4.14 ERA, El Duque is unlikely to pitch .500 ball given that Mets lineup. And I believe he'll beat those ERA projections. He averaged over one strikeout per inning and his 164 Ks in '06 were the most he ever logged in the Majors. So, it's hard to argue that Hernandez's stuff is fading. Hernandez's ERA for the Mets last year should have been 3.60 and that's right about what I expect in 2007 for as long as he's able to stay off the DL. Over the 163 innings James projects, that should be good for 12 or 13 wins out of 20 or so decisions. Note that Hernandez had a 3.00 ERA and allowed 1.02 baserunners per inning at Shea last year.

James has Maine pitching 130 innings of 4.36 ERA. PECOTA gives Maine a slightly lower ERA (4.22), but fewer innings (119 2/3). James and PECOTA both project a 7-7 record. It seems like a .500 record is the default projection for any sub-200 inning pitcher with a middling ERA regardless of the team behind him.

I have no firm conviction of what Maine is going to do in 2007. But there is one substantive reason for Mets fans to be excited. Last year, only one starter in all of baseball held batters to a lower batting average plus slugging average on fastballs -- Roger Clemens. Maine's BPS allowed on fastballs was .529, Clemens's was .518.

No comments: