Will Cunnane: Misused and unlucky?
After my last cutting edge analysis of catchers I decided to shelve that awesome endeavor to a later date. A much less daunting task fell to me courtesy of the work of Voros Mcracken. For those not familiar with his work, you should be, but I'll provide a quick summary. Two years ago he started a project to quantify a pitchers ability to control opponents batting average. His finding were that a pitcher had no control over the hit rate of a ball put into play. In short if the pitcher doesn't strike out the batter or give up a home run the only influences on whether or not the ball becomes a hit are the skill of the defense, and the skill of the hitter. I won't go into his methodology other than to say it is very complex, and as far as I can see accurate and appropriate. Since his initial findings he has worked to expand his sample size and to discover other factors that may have small impacts (very small). His most recent data indicates that his initial findings were insufficient. There does appear to be a very small effect that the pitcher has on balls in play.
All that said his truly intriguing contribution is publishing ERAs for each pitcher that are essentially totally independent of any effect park, defense, luck. In short they allow for quick and easy comparison of the pitchers contributions to the team. To that end I thought I'd pull those stats together and examine the ERA of our pitchers as measured conventionally (MERA) and compare it to the defense independent ERA that was published (DERA). Using DERA and the innings pitched for each pitcher I was able to determine how effective our defense was compared to average. During the regular season our pitchers allowed 740 earned runs (ER). Based on the DERA we should have allowed a park adjusted figure or 760 ER. This gives a difference of 20 runs that our defense saved over average. That understates things a bit since the actual ER number wasn't park adjusted. Adjusting for the slight offensive nature of Miller Park means that the defense was actually worth 25 runs last year. Using this data I was able to calculate yet another ERA number (PERA) this number is the DERA multiplied by the defense adjustment (0.9737). This calculates what the pitchers ERA should have been measured as last year. It eliminates biases due to bullpen help or hindrance, and other luck factors. To review DERA is the number that reflects the pitchers value independent of factors outside his control. MERA is what you would normally call ERA. PERA is what the pitchers ERA should have been with the defense accounted for, but not other luck factors. The final number I give to you is LERAC, which is merely the difference between PERA and MERA. A positive number means that his ERA should have been higher than it was last year, and a negative number means that his ERA should have been lower last year. I also included at the end a few extra pitchers who didn't play for Milwaukee last year but we have discussed. Their PERA values reflect what they would have done playing for us last year (and presumably close to what they would do in 2002).
There are many interesting things to notice. I'll start with the pitchers who really got shafted last year. Will Cunnane should have posted a 4.16 ERA as a long reliever instead got hung out to dry for a 5.40. Misused by Davey and bad luck means another team got a good pickup. Jimmy Haynes should have posted a 4.42 screwed with a 4.85. Conclusion probably just his career year, but he out pitched Wright by a lot. Mac Suzuki numbers say 4.65 hosed with a 5.30. Kyle Peterson should have been credited with a 4.24 instead he became possibly the best rule V pick because of his 5.52. My final mention goes to Allen Leverault should have had a 5.68 and got tagged with a 6.08. Conclusion he still wasn't that great, but not really a bad first year in the majors. Many other players had either big differences and small playing time or the opposite.
The next group is the beneficiaries, those players who had superficially good looking years. Basically our entire stud bullpen built a lot of it's reputation by screwing the starters over or a select few other relievers. Fox and Stormy both had a difference of greater than 1 full ERA point, while Leskanic and Dejean were both close to 1 point differentials. My conclusion is that we might want to deal a few more of these guys while people have over inflated opinions of them especially since we already let 3 guys go who could have done pretty well in the bullpen with proper use (Suzuki, Cunnane, and Haynes), and Dean seems good at finding others.
My special category gives mention to Ray King and Ben Sheets as the players who were closest to their predicted scores. It also notes that Wright's lackluster season was understated as his ERA should have been a 5.13 last year. My conclusion is to deal him fast, since it would take a complete recovery from injury and Stewart working magic to make him good enough to be worth 4.5 million to us this year plus note that Wright was extremely dependent on the double play last year and with Belly moved it just might get ugly.
Especially when I can talk about some other guys to do the job. First a mention of Rusch. The numbers say that he should have posted a 3.87 ERA if he played with us last year. So unless he betters that mark don't give Dave Stewart one once of credit for his success next year. The much mentioned cheaper younger Chen would have posted a lackluster 5.02 ERA. Though he is still young enough to improve greatly it would seem that Dean's scouts may have been on to something. My much suggested Omar Daal (for the low cost of Loretta which makes him free) would have given us a middle of the road 4.72, which might be worth something at the deadline and might not. The final name that I bring up once again and he's still available if we deal Wright is Ron Villone who would have only posted a 4.17 ERA last year. Call me crazy, but with Ron and Rusch stablizing the rotation plus improvements from Sheets, Que, and Nuegie it could be a pretty darn good rotation.