The Ten Win pitcher
No, not a pitcher credited with 10 wins in a season. Obviously no big deal, as thousands of pitchers have done that. I'm talking about a pitcher whose season rates as ten wins above replacement level. It is an extraordinary accomplishment, and the next pitcher who does so will only be the 10th since 1954.
I've been working on pitcher ratings, and have gone farther in adjusting for what needs to be adjusted than I've ever done before. I'm sure there are still more things that can be adjusted for, but here is where I am right now:
1. Start, simply, with runs allowed (total, not earned) and innings pitched. I'm not looking at any kind of defense independent stats or component ERAs or anything like that. Such measures are useful for projections going forward, but not necessary for evaluating what happened in the past.
2. Find out how many runs a theoretical replacement level pitcher would have allowed, given the ballpark, the defense he pitched in front of, and the opponents he faced.
Ballpark factors are the pitchers park factors from the database available from baseball databank. I may consider using another source, or reworking them later, but here I took them as is. Defense is a combination of Totalzone, outfield throwing, infield DP, and catcher ratings, all calculated from retrosheet data. A pitcher's share of the team defensive rating is simply the team rating multiplied by the pitcher's percentage of contact allowed, or (batters faced - strikeouts - walks - hbp - homeruns).
Say a replacement level pitcher, based on league average and park factors, would have allowed 110 runs. The pitcher pitched 10% of his team's contact allowed, and in front of a +50 defense. Replacement level would be adjusted downward to 105 runs, as a replacement level pitcher would also have benefited from the good defense.
Instead of using league runs allowed per game, I use a weighted average, by inning pitched, of the runs scored per game for all of the opponents a pitcher faced. Roy Halladay a lot against Boston, New York, and Tampa Bat, while Cliff Lee pitched quite a bit against Cleveland. The opponent adjusted league figures for the two are 4.85 for Halladay, and 4.62 for Lee, though even with that adjustment Lee rates as the AL top pitcher last year.
3. Once I've done all that to get runs over replacement, I figure a custom runs to win conversion - in a low scoring run environment fewer runs are needed to get an extra win.
So, without further ado, here are the 10 win pitchers:
9. 10.1 Bob Gibson, 1969 - 20 wins, 2.18 ERA, 314 innings
8. 10.3 Pedro Martinez, 2000 - Amazing year, 1.74 ERA in an hitter's park, in a high offense league. He's got the best rate stats ever for a starting pitcher, and saved 115 runs above replacement, the record from 1954-2008. And he does this with only 217 innings pitched, most of the other pitchers on this list had over 300. The thing that drops him down to 8th is the runs to win converter - in his high offense environment it takes 11.2 runs to convert to a win.
7. 10.4 Sandy Koufax, 1963 - 25-5, 1.88 ERA
6. 10.7 Wilbur Wood, 1971 - What a horse, 334 innings of a 1.91 ERA. He's getting some credit for doing this in front of a subpar White Sox defense.
5. 10.7 Roger Clemens, 1997 - 264 innings of a 2.05 ERA with the league going homer happy.
4. 10.8 Sandy Koufax, 1966 - His last season, 27 wins and a 1.73 ERA.
3. 11.0 Doc Gooden, 1985 - This one goes to 11. 1.53 ERA and a 24-4 record. He did it against a tough schedule too. His opponents averaged 4.16 runs per game, against a 4.07 league average.
2. 11.2 Bob Gibson, 1968 - The 1.12 ERA season, even in the year of the pitcher, was one of the greatest pitching feats ever. Gibson saved 85 runs above replacement, but it only took 7.6 runs to earn an extra win that year.
1. 11.6 Steve Carlton, 1972 - 27 wins on a 59 win ballclub. 346 innings, 1.97 ERA.
While there are two pitchers from the recent high offense era represented on the list, most of these guys pitched in the late 60's/early 70's pitching years. A low run environment allows pitchers to keep pitch counts down, coast through the Ray Oyler/Roger Metzger/opposing pitcher part of the order, and throw a larger percentage of their team's innings.
As a consequence, a pitcher like Koufax will be more valuable to his team, in the sense of earning more wins, in a pitcher's league. As long as we measure greatness in wins added, I don't think there is any way to get around this.