=== Baseball-Reference.com Comps ===
If you look up the guys that B-Ref.com has as comps for Lee, they're not much use. The list of 10 has an aggregate ERA+ of 105, because James' formula uses career totals. Cliff Lee, of course, was one pitcher before 2008, and another pitcher after 2008.
John Burkett, Kirk Reuter and Charles Nagy are in there because they didn't walk many guys. But as you can see, "doesn't walk many guys" is not a defining characteristic of a pitcher family.
Let's assume, for a moment, that the Cliff Lee of 2008-09 is the Real Cliff Lee. He did throw 300 innings last year, including spring training, so it's not going to be a shock if he feels the usage in 2010. But that's about the only quibble you could come up with.
The way that Lee pitches is very repeatable. He's 30; he has mastered his mechanics; there's nothing about his game that depends on a phase-in, phase-out feel for the baseball on a given night.
.
=== Defining Characteristics ===
James broke down pitchers into "families" so that he could identify useful similarities between pitchers.
His "picture-perfect" family was for righties -- Bunning, Saberhagen, Sutton, and guys like that are in it. These were guys who did not have huge fastballs, but had several plus pitches, plus command, plus everything...
"You watch them on a good night and you think, I don't know how this pokey ever loses. But they do lose, because they're always around the plate" and so usually give up their share of HR's.
Notice that Lee, before 2008, gave up 1.5 homers a game and few walks. Lee, after 2008, gave up 0.5 homers a game and few walks.
.............
Anyway, LaMarr Hoyt isn't a comp for Cliff Lee. A comp for Lee 2008-09 would be this guy:
- Left-handed
- Average-solid fastball: 88 > FB > 94
- Great change-speed game and pitchability
- Tremendous command: 2.5 > BB
- Nice but not overwhelming K's: 6.0 > K > 8.5
(All numbers would need to be adjusted for era. They didn't strike out 9 guys a game in the 1950's.)
Take out any one of the above 5 characteristics and jemanji utterly rejects the pitcher as a comp.
Obviously soft-tossers like Jamie Moyer and Mark Beuhrle are a different kind of pitcher than Lee is.
Similarly, Randy Johnson, Sandy Koufax and Lefty Grove might have the same 4:1, 5:1 control ratios, but they are not getting hitters out the same way. For this reason I would throw out Mark Mulder, though that's arguable.
.
=== Top Ten Comps List ===
So departing from the PECOTA and B-REF.com auto-generated comps, here is a Seattle Sports Insider list of Cliff Lee comps. They would be used the same way: you would plot career arcs with them, identify their best contexts, etc.
Following are most of the guys who qualify as post-1970, Picture-Perfect Lefties in the Cliff Lee mold:
Johann Santana (last coupla years, with less steam in his arm)
Greg Swindell, an underrated pitcher who came up with the sad-sack Indians
David Wells up until about 2002
John Candelaria, who pitched before the offensive explosion
John Smiley, a half-notch inferior to most guys on this list, but a bit scary...
Teddy Higuera, a small notch more of a power pitcher than others here
Sid Fernandez, definitely more of a power pitcher, but has the great CTL
Denny Neagle, a beautiful comp, and the only comp b-ref.com got right (and #1 on their list)
Jon Matlack, who was Tom Seaver's #2, so particularly apropos w/r/t Felix
Andy Pettite, yet another 91 mph Picture-Perfect Lefty
If we were going to throw somebody out, it would be El Sid, who used two pitches and had an unfair slider, as well as too many K's. El Sid is more in the Koufax/Johnson family. But then we wouldn't have 10. :- ) And besides, El Sid got it done with a 90-92 fastball.
So there are ten pitchers who are pretty fair stylistic comps to Cliff Lee. We're going to be a lot better off working with an edited list than with one generated by computer.
.
Part 2