Thursday, May 31, 2007

Examining Roger Clemens

I’m sure everyone has heard by now that Roger Clemens will be suiting up in pin stripes this season.  The Yankees signed Clemens to a prorated contract of $28 million (or roughly $18 million) at the beginning of the month, and the Rocket is set to take the mound for his first start on Monday against the Chicago White Sox.  At 44 years old, how will Clemens do, especially now that he is back in the American League?

In the past, Clemens has been nothing short of dominant, even on the wrong side of 40, but most of those years were played as a Houston Astro in the National League.  In 2004, Clemens posted a 9.15 K/9 and a 3.32 BB/9.  In 2005, he put up a 7.88 K/9 and a 2.64 BB/9.  Last year, he had an 8.10 K/9 and a 2.30 BB/9.  He consistently put up a very good groundball rate of 49% during those years.  But taking into consideration that they were played in the National League, without the DH, and with a much better Astros defense (compared to the Yanks), we are sure to see some sort of decline from Clemens, who is also a year older now.

If he were playing for the Astros again, it might be reasonable to expect a K/9 around (or perhaps slightly under) 8.00 and a BB/9 around 2.60.  Add in a 49% groundball rate, and you’re looking at an ERA in the low 3s.  But Clemens is not playing for the Astros, so we need to adjust for this.  With some help from David Gassko, I think we can come up with a pretty good estimate.  Thanks, David, for these numbers!

To start, American League hitters are .5 runs per game better than National League hitters.  In addition, the Designated Hitter in the AL accounts for another .4 runs per game.  Now, if we take the change in defenses into account, we see that Clemens should let up even more runs per game.  With all the ground balls Clemens induces, the switch from Adam Everett to Derek Jeter should have a somewhat negative effect on his stats.  In the THT 2007 Season Preview, the defense of the Yankees was projected to be roughly .5 runs per game worse than the Astros.  If you look now, though, the Yankees defense has been playing a bit above average in 2007.  There’s no guarantee they will keep it up, but it is certainly a good sign for Clemens owners.

So how does all of this affect Clemens?  Well, if we add up the extra runs Clemens should accrue, we find that he should be letting up anywhere between .9 and 1.4 extra runs per game.  .9 seems to be on the low end, since that would assume the defenses of the Astros and Yankees are equal.  Since the Yanks defense is performing better than expected, though, 1.4 seems to be on the high end.  Overall, I think 1.2 seems like reasonable number to settle on.  That would say the Yankees’ defense is .3 runs worse than the Astros’.

So, based on these numbers, we should see a decrease in Clemens’s strikeout totals, possibly to the mid 7s, and an increase in his ERA.  An ERA around 4.40 looks to be about right, taking all of these factors into consideration.  There is one more factor we should consider though.  With the Yankees starting pitching a disappointment (Mike Mussina‘s collapse, Philip Hughes‘s injury, etc.) so far this year, and with the poor performance of the bullpen, how hard will Joe Torre and Ron Guidry push Clemens?  How deep will they let him pitch into games, and how will this affect him?

As Yankees, they will always have one eye looking toward the postseason.  As such, they will not want to ride Clemens too hard, but they will want to get the most out of him.  Let’s play devil’s advocate for a minute and say that they will allow him to throw over 100 pitches nearly every game.  How has Clemens done when throwing this many pitches in the past?  I took a look at his numbers from the past three years (lumped together, to get a good sample size.  The Astros didn’t let him throw over 100 pitches very often).  The stats I looked at cut the pitch counts off at 90 and 105, so I created two groups.  One group is broken down into pitches 1 to 105 and 106 to 135, and the other group is broken down into pitches 1 to 90 and 91 to 135.  The breakdown can be seen below.

GROUP 1
1-105 pitches
TPA - 2060
K/TPA - 23.64%
BB/TPA - 7.96%
HBP/TPA - 0.58%

105-135 pitches
TPA - 74
K/TPA - 24.32%
BB/TPA - 8.11%
HBP/TPA - 1.35%

GROUP 2
1-90 pitches
TPA - 1812
K/TPA - 23.84%
BB/TPA - 8.17%
HBP/TPA - 0.50%

91-135 pitches
TPA - 322
K/TPA - 22.67%
BB/TPA - 6.83%
HBP/TPA - 1.24%

Surprisingly, we see that in the first group Clemens actually strikes out more batters with later pitches.  That group is looking at a small sample size, however, so we’ll focus on the second group.  In that group, we see a more noticeable drop in strikeouts.  We also see that he tends to hit more batters but - oddly - walks far fewer.  Overall, I’d say the biggest thing to worry about, should the Yankees decide to work Clemens hard, is his durability as it pertains to future starts.  He doesn’t seem to pitch terribly worse when throwing a lot of pitches, and I doubt his numbers would suffer much - in the short term anyway - from the Yankees occasionally have him throw 100+ pitches.  Perhaps we could bump his expected ERA up to around 4.50 if the Yankees decide to pitch him deep into games often.

I think that’s enough about Clemens, for now.  Overall, I’d say he is a good player to own, but won’t end up with Top 10 rate stats.  If someone in your league thinks he will put up numbers like he did with the Astros, sell him.  Otherwise, keep him for yourself and see how things go.

If you’re interested in reading more about the differences between American and National League hitters, Mitchel Lichtman’s article Is the AL Really Superior? (Part 3) is a good read.