Final NL Cy Young Thoughts

UPDATE: The voting is in — Carpenter wins, Willis second, Clemens third.

It’s time for my thoughts on the National League Cy Young Award, just in time for the official announcement on Thursday, November 10.

First of all, let me tell you the statistical categories that I think are important when determining the Cy Young Award winner, as well as a few that I don’t think are important.

Important:

Earned Run Average

ERA is definitely the most important stat in determining who the best pitcher in the league is. This stats tells us how many runs a pitcher allows, on average. Since the only thing a pitcher can do to help his team win is give up as few runs as possible, the best pitcher is the one who gives up the fewest runs.

Innings Pitched/Complete Games

The more innings a starting pitcher pitches, the fewer the bullpen has to pitch. That is a simple formula that is always true. So insofar as the pitcher controls how many innings he pitches, this stat is a great reflection of the value of a pitcher to his team. I believe that this stat holds less importance in the National League than in the American, though, because a pitcher’s ability to stay in a game is often impacted by his offense’s ability (or lack thereof) to score runs behind him. For example, Roger Clemens was pulled for a pinch hitter in the seventh inning or later of a close game (two runs or less) a whopping twelve times this season, including eight times when he had thrown 110 or fewer pitches (and three more where he was still well below 120). And even in the AL, a guy like Tim Wakefield can be fifth in the league in innings pitched just because he has a rubber arm, not because anyone would confuse him with an actual Cy Young candidate. So it is not the perfect stat, but it can tip the scales between two fairly equal candidates.

Shutouts

If ERA is the most important stat, obviously shutouts are very valuable. The only problem is that a shutout isn’t a shutout unless it is also a complete game, which means National League pitchers don’t completely control their shutout destiny.

Opponents’ OPS and batting average

This is another one of those tie-breaker stats. Let’s say you have two pitchers: the first one has a .250 average against him, but he only gave up two runs a game; the other guy has a .210 BAA, but his ERA is around 3.00. Obviously, all else being equal, the guy who gave up fewer runs consistently put his team in a better position to win games. (Please note that I am talking about past performance. If I were a GM looking for a starting pitcher, I might prefer the guy who allows fewer baserunners, just because giving up more baserunners but fewer runs could be a statistical anomaly.) But if you have two guys with 2.20 ERAs and one has a significantly lower BAA, the scales tip towards him.

Unimportant (or at least significantly less important):

Wins/Losses/Winning Percentage

Might as well start at the top, since this stat right here is the reason that I will often disagree with the results of the baseball writers’ voting. Do I think wins are important? Absolutely, they are the most important team statistic there is. But the Cy Young Award is not a team award — it is an individual award given to the best pitcher in the league. Does a pitcher have an impact on whether his team wins or loses? Absolutely, and if the win/loss record was the only statistic we had access to, we would be forced to make a lot of assumptions about the magnitude of that impact. But we need to look a little deeper. HOW does a pitcher impact the outcome of the game? By giving up as few runs as possible. Which means that ERA is the important stat. Assuming that the pitcher with the best W/L record is the best pitcher is like assuming that the team that wins the Super Bowl has the best quarterback, or that the Best Actor Oscar should go to the lead actor in the movie that made the most money. It might be true, but not necessarily.

Strikeouts

Strikeouts are the sexy stat for pitchers, and it makes sense. The less often an opposing player puts the ball in play, the less often he even has a chance to get a hit. If you have two pitchers with identical stats except that one has significantly more strikeouts, you would probably take the strikeout guy. But if the two pitchers are allowing the same opponent batting average, then it follows that the guy with fewer strikeouts actually gets out a higher percentage of players who put the ball in play. Does that make him better than the guy with a lot of strikeouts? No, but it certainly evens the score. Here’s an interesting tidbit: Nolan Ryan played for 27 seasons and was a first-ballot Hall of Famer, but he never won a Cy Young Award. Strikeouts are not enough.

Quality Starts

A “quality start” is any start in which a pitcher works six or more innings while allowing three or fewer earned runs. I like this statistic, although what it really boils down to is “a game when the pitcher does what you expect of him.” I think this stat is VERY useful if you are a general manager looking for a non-number-one starter. If you are looking for a number-one guy, or a Cy Young Award winner, I don’t think it tells enough of the story. Certainly a very accurate stat for what it measures, but not extremely important in comparing great pitchers, since a pitcher could, theoretically, have a quality start every time out and still have a 4.50 ERA.

So now that I have made it clear what I look at, let’s throw in one more stat that I made up: Adjusted Win/Loss (AWL). This record is compiled by figuring out the per-game ERA and comparing it to the league average for runs scored per game. So essentially, we are comparing every number to 4.59, the average number of runs scored per game throughout the Majors. If a pitcher’s game ERA was below 4.59, that game counts as a win. If it was over 4.59, it’s a loss. This is, in a way, even more accurate than ERA, because a pitcher could, theoretically, pitch 17 shutouts and give up five earned runs in each of his other 17 starts. He would have a 2.50 ERA, which is good, but a 17-17 AWL, which is a more accurate representation of how he pitched in those 34 games. On the other hand, a pitcher could give up four runs in every game and have an AWL of 34-0, despite being pretty mediocre. So it’s not a perfect stat, but it can be a handy tool in our Cy Young-picking toolbelt.

In the National League, I looked at each of the seven pitchers who finished the season with an ERA under 3.00, and here is a listing of their AWLs, along with their actual records and ERAs:

Name	AWL	AWL%	ERA	W/L
Andy Pettitte	30-3	.909	2.39	17-9
Roger Clemens	28-4	.875	1.87	13-8
Chris Carpenter	27-6	.818	2.83	21-5
Dontrelle Willis	27-6	.818	2.59	22-10
Roy Oswalt	27-7	.794	2.86	20-12
Pedro Martinez	23-7	.767	2.82	15-8
Jake Peavy	22-8	.733	2.88	13-7

As we see, if AWL were the only judge of a pitcher’s quality, Andy Pettitte would be the best pitcher in the league. But given the limitations of this stat, we need to look at AWL in conjunction with more traditional statistics, which is why I only included the seven pitchers with ERAs under 3.00 in the analysis.

In my opinion, the AWL combined with the ERA narrows the list down to four options: Pettitte, Clemens, Carpenter, and Willis. Other than Pettitte, these are the guys you hear talked about for the award all the time. Let’s see what they each have going for them among the stats that I listed as important above:

Name	ERA	IP	CG	SHO	BAA	OPSA
Andy Pettitte	2.39	222.1	0	0	.230	.616
Roger Clemens	1.87	211.1	1	0	.198	.544
Chris Carpenter	2.83	241.2	7	4	.231	.623
Dontrelle Willis	2.59	236.1	7	5	.243	.644

Now let me just say this up front: I will be very surprised if Willis and Carpenter do not finish 1-2 or 2-1 in the voting. The award is named after Cy Young even though he isnââ?¬â?¢t the best pitcher of all time–it is named after him because he had 511 wins in his career. The “most wins equals best pitcher” equation has been around for a hundred years, and as evidenced by the fact that Bartolo Colon won the American League Cy Young Award, it is still around. (More on that vote in a bit.)

Okay, so going back to the stats that I believe are important to the Cy Young voting. Since this is the National League race we�re talking about, we need to reduce the emphasis we place on innings pitched, complete games, and shutouts, since those things are not entirely within the control of the pitcher. We will not eliminate them completely, but we need to look honestly and fairly at how well each man did within the constraints afforded him by his offense. But first, let�s look at the three stats that are left: Earned Run Average, Batting Average Against, and OPS Against. (In case you don�t know, OPS is On-Base Plus Slugging percentage, a simple sum of the on-base percentage and the slugging percentage.)

ERA: Roger Clemens clearly dominated this category, with an ERA over a half-run better than the second-place finisher, his teammate Andy Pettitte. Clemens bested Willis by three-quarters of a run and Carpenter by nearly a full run. I�ve already told you that ERA is the most important stat for Cy Young, so based on this, Clemens has jumped out to a huge lead.

BAA: Clemens led the league in Batting Average Against, also, at .198. It is nearly unheard of for a starting pitcher to go through a full season with hitters batting less than .200 against him. Pettitte�s .230 BAA came in just ahead of Carpenter�s .231, with each man a bit ahead of Willis�s .243. More points to Clemens.

OPSA: Again, Clemens delivered a dominating performance, with a .544 OPSA. That is a full .100 better than Willis (.644), with Pettitte (.616) and Carpenter (.623) each about .075 behind Clemens. Still more points to Clemens.

Now let�s look at the other three: innings pitched, complete games, and shutouts. Basically, we have to see if there is any reason in these three stats to give the award to someone other than Clemens.

IP: Carpenter led these four guys with 241.2 innings pitched, followed by Willis (236.1), Pettitte (222.1), and Clemens (211.1). As mentioned above, though, Clemens was pulled for a pinch-hitter for reasons unrelated to his pitching somewhere between eight and twelve times this season, which cost him somewhere between 16 and 36 innings pitched. Considering that he was only 30 IP behind Carpenter, that is very substantial.

CG: Carpenter and Willis each had seven, Clemens had one, and Pettitte had none. There is no way to tell how many complete games Clemens could have had, but I think it is fair to say that of those eight-to-twelve games where he was pulled early due to lack of offense, he certainly could have finished at least four or five of them, which would have put him right up there close to Carpenter and Willis.

SHO: Willis led the league with five, Carpenter had four, and Pettitte and Clemens each had none. Clemens, however, was pulled for a pinch-hitter in three different 0-0 games, which means he easily could have had three shutouts if his team had been able to score for him. Carpenter�s four shutouts had scores of 3-0, 4-0, 7-0, and 8-0, which means he was never really in danger of being pulled for a pinch hitter in any of those games. (For the record, Carpenter was never pulled for a pinch hitter while pitching a shutout.)

It is clear that Roger Clemens was the best pitcher in the National League this season. By all rights, he should have won at least 22 or 23 games this season, maybe more (28 by my AWL calculations). It is as simple as this: if the Astros had scored more runs and he had won 23 games, he would be the unanimous Cy Young Award winner. Did you get that? Clemens could be the unanimous winner without doing anything different from what he did! To not give the award to Clemens simply because his team had a lousy offense is to completely skew the purpose of the award, which is supposed to go to the best pitcher.

If I were voting, I would go in this order:
1. Roger Clemens
2. Chris Carpenter
3. Andy Pettitte
4. Dontrelle Willis

I debated about the order of Carpenter and Pettitte, because Pettitte had better numbers in the three big categories. I put Carpenter ahead of Pettitte because Pettitte didn�t dominate in any of the big three the way Clemens did, which left room for Carpenter to sneak ahead with his dominance in the IP, CG, and SHO categories.

How do I think they will actually finish? Well, I don�t expect Pettitte to crack the top three. I expect it to end up Carpenter, Willis, Clemens, in that order, but I wouldn�t be surprised to see Willis win.

Now, let�s talk briefly about the AL. Bartolo Colon won, but my vote would have gone to Johan Santana, with Mariano Rivera second and Mark Beurhle third. Colon would have finished fourth in my voting. Jayson Stark wrote an excellent column for ESPN.com explaining why he would have picked Santana, and I agree with everything he wrote. I won�t rehash it here, but I definitely recommend that you read that article.