Inside the numbers -- Breaking down this year's NL Cy Young Race

As we wind down the season the biggest debate concerning pitching in September is who should win the Cy Young awards. The award ostensibly goes to the best pitcher in each league, but much like the Heisman Trophy which is intended for the best college football player but invariably goes to the most-hyped, quarterback or running back for a major school with a good record, the Cy Young usually goes to the starting pitcher with the most wins, and failing that, the best ERA. When it doesn't, it then goes to the closer with the most publicity.

For years I've railed that since wins and ERA are mostly out of the hands of starting pitchers and in the gloves of their defense, the bats of their offense, and the arms of their bullpen, using those stats to determine the best pitchers is like using height to pick a president. Let's see what got us here, and what we should do to correct the error of our ways.

The award was created for the 1956 season and was only awarded to one pitcher in the Major Leagues until 1967, when the practice of giving separate league awards began. The award was named after a pitcher who holds the unbeatable career record for wins (511) and career innings pitched (7,355). However, he also holds the record for most career losses (311). To put that into perspective, he had as many losses as Tom Seaver had wins.

However, what most likely got Young immortalized with the 1956 eponymous award was the shrewd decision to die in 1955. If the award had been created in 1926, perhaps it would have been called the Christy Mathewson Award. "Big Six" would have been a better candidate to represent all-around pitching talent. With only 65 percent of the innings Young threw, he still had a better winning percentage, better ratios, and more shutouts and is tied with Grover Alexander as the NL career wins leader. He also was instrumental in creating the newly-formed MLB with his star power and good character, and openly accused the Black Sox of throwing the 1919 World Series. His career was cut tragically short when he was accidentally gassed during training in World War I.

But since the award was named after the MLB career wins leader, the idea of honoring the season wins leader was likely hard-wired into the heads of the Baseball Writers of America (BBWA), who were given the job of voters.

For the next three seasons, the voting process was a simple popular vote (like it had been since the award's inception), with balloters naming a winner in each league. For those six Cy Youngs (three seasons each for the AL and NL), the wins leader took five (83.3%) of those trophies. After that, the award was based on a weighted ballot, with members of the Baseball Writers Association of America listing their top three candidates.

Including all of those post-1966 seasons up to 2007, we get a very interesting statistic: in both leagues, in 41 seasons, the wins leader won the Cy Young 25 times (61.0%). However, in the NL, of the other 16 remaining seasons, the ERA leader won it 7 times (78% combined), while in the AL, the ERA leader won 6 times (75.6%). And to add to the statistical oddity that the AL and NL have mirrored each other, of the remaining seasons, both leagues awarded the Cy Young to relievers with 25 saves or more. And incidentally, if someone can tell me how to characterize Mike Marshall's fatigue-inducing 1974 NL Cy Young-winning season (106 G, 208.1 IP, 15-12, 21 Sv), I'm all ears.

As for the coveted 20 wins plateau, the first non-20 game winner in the NL to win was Seaver in 1973, which made Ron Bryant (24-12) the first 20-game winner to get stiffed in either league. The difference was likely their ERAs -- 2.08 for Seaver vs. 3.53 for Bryant -- which helped Seaver make up for five less wins (Bryant would have only three more wins in his MLB career). In the AL, the first starter (now follow me on this) not to achieve 20 wins in a non-strike shortened season where a starter did reach 20 wins and win the Cy Young was Roger Clemens who in 1991, beat 20-game winners Scott Erickson and Bill Gullickson.

When a player changes leagues during the season, the player's statistics are kept separate, so any starter changing leagues already has once strike against him, perhaps literally. It takes a stellar performance to then be considered a true contender for the Cy Young. In 1984, Rick Sutcliffe went 4-5 for the Indians before a trade to the Cubs saw him go 16-1 the rest of the way, which was good enough for the NL Cy Young. That meant he wasn't the official wins leader for the NL, but his combined 20-6 would have tied leader Joaquin Andujar's 20 (for the statistics in the previous section, I counted Sutcliffe as a wins leader in 1984).

So technically only the pitcher's numbers in the league he's being considered in are counted, but it's my belief the entire season should be considered. Using this criteria, Randy Johnson would have given Tom Glavine a run for his money in 1998 when his 9-10 record in Seattle combined with his 10-1 record in Houston would have given him only one less win than Glavine, who won the Cy Young Award that year. The Big Unit's 329 combined strikeouts coupled with his stingy 12 ER in his 11 Houston starts made for a compelling case.

Yes, all relief pitchers should be considered for the Cy Young, not just closers. As noted above, the only non-starters to win Cy Youngs have been closers, who are judged by the partly-bogus statistic, the save. While taking the mound in the ninth with the tying run at third and the winning run at second and preserving the win is an admirable quality deserving our respect, getting three outs in the ninth with no one on base and a three-run lead isn't, especially if you give up two runs. But both situations will earn a closer a save.

So if we're going to reward that behavior, we should reward any pitcher who takes the mound -- in a pressure situation or not -- and consider middle relievers as well. Think about the dominant setup man who would likely close for another team and consistently ends the game in the eighth. Think about the long reliever who gives his teammates a rest and can keep them in a game after a starter's early exit. Think about reliever John Hiller for the 1974 Tigers who went 17-14 in 150.0 innings with a 2.64 ERA. That's better than most starters with the same amount of innings. He was well behind AL Cy Young winner Catfish Hunter's 25 wins that season, but only incrementally higher than his 2.49 ERA. And anyone who reads my columns knows that I don't buy the argument that vulture wins are wholly unpredictable.

So yes, relievers should be eligible as the award is for the best pitcher of the season, but we shouldn't over-value saves.

First Criterion: Don't ignore wins, saves and ERA, but put them into the proper perspective. They are dependant on the pitcher's team as much as the pitcher (I say even more so), so there is an element of good (and bad) luck involved.

Second Criterion: Look at statistics that measure a pitcher's ability to keep men off-base, such as WHIP (I still maintain WHIP should be modified to include hit batsmen) and his control, such as K/BB and K/9IP. I know the latter is a penalty for ground ball-inducing pitchers like Chien-Ming Wang, but good ones are rare enough that we can analyze them on a case-by-case basis.

Third Criterion: Since external factors affect a pitcher's key stats, we need to look at pitchers against the other starters on their own team. Conceivably a 162-game season means statistical bias should dampen out (but not be wholly eliminated, of course), so pitchers on the same team should get the same level of defense behind them, the same amount of run support, and the same bullpen effectiveness, all factors in wins and ERA. Look at the wins and ERA of the candidate against his teammates in the rotation (this obviously doesn't work for saves). Then compare the candidate to pitchers on other teams by looking at the differences in the teams' runs scoring and bullpen effectiveness.

Other key accomplishments like shut outs, complete games, high innings pitched and strikeouts may also be considered, but are more likely to come into play when the race is close.

There are still two weeks of the season to go and a surge or decline by one candidate or another could happen. This week I'll analyze the NL Cy Young race, and next time, I'll analyze the AL.

The NL has been infamous for not generating a 20-game winner since 2005, but Brandon Webb has finally broken through with his 20th win. I suspect that most BBWA NL ballots will start with Webb, so we'll use him as our presumptive winner to eliminate pitchers that weren't as good as Webb.

For starters (for starters) I used the typical stats discussed above for the first and second criterion, and for the third, I used the following calculations to compare the pitcher to his team:

W%(Rot): Win Percentage of Rotation, or W(starter)/W(rotation), which tells us what percentage of the wins by all the team's starters were achieved by the pitcher

ERA (Diff): ERA Differential, or [ERA(starter)-ERA(rotation)]/ERA(rotation), which tells us how much better (if negative value) the pitcher's ERA was than the whole rotation's ERA.

R/G: Team runs scored per game

For relievers, I used the same stats for the first and second criterion, but for the third added in SV%, which is simply the percentage of the team's wins that were saved by the reliever. For non-closers, this doesn't tell us anything, so I omitted it.

Keeping in mind that Webb is our prohibitive favorite, we can take a look at the relievers and see if any merit further discussion. We can eliminate Wilson and his awful ratios, and while the two middle relievers -- Kuo and Marmol -- have done excellent jobs, they've done so for first place teams and nothing screams "special" about their seasons. Fuentes has done a stellar job for a bad Rockies team, but giving the Cy Young to someone with a single win and less than 30 saves doesn't pass the red-face test. As for Valverde, I owe him a public apology as I rode him hard the first week of the season and declared his imminent collapse. He has been instrumental to an Astros team that just won't go away. However, those ratios aren't good enough to warrant consideration. That just leaves Lidge, who has been incredible this year and big part of the Phillies' success. While I doubt it can argued successfully that he is a better candidate than Webb, he likely has merited a top five spot in my voting.

For the starters, let's eliminate the two Cubs -- Dempster and Zambrano -- as neither is leading the league in wins despite playing for the best team in the NL. Next goes Haren because we can't even argue that he's the best starter on his own team. Next goes Billingsley because he plays for a better team than Volquez, but still doesn't have as good numbers. Hamels also walks the plank because even with the best bullpen in the NL, he's only mustered 13 wins, and in many respects, the "I've Learned My Lesson" version of Brett Myers has been the Phillies best starter (but his horrible start keeps him off this list). Sheets is out because his team scores runs and his bullpen prevents them, but he still has only 13 wins. Finally, Nolasco and Santana leave hand-in-hand for two reasons: (1) even though they've been hard luck pitchers, they're still too many wins away from Webb, and (2) I just wanted to remind those fantasy players that paid $40+ for Santana that they could have gotten slightly better numbers from Nolasco, who went undrafted in many leagues.

This leaves Cook, Lincecum, Oswalt, Sabathia, Volquez and Webb. Oswalt overcame an abysmal start this year to be one of the most consistent post-All Star Break pitchers in the NL. Cook and Volquez have awful WHIPs for Cy Young contenders, but both played very well for awful teams, but Volquez gets the nod over Cook because of the strikeouts (and yes, I'm penalizing Cook for strikeouts). Sabathia has mixed stats from playing in the AL and NL this year, but has just eaten the NL. Webb was phenomenal to start the season and sputtered as his team gave up more runs in the ninth than it could score in the previous eight innings. But at least he played for a team with talent, as opposed to Lincecum, who toiled for a team many thought would be the worst in the NL (surprise, they could finish second in the NL West). Besides being able to squeeze 17 wins out of that team, what really stands out is his ERA was almost 44% better than the entire Giants rotation's and he had a gaudy K/9 above 10.0.

So, based on the numbers in the three criteria, here's my final vote:

1. Tim Lincecum (SF)

2. Brandon Webb (ARI)

3. CC Sabathia (CLE/MIL)

4. Edison Volquez (CIN)

5. Brad Lidge (PHI)

David Young is a fantasy baseball expert who has written weekly columns for, among others, ESPN and Sports Illustrated's websites. He also won KFFL's Expert League in 2007. Send him a comment or question at davebaseball29@yahoo.com.