A recent article by Aaron titled "Best Individual Pitching Seasons" has generated a great deal of debate. Much of the discussion has centered on Aaron choosing Pedro Martinez's 1999 season as his best - and the all-time best. Many readers have argued that his 2000 season was even better. In the analysis that follows I hope to show, by digging into the numbers a little deeper, that Pedro's 1999 season is indeed his best, and that frankly it isn't as close as you think.
First let's review the situation by looking at the basic stats for these two seasons.YearWLGIPBFHHRBBKHBPERALGERAWHIP199923431213.383516093731392.075.020.923200018629217.08171281732284141.745.070.737
Before we move on, let us all pause for a moment and stand in awe of what we have witnessed. For two years, while the rest of the AL was allowing FIVE runs per game, Pedro Martinez was allowing LESS THAN TWO. This is stunning. This is unprecedented. ‘Yes' his team failed to win the World Series, and ‘Yes' he had the misfortune of compiling these statistics while one of the greatest teams of all-time played in the same division, and ‘Yes' that team happened to play in the media capital of the free world, BUT... How is it that nobody was writing epic poetry about what this man was doing?
Why was there not a "SportsCenter Special" on the complete and utter dominance of Pedro Martinez? All of us who are sports fans live for those events that we know we will be talking about when we're old and gray and living at Del Boca Vista. Pedro Martinez at the turn of the century is such an event.
Back to the issue at hand. Some have argued that the 2000 season was the greater season because Pedro's ERA and WHIP were both lower, and the extra wins he compiled in the 1999 season are the result of the Red Sox providing him better run support. While these assertions are true, further analysis shows that the conclusion is not.
The first step to getting further inside the numbers is to calculate Pedro's Batting Average on Balls In Play (BABIP) for these two years. Simply put, most pitchers (less knuckleballers and extreme groundball pitchers) do not have a repeatable ability to turn batted balls into outs. The year-to-year variation in BABIP is the result of the quality of the fielders behind the pitcher and blind luck. Calculating BABIP is a fairly straightforward task. All we have to do is divide the number of hits that didn't end up in the stands (H-HR) by the total number of balls in play allowed by the pitcher (Batters Faced-K-BB-HBP-HR). The result for Pedro:
- 1999 - .323 BABIP
- 2000 - .243 BABIP
The reason Pedro's 2000 WHIP is so much lower than 1999 is simply because his fielders turned far more of the batted balls he gave up into outs. Since the majority of the fielders behind him were the same in both years (Stanley, Valentin and Lewis were replaced by Daubach, platoon and Everett), the fact is that Pedro was simply luckier in 2000 than he was in 1999. His decreased WHIP and, since this led to fewer base runners and scoring opportunities, ERA are not signs that he was a better pitcher in 2000. They're signs he was luckier.
Rather than using standard ERA, or even ERA(plus), let's use Fielding Independent Pitching (FIP). This statistic is exactly what it sounds like; a measure of a pitcher's effectiveness, normalized to look like an ERA, that only takes into account the things a pitcher can control. The formula is freely available out there on the internet so I'll refrain from reprinting it here and just provide the numbers.
- 1999 - 1.46 FIP
- 2000 - 2.05 FIP
By this measure, Pedro was about half a run better in 1999 than he was in 2000. His "elevated" hit total in 1999 - a statistic that is more the result of luck than anything to do with Pedro - is what makes the standard numbers appear to show the opposite conclusion. Since FIP allows us to take luck out of the equation we can overcome this illusion. Is FIP the "ultimate statistic?" No it isn't, but given the fact that the rest of Pedro's numbers (BB/9, K/9 etc) are basically the same in both years and the fact that park and league factors essentially cancel out, in this case it is telling.
The great thing about an article like the one Aaron wrote is that it will always inspire debate because there's no way to be "right". Too much of the answer to the question relies on a subjective analysis of the performance recorded by the player. I'm certainly open to the possibility that there are subjective factors not reflected in the numbers that would lead an observer to conclude that Pedro's 2000 season was greater than his 1999 season. I am not, however, open to an objective statistical analysis arriving at any conclusion other than the one at which Aaron arrived. Pedro Martinez' 1999 season was the greatest I have ever witnessed, and it is highly unlikely I will ever witness a greater one.
By Steve Caimano of www.DugoutCentral.com