Sociologists sometimes refer to a phenomenon called "The Pendulum Effect," which refers to the tendency of folks on opposite sides of an argument to make increasingly extreme statements until we reach a point where all Democrats are baby-slaying, liberty-hating communists and all Republicans are Muslim-bashing, sexually-repressed masochists. The effect is not limited to politics; it exists in all societal groups, including baseball pundits, where the schism du jour falls between the new school "Stat heads," who rely heavily on a number of advanced metrics whenevaluating and predicting performance, and the old school "Purists," who prefer things like ERA and eye-witness experience to things like OPS and xFIP. Witness the reaction to the NL Cy Young Ballot cast by ESPN.com columnist Keith Law, a former writer for Baseball Prospectus who voted Javier Vazquez No. 2 ahead of more conventional candidates like Adam Wainwright and Chris Carpenter. In choosing Vazquez, Law relied on metrics like xFIP and WAR, the second of which -- in addition to serving as an acronym for Wins Above Replacement (player) -- provides a pretty accurate summation of the conflict between the two parties. The Purists think the Stat Heads are taking a sport that features numerous incalculable subtleties and turning it into a glorified game of Strat-o-Matic. The Stat Heads think the Purists are dinosaurs blinded by their infatuation with Romance and their freqent trips to the seventh-inning hot dog table in the press box dining room. It's like Grumpy Old Men, except Jack Lemmon and Walter Matthau have Hall of Fame votes.
As often happens in world conquest, a seemingly harmless entity finds itself an unwitting linchpin, subjected to each side's repeated attempts to claim it as its own. It is kind of like two kids fighting over a Gumby doll: Both sides grab one arm and pull as hard as they can in the opposite direction until the object in question looks nothing like reality and more like Manute Bol with elephantitis of the arms.
In this case, Gumby is J.A. Happ, the 26-year-old left-handed starter who last season went 12-4 with a 2.93 ERA for the Phillies. Happ finished second in the National League Rookie of the Year balloting, which is a pretty good measure of how the Purists view Happ, since the Baseball Writers Association of America votes on the awards, and the BBWAA is still largely composed of Purists. This, of course, prompted guffaws from the Stat Heads, who all season had been warning that Happ was not as good of a pitcher as his ERA or record (12-4) indicated.
Law -- the aforementioned ESPN.com columnist who, by the way, maintains a pretty interesting blog that combines literature and food criticism with baseball -- has led the charge since Happ started garnering pub around the All-Star Break.
After watching Happ allow four runs in a loss to the Giants in early September, Law labelled Happ "a back-end starter, maybe a No. 4 if you really believe in his deception and his feel for pitching, otherwise a very good No. 5."
Judging by my email, many Phillies fans disagreed with this assessment. And, up to this point, both sides had legitimate arguments.
Stat Heads pointed to things like the disparity between Happ's ERA and his FIP, short for "Fielding Independent Pitching," a metric that attempts to measure things for which a pitcher is reponsible (home runs, walks, strikeouts) while attempting to factor out the defense behind him. Happ's FIP was 4.49, and his xFIP -- a slightly adjusted version of FIP -- was 4.33.
They pointed to his abnormal success with Runners in Scoring Position -- at the time Law labelled Happ a good No. 5 starter, he had held opponents to a ridiculous .120/.225/.157 line with RISP. They pointed to the fact that he stranded 85.2 percent of his runners, well above what is considered "normal."
Many Purists who had watched Happ pitch regularly argued that many of the aforementioned statistics were the result of his unique poise on the mound, as well as the deception on his fastball.
At this point, the two sides could have agreed to disagree and compare notes in a couple of seasons, when Happ has a couple more seasons under his belt.
But the Pendulum was already in motion.
Law left Happ off his hypothetical Rookie of the Year ballot. Phillies fans felt persecuted.
Which brings us to this past week, when Law labelled Happ's 2009 season "a raging fluke driven by an unsustainable performance with men in scoring position that was about luck, not skill."
This prompted one user on the always-entertaining Phillies message board at PhilaPhans.com to label Law an "ass."
So which is it: Is Happ headed for an epic slump after a raging fluke of a season? Or are Keith Law and Co. asses?
In the landscape of baseball projection and evaluation, I consider myself to be a Moderate. I know that is not a popular concept in modern day America, but I suggest that it is the proper stance to take.
My moderate's creed:
- I believe in advanced metrics.
- I believe that there is no infallible metric.
- I believe that metrics are valuable indicators.
- I believe there are aspects of baseball - pitch sequence, deception, poise, intelligence - that are best evaluated through personal experience.
- I believe that personal experience can be impacted by human biases and emotions.
- I believe that the best evaluations are formulated when metrics are combined with personal experience.
In taking a look at the Curious Case of J.A. Happ, I suggest we work backwards. We'll start with Law's contention that Happ's 2009 campaign was a "raging fluke" and that he is "headed for a big regression in 2010" and that "a sophomore slump could really hurt the defending NL Champs." And we'll progress toward Happ being a frontrunner for the 2010 NL Cy Young.
1) Statistical projection is like Black Jack. You aren't betting on what is GOING to happen. You are betting on what is LIKELIEST to happen, based on your evaluation of the table. The more scientific your system, the better your chances of being correct. But even card-counters don't know exactly what the Dealer is going to turn over.
And I don't need to count cards, or xFIPs, to tell you that Happ is likely in line for a statistical regression. From 1995-2008, 13 pitchers have posted an ERA of 3.00 or below (min. 150 IP) in one of their first three seasons in the majors. One of those pitchers, Alan Benes, missed the entire next season.
Of the remaining 12 pitchers, eight saw their ERA rise by at least 0.75, seven by at least 1.00, and three by at least 2.00. Eight of the 12 saw their WHIP rise by at least .200.
Only four pitchers posted ERAs and WHIPs that were close to their previous year's: Tim Lincecum (2.48 ERA, 1.047 WHIP in 2009, 2.62 ERA, 1.172 WHIP in 2008), Jake Peavy (2.27 ERA, 1.196 WHIP in 2004, 2.88 ERA, 1.044 WHIP in 2005), Barry Zito (2.75 ERA, 1.134 WHIP in 2002, 3.30 ERA, 1.183 WHIP in 2003) and Hideo Nomo (2.54 ERA, 1.056 WHIP in 1995, 3.19 ERA, 1.161 WHIP in 1996).
2) The question, of course, is what counts as a "major regression?" Is Andy Pettitte going from a 2.88 ERA and 1.240 WHIP in 1997 to a 4.24 ERA and 1.447 WHIP in 1998 "major." Or is "major" more along the lines of Kyle Kendrick, who posted a 3.87 ERA and 1.273 WHIP in 20 starts in 2007, then saw those numbers balloon to 5.49 and 1.612 in 2008?
Let's call a Kendrick-like regression "major," and a Pettitte-like regression "moderate."
3) For the time being, let's assume that J.A. Happ is in line for a regression that is at least "moderate." Let's assume that he doesn't maintain his .181 batting average on balls in play with runners in scoring position, and therefore does not maintain his .158 average with runners in scoring position, and therefore does not strand 85.2 percent of his baserunners.
Let's assume that The Hardball Times, a renowned stats journal, is correct when it says that Happ's LOB percentage is a product of significant luck, and that a "normal" figure for him last season would have been 70.4 percent. If Happ's LOB% was 70.4 this season instead of 85.2, he would have allowed 27 more runs, and finished the season with an ERA of roughly 4.45 (No, I'm not going to show my work, but you can email me if you really want the details).
4) So, would a 4.45 ERA out of J.A. Happ spell trouble for the Phillies? Consider: Only one NL team, the Rockies, had five pitchers who threw at least 150 innings and posted an ERA of 4.45 or under last season. Three more had four pitchers. So Happ's production - at least according to the most basic measurements - would have been better than that of the fourth starter on 12 of 16 NL teams, including the Dodgers, whom the Phillies defeated in the NLCS.
So even if Happ's ERA was what things like FIP and xFIP suggest it should have been, he still would have been one of the more productive No. 4 starters in the NL.
5) Bill James, the Grand Pooh-Bah of projection, forecasts the following figures for the Phillies' rotation this season:
1. Roy Halladay 17-10, 3.23 ERA, 240.0 IP
2. Cole Hamels 15-9, 3.43 ERA, 210.0 IP
3. Joe Blanton 13-11, 4.06 ERA, 213.0 IP
4. J.A. Happ 10-11, 4.31 ERA, 188.0 IP
Since Wild Card play began in 1995, eight NL teams have finished the year with at least four players tallying more than 185 innings pitched. Two of those teams appeared in the World Series. Four others appeared in the NLCS. Seven of the eight made the playoffs. The lone exception? The 2003 Phillies, who had Kevin Millwood, Randy Wolf, Brett Myers, and Vicente Padilla. . .and David Bell hitting .195 at third base and Pat Burrell hitting .209 in left.
Since 1995, only two NL teams have finished the season with three pitchers who had at least 210 innings pitched and a fourth who pitched at least 185: The 1997 Braves and the 2003 Cubs, both of whom lost in the NLCS.
So even if J.A. Happ does regress to the norm, he still will provide pretty good production out of the fourth spot.
James also projects Happ have to have a 4.43 FIP and 1.38 WHIP. Since we used Pettitte as an example earlier, his 1998 FIP was 4.28 and his WHIP was 1.45.
He was the No. 4 starter on the Yankees that season, when they swept the Padres in the World Series.
My point? Even if Happ isn't the wunderkind we saw last year, he can still be a valuable piece of this rotation.
A little later, though, we'll take a look at some caveats to consider before you downgrade Happ too heavily.