WBC Study Proves Inconclusive

USA Today has published research that attempts to quantify the dangers of pitchers participating in the World Baseball Classic, based on the pitchers who participated in the 2006 WBC:

USA TODAY analyzed regular-season statistics of every major league pitcher who participated in the spring tournament.

The results show nearly four out of every five pitchers recorded a higher ERA in 2006 than in the previous year (among the 59 who pitched at least 20 innings each year). And more than one in three spent time on the disabled list in ’06.

ERAs went up 5.6% throughout MLB in 2006, but the increase was significantly higher among WBC pitchers: 18.4%.

Of the 59 who pitched 20 or more innings in the majors in 2005 and ’06:

•78% posted a higher ERA in 2006;

•42% registered an ERA that increased by one run or more.

•The average ERA of the WBC group jumped from 3.69 in ’05 to 4.37 the following season, while the ERA throughout all of the majors increased less than a quarter of a run (4.28 to 4.52).

The U.S. team was strongly affected, as 12 of the 13 pitchers recorded higher ERAs; eight went up more than a run.

Sorry, folks, but this is correlation, not causation. Let’s take a closer look, just using the Team USA pitchers from the 2006 roster as an example. I’ve excluded Al Leiter since, as the USA Today writers point out, he did not pitch for a major league team during the 2006 season. Each player’s career ERA is through the end of the 2005 season.

Roger Clemens 1.87 2.30 3.12
Chad Cordero 1.82 3.19 2.36
Brian Fuentes 2.91 3.44 3.67
Todd Jones 2.10 3.91 3.94
Brad Lidge 2.29 5.28 2.71
Gary Majewski 2.93 4.61 3.11
Joe Nathan 2.70 1.58 2.70
Jake Peavy 2.88 4.09 3.33
Scot Shields 2.75 2.87 2.80
Huston Street 1.72 3.31 1.72
Mike Timlin 2.24 4.36 3.50
Dan Wheeler 2.21 2.52 4.18
Dontrelle Willis 2.63 3.87 3.27

See what’s going on here? Not only did every pitcher on the list outperform his career ERA in 2005 (Street was a rookie in 2005, so throw him out), but every single one of them had an ERA under 3.00 in 2005, including seven with sub-2.50 marks. No rational person would have bet on the 40-year-old Timlin having an ERA lower than 2.24 in 2006, and no rational person ever wager money on virtually any player in this decade posting an ERA below 2.00, as Clemens, Street and Cordero would have had to have done to lower their ERAs.

This is simply selection bias and regression to the mean. The Team USA pitchers were probably picked in large part based on their sparkling 2005 ERAs. A pitcher’s ERA is part signal, part noise, a cocktail of his own skill, the fielders behind him, his home park and other factors beyond his control. Pitchers who have an outstanding ERA one year tend to follow it up with a worse ERA the next season, relative to the outstanding year. That’s regression to the mean, which is going to be especially strong for this group of pitchers because 10 of the 13 pitchers were relievers (who of course pitch fewer innings than starters), which means less signal, more noise, more year-to-year volatility and more regression to the mean.

So in 2006, if we toss out Street, seven pitchers had ERAs higher than their career ERAs (not that that’s the best indicator of their true talent level, I know), while five outperformed their career ERAs. It’s a wash. Team USA pitchers were not "strongly affected" by the WBC, but by regression to the mean.

The WBC might have an effect on pitchers, but comparing their 2005 ERAs to their 2006 ERAs doesn’t tell us anything, at least not enough to extrapolate causality.