Growing up as a Minnesota Twins fan and teenager in the late 80s and early 90s, it was hard to miss those great Kansas City Royals teams with George Brett, Frank White, Hal McRae, Bret Saberhagen, and submarine throwing closer Dan Quisenberry. Other than the fact that, much like the Twins, the Royals always seemed to perform at a level far above the sum of their parts, the curious case of Bret Saberhagen intrigued the sports world. For a solid 10 years with the Royals and New York Mets, from 1984 to 1993, without a single exception, Saberhagen managed to lower his ERA over the previous year in each and every odd numbered season, and his ERA increased in each and every even numbered season. In addition, over the same time period Saberhagen never posted a winning record in even numbered years, and he never posted a losing record in odd numbered years. Quite the Dr. Jekyll and Mr. Hyde situation for Mr. Saberhagen. While this is not an occurrence with astronomical odds against, similar to flipping a coin 10 times and coming up "heads" each time, 1 in 1,024 odds, for a pitcher of Saberhagen's profile, Cy Young Awards (both in odd numbered seasons, of course) the differences were striking.
The Twins have a similar even-odd situation in Nick Punto. Since his first meaningful season in the majors (2003 with Philadelphia), Punto's batting average and wOBA have been higher in every even numbered season and lower in every odd numbered season. While his seven seasons haven't quite matched Saberhagen's 10, the differences have been quite striking for the Twins infielder. After the jump, I'll compare and contrast the two cases, as well as attempt to identify some root causes for this effect on Punto's performance.
First, the raw numbers. For Bret Saberhagen:
And for Nick Punto:
Or for those of you who are numerically challenged, pictures are worth a thousand words. For some reason, it was more difficult to find a "good" Punto picture...
Than a bad one...
The much more difficult question is why there are such regular fluctuations in performance? Do the peripherals show the same peaks and valleys as the resulting ERA and AVG? Or does an investigation of peripherals point to a different root cause?
First, Saberhagen. His Fielding Independent Pitching (FIP) does not correlate as well with his ERA during the given time period. For example, in 1987, his ERA dropped from 4.15 to 3.36 while his FIP jumped from 3.18 to 3.66. Basically, Saberhagen appeared to benefit from a low 68% LOB% in 1986, fueling an ERA lower than his peripherals. However, looking at individual peripherals we see the same even-odd effect for BB/9 and BABIP, with single exceptions for 1987 and 1984, respectively. Then again, while this is a bit of a chicken-end-the-egg situation, the entire Royals team showed the same even-odd behavior with W-L record, with the single exception of 1988 (up from 83 to 84 wins), with an average year to year difference of over 10 wins! Whether Saberhagen's performance worsened in even years due to a poorer team behind him, or the team struggling as its Ace struggles, it's difficult to determine due to lack of good defensive metrics from this time period
But this isn't "Bret Saberhagen Day". Let's talk about Nick Punto. Considering he's a singles hitter with a career isolated slugging (ISO) of .076, we would expect his batting average to be driven mostly by BABIP and SO%. Which it does. Both Punto's BABIP and strikeout rate have improved in every even numbered season, and declined in every odd numbered season. I'll get into BABIP in a little bit, but his strikeout rate is interesting. During Punto's two best offensive seasons (2006 and 2008), he struck out at rates (14.8%, 16.9%) well below his career 18.7% rate. And in other years he was above the career average. Why is this? For the most part, we don't see similar behavior in Punto's walk rate, but looking at Fangraphs' per pitch performance, it appears that he has struggled mightily against the fastball in his "poor" years (-1.65, -0.92 wFB/C in '07 and '09) and more successful in his "better" years (-0.06, -0.35 in '06 and '08). This would indicate a potential issue catching up to the fastball in some years if it wasn't for similar performance against curve balls and change ups, pretty much across the board.
From a batted ball standpoint, Punto has been more likely to hit more groundballs, and average of 4% more often in '05, '07 and '09 than in '06 and '08. And has has been less likely to hit line drives by a similar 4% margin. So basically, instead of hitting line drives in his "off years", he hits more ground balls. And at a career 47.9% ground ball rate, he already hits a ton of ground balls. I also expected to look at infield popups and see a similar increase in off years. But for some reason, Punto did just the opposite. During his solid 2008 season (.284 AVG), Punto popped out 21 times (22.1% IFFB). During his 2009 "off year", he popped out only 8 times (8.6%).
For what it's worth, Punto's fielding (as measured by UZR) has not correlated with his offensive struggles. But his offense has clearly fluctuated pretty wildly from season to season. Fortunately for the Twins, 2010 is an even numbered season, so if the pattern holds, we will see a solid offensive season from Nick Punto this year. Then again, perhaps this is all a coincidence and Punto will repeat his 2007 black hole season. In any case, this is why we play the games, and watch Punto slide headlong into first base...