Giving Up The Long Ballby Tom Meagher
March 31, 2005
FIP - an invention of Tangotiger and one of the stats we keep track of here at The Hardball Times - is a great quick way to gauge some of the luck involved in a pitcher's earned run average. FIP, inspired in part by Voros McCracken's DIPS work, looks at a player's walks, strikeouts, home runs, and innings pitched to quantify a pitcher's performance. Tangotiger first used FIP as a way of calculating defense-independent runs saved relative to average, but you can add a league normalizing factor to FIP to make it approximate ERA pretty well.
The reason why FIP is a very good quick and dirty estimation of a pitcher's talent level is because its components are less subject to random fluctuation than the other things that go into a pitcher's ERA like fielding and the timing of offensive events. Recently, Dan Fox took a look at the correlations for various pitching categories between 2003 and 2004. The correlation for batting average on balls in play is low while the correlation for strikeouts and walks is quite high. ERA has an okay correlation, but Fox found that a player's component ERA or DIPS ERA in 2003 correlated quite a bit better with 2004 ERA than ERA itself, with DIPS doing the best job of the three.
Because a player's defense-independent pitching numbers correlate so well with future ERA, there is a tendency (which has certainly had an effect on me) to look at a player's FIP in one season or even several seasons and use it as a pretty strong index of a player's performance in future seasons.
There is, however, a bit of a weak link in the predictive prowess of FIP. That weak link is a pitcher's ability to prevent home runs. In the correlations Dan did, a pitcher's home run rate correlated much better than batting average on balls in play, but nowhere near as well as strikeout and walk rates. This is not surprising, since we'd expect home runs allowed to have a pretty low correlation as there are so few of them; last year Jamie Moyer led the major leagues with 44 home runs allowed. While home run rate tends to stay in the same ballpark for any given pitcher, a player's home run rate in any season or even over a few seasons has much more room for substantial deviation from that player's true talent level than K and BB numbers do.
Nate Silver of Baseball Prospectus has argued (subscription required) that a pitcher's groundball rate may in fact be a better predictor of home run rate in future seasons. A pitcher's groundball rate does not fluctuate too much from year to year, and pitchers with low groundball rates give up home runs much less frequently than fly ball pitchers. In fact, the reason that home runs correlate fairly well from year to year may be partially that they're a proxy for a pitcher's groundball/flyball rate.
So before you look at a player's FIP as a proxy for future ERA, you might want to consider whether that player's home run rate is unsustainable given the type of groundball/flyball pitcher they are. Of course, this lesson doesn't just apply to FIP. Even if you want to look at a player's 2004 ERA to approximate their 2005 ERA, you should keep in mind that the pitcher may have allowed home runs with fortuitous frequency.
Here are some players whose home run rates are probably set for a rise. Each one ranked in the top third of pitchers in non-infield fly balls per batted ball and in the bottom third for home runs per batted ball:
2004: 137 IP, 4.59 ERA, 4.39 FIP, 0.85 HR/9
There's some reason to believe Bedard may be an outlier. In 335.1 minor league innings, Bedard only gave up 9 home runs. While that's certainly impressive, he also typically accomplished that against younger competition and had very high strikeout rates; it was difficult to put the ball in play against him, much less put it over the fences. Now that he's in the major leagues and probably can't manage a strikeout per inning, he's probably set to see average or worse home run rates.
2004: 72.1 IP, 1.62 ERA, 2.40 FIP, 0.37 HR/9
As angry as it will make Aaron, Nathan is probably set for some decline. Entering 2004 with a career major league HR/9 of 1.21 in 266.1 innings, Nathan allowed only 3 HR despite being among the most extreme fly ball pitchers in baseball. While Nathan will almost certainly see a HR/9 lower than 1.21 because he strikes out a lot of batters now that he's a full-time reliever, it's unlikely that he can only yield one home run every 24 innings.
2004: 53.2 IP, 2.18 ERA, 2.69 FIP, 0.17 HR/9
After a disastrous 2003 season, Carrara eventually made his way back to the Dodgers in 2004. He was lights out. However, Carrara certainly had his share of luck in giving up only one home run over 53.2 innings. That was quite a reversal from Carrara's career major league 1.77 HR/9.
2004: 69.1 IP, 3.38 ERA, 3.51 FIP, 0.39 HR/9
Bottalico had a solid season with the Mets, and the Brewers picked him up this offseason for a mere $800,000. While Bottalico has done a respectable job avoiding the long ball throughout his career, he has done so by striking out a ton of batters. With his K rate in decline, don't expect anything approaching what he did last season. If you've picked up Botallico for your fantasy team in hopes that he'd usurp the closer role from Mike Adams, you may have too many problems for this to concern you, but it's there.
2004: 120.2 IP, 4.77 ERA, 5.11 FIP, 0.90 HR/9
The notion that things could get worse for Brazelton may be disturbing enough, so I'll refrain from much comment. On the bright side, He has had outstanding home run numbers in his minor league career, though the majority of those innings were compiled in the Southern League. Still, it's impressive that he only allowed one home run in 80.1 innings pitching for Durham.
* * *
Now, let's take a look at some players who might be due for a rebound. Each one ranked in the bottom third in non-infield fly balls per batted ball and in the top third for home runs per batted ball:
2004: 212.2 IP, 4.02 ERA, 4.38 FIP, 1.48 HR/9
At the tail end of a career with one of the most impressive home run rates in history, Maddux had never allowed one home run or more per nine innings pitched in any season in his major league career entering 2004. Leaving Atlanta and Leo Mazzone behind, Maddux' performance took a hit in Chicago. While Maddux won't reach the levels of dominance he achieved at his peak, his performance in 2004 was pretty much the same as it had been in his last three years in Atlanta, as his K, BB, and G/F rates all were roughly what they had been in Atlanta - in fact, his K/9 was his best since 2001 and above his career average. The only thing that changed was his ballooned home run total. While blaming the Friendly Confines - which helped a lot of balls over the fence last season - is tempting, Maddux actually allowed 19 on the road and 16 at home despite pitching a lot more at home.
2004: 68.1 IP, 3.95 ERA, 4.84 FIP, 1.58 HR/9
Graves gets a lot of flak for being a perhaps mediocre pitcher who gets unfair praise for amassing 41 saves last season. One common slam is that, while he put up a 3.95 ERA, Graves had an ugly 5.14 run average. However, Graves has reason to expect a rebound. After a career of success groundballing it in relief - with the requisite low home run tallies - Graves was converted to a starter in 2003. The role didn't suit him well, and he put up a 5.33 ERA and a 3.20 K/9. His HR/9 shot up to 1.60. When returned to relief in 2004, Graves regained his ability to fan batters and displayed excellent control. That was masked, though, by 12 home runs. Expect Graves to bounce back (somewhat).
2004: 176 IP, 5.52 ERA, 5.20 FIP, 1.59 HR/9
Myers might be a decent candidate for a breakout season. A groundballer with a solid K rate and improving control, Myers season was bloodied by 31 home runs. The Phillies used to play in a Veteran stadium but moved to rookie Citizen's Bank Park last season, and that probably played some role in Myers increased rate. Still, Myers gave up 15 home runs on the road in 102.2 IP, and he's unlikely to sustain a home run rate that high.
2004: 202 IP, 4.72 ERA, 4.94 FIP, 1.56 HR/9
Underrated on the basis of one season. While his shrinking K rate is concerning, it didn't shrink by that much, and his BB rate was better than his career average. The only big change was a fat HR total after years of skinny ones. With the Cardinals' infield defense, expect a solid year from Morris in 2005. He probably won't match the great performance of 2001-2002, but he can do what he did from '97-'00 and in '03 again.
2004: 196.2 IP, 4.90 ERA, 4.76 FIP, 1.37 HR/9
Robertson isn't exactly young or promising, but his home run rate was a lot worse than you'd expect from a groundball guy in commodious Comerica Park. He had an undistinguished minor league career, so I doubt that he can match the 2.35 K/BB ratio he boasted last season. If he can, though, he's a nice sleeper.
Tom Meagher jots occasional thoughts on baseball and the Dodgers at The Fifth Outfielder and can be reached via email.