SIERA: Sleepers and Busts

Earned Run Average can't be trusted in small samples. It has little predictive ability over short spans of time, given the number of things that can influence it—defense, park effects, batting average on balls in play, or just plain old dumb luck. When enough innings are looked at together, ERA has its merits, but with less available data, run estimators—like SIERA—are the better bet for year-to-year predictions.

That means that SIERA is useful for seeing which players over- or underperformed in terms of their ERA. (No, Matt Cain was not one of them.) Using this tool, we can predict some likely 2011 sleepers, as well as some pitchers who will miss the good old days when extra luck and quality defenses saved them inning after inning.

SIERA Darlings

Josh Beckett, Boston Red Sox: 5.78 ERA, 3.84 SIERA

Beckett signed a lucrative long-term extension with Boston at the start of the 2010 season, and subsequently pitched like a guy looking to get released. His 5.78 ERA was a career worst, and he posted his lowest K/BB ratio since his first year in Boston. The changes to the Boston defense didn't help him much, either—thanks to a flood of injuries that plagued the club throughout the year, the Sox rarely had their optimal defensive unit on the field behind Beckett.

The difference between Beckett's ERA and SIERA was the largest of any hurler with at least 100 innings pitched. There are some underlying issues with Beckett's 2010 that may not yet be resolved, though: his back was an injury concern, and the pain he experienced kept him from commanding his normally excellent curveball. That led to more fastballs and more counts that started out in the batter's favor—the second-highest home runs per nine innings rate of Beckett's career didn't come out of nowhere.

If his back is healthy, Beckett will be one of the American League's top starters once again. From 2007 through 2009, the mostly healthy Beckett threw 587 1/3 innings with a 3.71 ERA and a 4.4 K/BB ratio—remember, while ERA doesn't work well in small samples, with more data, it paints an accurate picture. The 2007-2009 version of Beckett is who he is—as long as he's feeling well on game day.

Brandon Morrow, Toronto Blue Jays: 4.49 ERA, 3.15 SIERA

Morrow was dealt to the Blue Jays before the season began, in exchange for Brandon League. Even without looking at his SIERA, this was one of the most lopsided swaps of the season. Morrow has nowhere to go but up: in his first year as a full-time starter, he logged 146 1/3 innings while striking out almost 11 batters per nine. His walk rate—both in 2010, and for his career—is worse than the league average, but thanks to his excellent punchout rates, he was well above average in K/BB rate. A .344 BABIP could be to blame for much of the difference between his ERA and SIERA; someone with his stuff shouldn't be that hittable, so chances are good we will see a vastly improved Morrow in 2011.

Aaron Harang, San Diego Padres: 5.32 ERA, 4.44 SIERA

Harang's days as an impact starting pitcher are over, but, despite what his ERA says, he's not quite done yet, either. Harang admits that his shoulder never quite recovered from an eight-day stretch in 2008 when he threw nearly 400 pitches (including warmups)  thanks to a 103-pitch start that was followed up three days later by a 63-pitch relief appearance. His mechanics were altered to compensate for the pain, which is never a positive for a pitcher. In addition to the bum shoulder, Harang had a .338 BABIP and the Great American Ballpark to contend with in 2010—considering the Reds had an excellent defense, that .338 BABIP is part poor luck, part Harang's health issues.

He's in San Diego this year, though, in the friendliest pitchers' park in the majors, and with his old mechanics intact after a winter of work. Those changes should improve his production, and that is before you take into account that SIERA already expects him to pitch better.

Dan Haren, Los Angeles Angels of Anaheim: 4.60 ERA (Arizona), 3.16 SIERA

Once Haren came to the American League, his fortunes changed. With a tolerable defensive unit behind him and home starts in a park that wasn't built for the sole purpose of making pitchers fail, he was able to post a 2.87 ERA. For the season, though, things look a bit worse: Haren's overall ERA dropped to just 3.91 due to his time in L'Anaheim, which makes him appear to be more of a Josh Beckett type, rather than one of the very best pitchers in the majors.

A full season with the Halos' defense and neutral home park will do Haren a world of good. Expecting another ERA under three is asking a bit much, but PECOTA doesn't have him pegged very far from that: Haren is projected to post a 3.11 ERA with his typically excellent K/BB ratio intact.

Regression is a dish best served cold –Old Jamesian proverb

Brian Duensing, Minnesota Twins: 2.62 ERA, 4.22 SIERA

Duensing threw 130 2/3 innings for the Twins in 2010, with more than half of those coming as a starting pitcher. His ERA was one of the biggest surprises of the 2010 season, given that Duensing strikes out a below-average number of hitters (5.4 per nine) and, while his control is good, his K/BB was 2.4—solid, but not the kind of thing that nets a pitcher an ERA like that. As a groundball pitcher, he's expected to put up an ERA that surpasses his SIERA by a bit—there are more unearned runs for groundball pitchers due to the additional balls that infielders need to field, so Run Average is preferred to ERA when evaluating their performance—but a 1.6-run difference goes well beyond the expected variation. Duensing has 214 2/3 innings in his major league career with an ERA of 3.02—a full season in the Twins' rotation this year should correct that number for the worse.

R.A. Dickey, New York Mets: 2.84 ERA, 4.04 SIERA

The knuckler's ERA made him look like the ace of the staff, but in reality, Dickey was very lucky. Citi Field helps him avoid costly home run mistakes on knucklers that don't knuckle, but a 2.2 walks per nine rate from someone twirling the butterfly pitch is one of those things that can't be sustained from year to year. Tim Wakefield has ventured into that territory a few times in his entire career, but his career walk rate is 3.4 per nine, and he's gone to the other extreme about as often—and while Dickey had a great 2010, he doesn't have quite the same knuckleball resume as Wakefield. This isn't to say Dickey will be bad in 2011—if his 2011 ERA matches his 2011 SIERA, Dickey will have loads of value to the Mets—but with his below-average strikeout rates, a spotty history of control and the ever-present long-ball issue that comes with the knuckleball territory, it would be foolhardy to expect him to come close to replicating 2010.

Livan Hernandez, Washington Nationals: 3.66 ERA, 4.91 SIERA

Hernandez doesn't induce groundballs. His BABIP was close to the league average at .287. He strikes out two full batters per nine less than the league average for starting pitchers. To sum it up, Livan Hernandez's 2010 season was not supposed to happen. The list of things he did genuinely well in 2010 was a short one: he threw over 200 innings, which has value. He can attempt to repeat that feat in 2011, but the quality of those innings is sure to diminish.

Washington's 2010 defense was below the league average, as they ranked 22nd in Defensive Efficiency. This problem—lucky for Hernandez—didn't affect him. The chances of that luck repeating, given Hernandez's spotty stuff, are about as slim as they come. With more runners on base, he'll be forced to throw strikes more often, and hitters are sure to catch up to those offerings.

Jon Garland, Los Angeles Dodgers: 3.47 ERA, 4.45 SIERA

Garland exploited Petco Park as well as anyone could. He attacked hitters up in the zone without fear of retribution, as he knew his environment would save him from the homers that would have resulted in many other stadiums. This helped him set a career high in both strikeout rate and total punchouts. He  also worked down in the zone often, attempting to induce grounders, knowing that the extra walks he would give up were not as dangerous in a place like Petco, where extra-base hits are hard to come by. SIERA accounts for a lot of things, but the quality of a defense and park effects are not on that list—Garland deserved his 3.47 ERA, but only in the context it was in. Outside of Petco, that SIERA stands as his actual body of work.

Now he moves back to Los Angeles,to a park that marginally slows home run production, but that looks like Coors Field compared to Petco. The chances of another ERA in the mid-three range are low—PECOTA has Garland down for an ERA of 4.29, which is closer to the league average pitcher he is, rather than the ace that Petco allowed him to moonlight as.

A version of this story originally appeared on ESPN Insider .