Who'll be Boston's fifth starter?

In 2011, the Boston Red Sox entered the spring with a solidified rotation featuring five proven starters. It was supposed to be the backbone to the most talented team in franchise history.

Few expected a 4.49 ERA from the rotation or Kyle Weiland to be starting key games down the stretch.

The certainty of a year ago has disappeared. Injuries opened up two slots in the rotation, but salary constraints are forcing the team to go with a scrap-heap approach.

That method worked for the New York Yankees last season (see Bartolo Colon and Freddy Garcia).
Can Boston find similar success?

Daniel Bard is expected to transition from relief, filling one of the two open rotation spots. The final spot likely comes down to a group of five, all of whom come with major question marks.

Let's see what the numbers suggest about the candidates:

Alfredo Aceves

The basic argument against moving Aceves to the rotation is that he's too valuable to a relief staff that will already be without Daniel Bard and Jonathan Papelbon.

Aceves had a 2.03 ERA as a reliever. While the rest of the team collapsed in the final six weeks, Aceves was a rare bright spot with a 1.21 ERA. Why mess with what worked?

There's certainly some validity to that line of thinking.

But another argument against the move isn't as flattering to Aceves. While his effectiveness cannot be questioned, there's reason to believe he won't be able to replicate it.

Aceves is a victim of the same stat that foretold a regression for Daisuke Matsuzaka in 2009 and Clay Buchholz in 2011.
While he posted a 2.61 ERA, his xFIP (expected fielding independent pitching) was 4.77. In other words, Aceves didn't pitch as well as his ERA indicated.

xFIP focuses on the outcomes over which a pitcher has direct control: strikeouts and walks. With 42 walks in 114 innings, that's not an area where Aceves shined.

But he's hurt even more by perceived luck in his home run rate. Among those who faced at least 400 batters, Aceves had the fifth highest fly ball percentage (47.5). Of the 160 fly balls he gave up, only eight were home runs. Tim Stauffer (18), Joel Pineiro (14) and Homer Bailey (18) all allowed significantly more home runs on about the same number of flies.

The theory is that pitchers have little control over where fly balls land. On average, 9.7 percent of fly balls were home runs. For Aceves, that was just 5.3 percent.

Those factors contributed to an xFIP that was far higher than his ERA. So why should anyone care? History indicates that the stat holds value as a predictor of future performance.

As noted in a similar analysis of Buchholz going into 2011, when an xFIP is at least 1.00 higher than an ERA, the next season tends to be ugly.

There have been 46 such seasons by qualifying starters since 2002. In 44 of those cases, the pitcher saw a rise in ERA in the following season. On average, the ERA increased by more than a point.

Certainly a 3.61 ERA out of Aceves would be welcome in the rotation. But that doesn't factor in the inevitable rise in ERA that comes with starting.

It also ignores just how big the difference was between his ERA and xFIP. Among those to throw at least 100 innings in 2011, Aceves' 2.16 difference between the two numbers was easily the biggest in the majors. Jeremy Hellickson (1.77) is next on the list.

A move to the rotation would simply accelerate Aceves' fall back to earth.

Felix Doubront

Barring a terrible spring, Doubront remains a good bet to break camp with the big club. Still just 24, his star has faded a bit after an uneven 2011. But without any options remaining, he's still too promising to lose.

But can Doubront snag the final rotation spot?

Doubront went into last season as arguably among the top five prospects in Boston's system. At 22, he'd posted a 2.81 ERA between Pawtucket and Portland, while acquitting himself well in 25 innings with Boston.

Unfortunately, 2011 felt like a lost season with injuries to his elbow, forearm, groin and hamstring. He struggled to a 4.22 ERA in Pawtucket and his limited stint in Boston was a disaster. Doubront fell to 16th in Baseball America's Red Sox prospect rankings.

Too much was expected too soon. Yet Doubront is the high-upside candidate of the five main options for the final rotation slot.

He's the same age Ivan Nova was last season. Jon Lester was also 24 in his first full season in the majors. Clay Buchholz similarly struggled at age 23 before rebounding in 2009.

The key for Doubront will be his performance against lefties. It's an issue he's struggled with on and off throughout his professional career.

In Pawtucket, lefties hit .309 against Doubront. Things got worse in Boston, where lefties hit .412, as he struggled with control and walked eight in 10 1/3 innings. Alarming numbers, given that Doubront himself is a lefty.

Vicente Padilla

Appearing on WEEI in February, Padilla described his reason for signing with the Red Sox.

"Other teams needed me as a reliever," he said.

For 10 seasons, Padilla was used almost exclusively as a starter, including four 14-win seasons. Among all the candidates, he has the best historic credentials.

But he's coming off an injury-ravaged year. Elbow surgery caused him to miss the first two weeks of the season. After a month in the bullpen, he landed back on the DL with a forearm injury. Then came surgery in June on a bulging disk in his neck that forced Padilla to miss the rest of the season.

After a lost season, it's hard to know what to expect.

Nothing indicates that Boston will get the All-Star from 2002 who posted a 3.28 ERA. Since 2004, Padilla has posted a 4.68 ERA and 1.42 WHIP. Those aren't promising numbers for a 34-year-old pitcher coming off major surgery.

However, Padilla is a reasonable stopgap option. He makes sense until Daisuke Matsuzaka returns, a young pitcher emerges or a trade is made.

It's all about catching Padilla in the right month. He went 4-0 in May 2008. The following May saw a 1.57 ERA. In July 2010, Padilla posted a 1.42 ERA in over 30 innings.

In short, he's capable of being a back-end starter, just not for a whole season. That makes him a prime candidate for the role recently held by Tim Wakefield. If nothing else, Padilla's 50 mph eephus pitch will bring back memories of the slow knuckler.

Aaron Cook

Perhaps the easiest decision of the offseason was made when the Colorado Rockies declined Cook's $11 million option for 2012.

That's the contract he signed after he went 16-9 with a 3.96 ERA in 2008. After a decent 2009, Cook hasn't been able to get healthy.

He's consistently battled shoulder issues, which have seen his velocity take a big hit. At his peak in 2008, Cook's fastball averaged 91.8 mph and his sinker averaged 91.1 mph. Last season, both were down to 88.

Over the last two seasons, he's had a variety of other maladies. In 2010, Cook sprained a toe in batting practice. That was followed by a fractured fibula suffered on a comebacker hit by Joey Votto. He also suffered from shoulder tendinitis and then broke a finger on a door during spring training. After debuting in June, he posted a career-worst 6.03 ERA. Among those who threw at least 90 innings, only one pitcher had a higher ERA. That, of course, was John Lackey (6.41).

Cook's problem (other than the injuries) is that he has so little room for error. Even in his higher-velocity days, he's never been a strikeout pitcher. In fact, his 3.83 strikeouts per nine innings are the fewest among active pitchers with at least 500 innings.

The last pitchers with a lower strikeout rate and at least 1,250 career innings? Bob Forsch and Bob Stanley, both of whom retired in 1989.

Cook is owed $1.5 million if he makes the roster, which seems like a long shot based on his downward trajectory.

Andrew Miller

Miller's candidacy has already taken a minor hit. He was scratched from a scheduled Thursday appearance due to elbow stiffness.

But a far bigger red flag is raised by his performance last season.

Miller stumbled to a 5.54 ERA while walking 41 batters in 65 innings. Opponents hit .302 with a .398 on-base percentage. Among those who pitched at least 50 innings last season, that's the third-highest on-base percentage.

The biggest issue came against right-handed batters, who posted a .548 on-base percentage against Miller. He only threw 50 percent strikes against them. That's a recipe for disaster at Fenway, where he had a 9.00 ERA in 21 innings.

It culminated in perhaps the ugliest September performance on a team that was full of them. Miller's 11.70 ERA was the highest for a Red Sox pitcher in September (minimum 10 innings) in the post-World War II era. At 10.64, Bard wasn't too far behind.

Kyle Weiland, who essentially took Miller's spot in the rotation in September, posted an 8.53 ERA in four starts. As bad as that was, it was still an improvement over Miller.

So why, with all these ugly numbers, is Miller even a candidate? It's hard to ignore his arm.

It's the arm that made him -- and not teammate Daniel Bard -- the ace of North Carolina's staff in 2006. Miller was selected one spot in front of Clayton Kershaw that year. Going into 2007, his promise made him a top-10 prospect.

That arm is still there. Miller was one of only 12 lefties to touch 98 mph last season, a group that includes CC Sabathia and David Price.

Just how long can the Red Sox keep hoping? Like Doubront, Miller is out of options. For his career in Boston, it's now or never.

Jeremy Lundblad is a senior researcher with ESPN Stats & Information. He provides statistical analysis for ESPNBoston.com.