How do the Sox improve their bullpen?

BOSTON -- The first trade the Red Sox made that bore Theo Epstein's fingerprints was engineered when he was still assistant GM and played a key role in prying left-handed reliever Alan Embree from his former employer, the San Diego Padres.

That was in June 2002, and it was a success, Embree playing a leading role in a bullpen that went to the American League Championship Series in 2003 and won it all in '04.

Epstein spent the better part of the next eight years trying to construct and maintain an effective bullpen, a frequently maddening and quixotic enterprise.

There have been years he has succeeded brilliantly -- such as the two World Series years, when the Sox's bullpen finished fourth (2004) and first (2007) in the American League in bullpen ERA.

There have been years he has failed miserably -- like in 2005, when the Sox finished last in the league in bullpen ERA (5.17). And don't forget the notorious closer-by-committee season, 2003, in which the bullpen was erratic for much of the season (4.87 ERA, 12th) before performing far beyond expectations in the postseason, a transformation that in its own weird way cost Grady Little his job.

And then there was the disintegrating bullpen of 2010, which only the year before had finished second in the league in ERA (3.80) but seemingly turned dysfunctional overnight. Hideki Okajima went from usually reliable to highly flammable, Manny Delcarmen and Ramon Ramirez pitched their way out of town, Jonathan Papelbon's "Shipping Out to Boston" became a signal to seek shelter, and the open auditions of spring training became a talent contest sadly bereft of talent.

Of all the many issues facing the Red Sox this offseason, Epstein answered "trying to fix the bullpen" when asked to list his priorities at his season-ending, state-of-the-team media session.

The question is, how?

There isn't an avenue that Epstein hasn't tried since he became Sox GM: trades, big-name free agents, minor league free agents, Rule 5 draftees, Japanese imports, homegrown talent.

Consider some of his acquisitions:

Trades: Embree, Byung-Hyun Kim, Scott Sauerbeck, Scott Williamson, Brandon Lyon, Mike Myers, Chad Bradford, Mike Remlinger, Guillermo Mota, Bryan Corey, Brendan Donnelly, Eric Gagne, David Aardsma, Ramon Ramirez, Billy Wagner.

Free agents: Chad Fox, Mike Timlin, Ramiro Mendoza, Keith Foulke, Matt Mantei, Rudy Seanez, Julian Tavarez, J.C. Romero, Craig Breslow, Takashi Saito, Joe Nelson, Scott Schoeneweis, Brian Shouse.

Imports from Japan: Okajima, Scott Atchison.

Rule 5 pickups: Javier Lopez (the first time), Lenny DiNardo.

Homegrown products: Papelbon, Delcarmen, Craig Hansen, Daniel Bard.

"What we hope we end up with,'' Epstein said, "are guys who lock down spots. Whether [those are] guys you would feel that way about at this moment, that
remains to be seen. We would like to sign guys with good track records, guys who have a great track record, and we may.

"But I also think to have a good bullpen, guys have to emerge, whether internally or guys who you buy low on or guys who you sign to minor league deals.''

Timlin was a reliever with a proven track record who paid dividends for the Sox. But Mendoza, who was heralded as a coup when Epstein signed him away from the Yankees with a two-year, $6.5 million deal, was a bust, putting up a 6.75 ERA in '03. So were Romero, who wound up being released midway through his only season with the Sox (2007), and Gagne, hailed as the missing link when Epstein acquired him from Texas at the trading deadline that same season.

No other job in baseball comes with such a degree of unpredictability. Ramirez and Lopez, both Sox castoffs, are now playing critical roles for the San Francisco Giants, who just advanced to the NLCS. That's why teams are reluctant to spend a great deal of money on long relievers and setup men.

"If you look at the history of higher-paid, nonclosing relievers on multiyear deals, it's not pretty,'' Epstein said.

The free-agent market will be teeming with relievers this winter, in quantity if not in quality. But there is a catch: To sign a potential Type A free agent like Scott Downs, the Toronto left-hander in whom the Sox had interest at the trading deadline but not for the price the Blue Jays were asking, Boston would have to surrender a first-round draft pick as compensation. Surrendering a No. 1 for a reliever who will turn 35 in spring training is not the Epstein way.

The internal options? Hansen, a No. 1 draft pick promoted to the big leagues the same season, was a spectacular flameout. Bard, a No 1 draft pick who struggled early as a pro, has been a spectacular success, a closer-in-waiting. Left-hander Felix Doubront showed promise, and while he will come to camp next spring competing for a starter's role, he could easily be shifted into the bullpen. Rookies Dustin Richardson and Michael Bowden produced less encouraging results.

Bard doesn't figure to inherit the closer's role next season. While acknowledging Papelbon's shortcomings (eight blown saves), Epstein noted that Papelbon had set an impossibly high standard for himself earlier in his career.

"It's hard to be perfect in the American League East as a closer,'' Epstein said. "He's still a really good closer. He's still going to help us win games. He's got a
little bit of work to do to get back to that elite level where he was at.''

So, the repair job begins, but, come spring training, the uncertainty will remain. It's the nature of the bullpen beast.

Gordon Edes is ESPNBoston.com's Red Sox reporter. He has covered the Red Sox for 12 years and has reported on baseball for 25 years. Ask a question for his next mailbag here.