Keith Foulke is finally starting to get some attention, but it's still far less than he deserves.

Now that he's been the White Sox' full-time closer for two years, Foulke's name has slowly begun to appear in discussions of the game's best relievers, even if he's still often relegated to the second tier behind Mariano Rivera, Robb Nen and Trevor Hoffman. In fact, Foulke's numbers show that he is one of baseball's elite relievers -- arguably the very best in the game -- and he was even before the coveted "Sv" abbreviation started showing up next to his name in the box scores.

	Keith Foulke Relief Pitcher Chicago White Sox Profile
	1999-2001 STATISTICS IP W-L Sv H K ERA 274.1 10-13 85/97 195 289 2.49

Foulke's relative obscurity comes partly from the fact that some of his best work was done in the thankless role of set-up man. In 1999, he had one of the best run-prevention seasons for a full-time reliever in recent memory. He was extremely effective (a 2.39 RA vs. the league's 5.25), he was a workhorse (105 1/3 innings in relief), and he even did the things that don't show up in the box score (he stranded 20 of 23 inherited runners). But because Bobby Howry got the lion's share of White Sox saves (28 to Foulke's nine), no one noticed.

Foulke's experience is shared by plenty of middle relievers throughout the league. The best set-up guys pitch innings that are just as high leverage as the closers' -- maintaining a tie in the eighth inning is just as important as protecting a one run lead in the ninth inning -- and they are often more effective than closers in preventing runs. But when it comes time to hand out plaudits, awards, or (most importantly) contracts, the guys with the saves come out on top. Some of the great set-up men have eventually gotten some recognition by getting a promotion to the closer role; Foulke, Mariano Rivera, Jeff Zimmerman, and Dave Veres are recent examples. Just as often, though, they continue to toil away in obscurity, adding zeroes to the opposition's line scores but not to their paychecks.

The source of this injustice is baseball's inexplicable fascination with the save statistic, a measure that distorts reliever evaluation and even, as Jayson Stark wrote last year, the way the game is played. There are many better methods for measuring a reliever's performance than saves, one of which is simply to measure the number of runs he keeps from scoring.

At the Baseball Prospectus web site, we track a statistic called Adjusted Runs Prevented (ARP) that measures the number of runs a reliever prevents compared to a league average pitcher. An extension of earlier work by Gary Skoog and Steve Schulman, ARP recognizes that a reliever's job is inherently situational: he's not just trying to keep his own runners from scoring, he's also trying to keep the runners he inherits from other pitchers from scoring.

I won't give any more of an explanation than that here, but you can find out all the details by looking at the current Reliever Report at the BP site and following the links from there. While the site's reports track only single-season numbers, here we'll provide the cumulative numbers for the past four-plus years (all the years for which I have the data) so we get a sense of which relievers have performed consistently over a larger number of innings. Here are the top 15 relievers in preventing runs since 1998, through Wednesday's games:

                ARP
Keith Foulke     85
Mariano Rivera   84
Derek Lowe       72
Scott Sullivan   72
Danny Graves     71
Jeff Zimmerman   61
Jim Mecir        56
Armando Benitez  56
Trevor Hoffman   56
Robb Nen         52
Greg Swindell    50
Mike Remlinger   49
Steve Kline      49
Billy Wagner     48
Dave Veres       48

Foulke finishes ahead of everyone, even the great Rivera, as the top run-preventing reliever of the past four years. To put those 85 runs prevented in perspective, that's about the number of runs above average that Brad Radke has been worth as a starter over the past four years. When you add in the fact that Foulke is doing nearly all of his pitching in high-leverage situations, it gives you a sense of the value a top reliever can add to a team.

Some of the names on the list will come as no surprise. Rivera, Hoffman, Benitez, Nen and Wagner would make anyone's short list of the league's best relievers. On the other hand, how often do you hear Scott Sullivan, Jim Mecir, Mike Remlinger and Steve Kline mentioned in that lofty company? Not often, yet those guys are more effective than most closers at keeping runs off the board in the late innings. Given their performance in recent years, it's hard to justify them serving second-class status to save-collectors like Troy Percival (18 ARP), Antonio Alfonseca (13), and Jose Mesa (-4).

Sullivan is just the head of a long parade of Reds relievers to finish near the top of the run prevention list. Danny Graves, Gabe White (44 ARP, some of it with the Rockies), and Scott Williamson (27) have also been major contributors to the Reds having the best team bullpen since 1998. Cincinnati's reliever factory is part of the reason I think Don Gullett is one of the best and most underrated pitching coaches in the game today.

One of the reasons we see so many set-up men on the list above is that they get more opportunities to prevent inherited runners from scoring. The typical closer outing involves coming in to start the ninth with a one-, two-, or three-run lead. Set-up men, on the other hand, are usually brought in to put out fires as they arise, and that often means entering mid-inning with runners on base. ARP includes how the reliever handles these inherited runners, and we isolate the handling of those runners in another stat, called Inherited Runs Prevented (IRP -- yes, the acronyms can be confusing).

IRP measures the number of inherited runners the reliever prevented from scoring compared to an average pitcher. In a sense, this represents the reliever's charity work: these are runs that would have been charged to other pitchers if they scored. By preventing them from scoring, the reliever is lowering someone else's ERA. Here are the top 10 since 1998:

                IRP
Scott Sullivan   19
Ricardo Rincon   17
Jeff Zimmerman   15
Greg Swindell    15
Dennis Cook      14
Steve Reed       13
Danny Graves     12
Bob Wells        11
Curt Leskanic    11
Mike Stanton     11

This list is practically closer-free, especially considering that Zimmerman had his big year with inherited runners in 1999 as a set-up man. The list shows that a lot of the value of the best platoon specialists comes from stranding inherited runners. Lefty killers Ricardo Rincon, Greg Swindell and Dennis Cook are here, along with righty killer Steve Reed. In Cook's case, for example, the measures show him preventing 14 inherited runs compared to just 26 overall.

These measures are fun for "who's best?" discussions and for understanding the relative importance of relievers, but of course there's also an important team-building lesson here. Teams that refuse to pay a premium for gaudy save totals and instead use their relief budgets to stockpile proven run preventers will have a competitive advantage. You're likely to keep more runs off the board with two Jeff Nelsons than one Troy Percival.

You can check out more work from the team of writers of the Baseball Prospectus (tm) at their web site at baseballprospectus.com. Michael Wolverton can be reached at mjw@baseballprospectus.com.