Bloomberg Sports' Eriq Gardner recently wrote a column entitled True Value of Great Relievers where he shows that the difference between Jonathan Papelbon (38 saves) and Fernando Rodney (37) in ERA and WHIP last year was slightly greater than the one between Wandy Rodriguez (3.02 ERA) and Jon Garland (4.02).

Let's look at this a different way and compare Mariano Rivera last year straight up to C.C. Sabathia. Last year, Rivera threw 66.1 innings, allowing 48 hits, 12 walks and 13 earned runs. Sabathia pitched 230 innings with 197 hits, 67 walks and 86 earned runs for an ERA of 3.37 and a WHIP of 1.148.

If we subtract out Rivera's stats from Sabathia's we get: 163.2 IP, 149 hits, 55 walks and 73 earned runs for an ERA of 4.01 and a WHIP of 1.25. That means in deeper leagues Sabathia was a good deal more valuable in those categories, but in shallower ones his impact above Rivera's there was marginal.

Elite middle relievers (Matt Thornton, Mike Adams) also contribute heavily in ERA and WHIP in shallow leagues, but without the saves from a closer or wins and Ks from a starter that you need to remain competitive in those categories. You cannot roster them in a 12-team or less mixed league except as closers-in-waiting on your bench.

The one exception is in Yahoo! leagues with that low 1250 innings cap. Because everyone in the league should hit the cap, the strikeout category is really K/9, and the wins category becomes Wins/9. In the former category Thornton, Adams et. al actually crush the top starters while making a |STAR|nearly|STAR| equal impact on WHIP and ERA (Nearly as good as Sabathia, but nowhere near as good as Tim Lincecum, Roy Halladay) But let's look at the question of whether they're liabilities or assets in Wins/9.

Thornton for example has averaged 5.5 wins in 70 innings over the last two years. Sabathia has averaged 18 wins in 241 innings. That puts Thornton at .079 wins per inning and Sabathia at .075.

In other words, if you project your 70 IP middle reliever to win five games and strike out a batter per inning, he's gold in your low innings cap yahoo league in four categories. And he's got a chance to be a closer and five-category monster like Rivera.

This will not work in uncapped shallow leagues because good starters can be found who will rack up more total strikeouts and wins without destroying your ERA. In deeper leagues, good starters also provide more ERA and WHIP value (because replacement ERA and WHIP are lower), but by using middle relievers you're giving up less strikeouts and wins to the bottom half of the rotation, and you're saving a lot of ERA and WHIP damage.

Just keep in mind one caveat before you try this: middle relief performance is at least as fickle as that of starting pitchers, if not more so. Rafael Perez was Thornton heading into last year, and Joe Nathan was Rivera. Moreover, new sources of middle relief dominance are plentiful on the waiver wire every year, so it's probably not smart to pay too much for these guys.