It can’t be that hard, can it? Draft a few established closers with great ERAs on great teams, maybe a couple elite guys. Then find some diamonds in the rough — look for a reliever or two with great peripherals behind closers with poor peripherals on a good team with a good pitching staff that’s often in close games. Boom, roasted. Teh savez.
Except it’s not true. None of it.
Here’s what the relevant research has found about saves. You might be surprised.
1) A reliever’s projected ERA, previous season ERA, and three-year FIP are not strongly correlated with his future amount of saves. [Card Sharks]
2) The difference between the expected amount of saves for the favorite in a committee and the underdog is negligible. [Card Sharks]
3) Closing experience only explains about 4% of the variance in saves in any given year. [Card Sharks]
4) Managers prefer righty closers as sole closers. [RotoGraphs]
5) Even the worst teams usually provide at least 30 saves a year. [RotoGraphs]
6) Team winning percentage only explains about 8% in the variance in save opportunities. [RotoGraphs]
7) Teams that play in close games are not more likely to have save opportunities. [RotoGraphs]
8) Teams with good overall pitching staffs do not provide more save opportunities. [RotoGraphs]
9) Good bullpens only explain about 5.5% in the variance in save opportunities. [RotoGraphs]
10) Team runs scored is the team stat best correlated with save opportunities — and they explain 12.1% of the variance in save opps. [RotoGraphs]
Kind of amazing how many tried-and-true tropes are turned on their heads here. Winning teams don’t provide all that many more save opportunities than bad teams. Even teams that play in close games don’t rack up the save opps. Great relievers that haven’t closed before and are entering the season in an unsettled situation aren’t worth large investments.
If you read between all the lines of negativity, there are some guideposts in there. Lefty closers are about half as rare as they should be. Predicting saves is tough, and you probably shouldn’t invest too heavily in a closer, especially if they don’t have the job yet. Team factors don’t matter much, but teams that score runs and have good bullpens should get a tiny tick in their favor. As a tiebreaker, perhaps.
Some of this is best filed for next year. It means that Mariano Rivera wasn’t that much better of an investment than Joel Hanrahan going into the season, and certainly no more than J.J. Putz. It means that a situation like the one the White Sox had going into the season is best left alone, unless you are spending a dollar or a late-round pick as a lottery ticket.
It makes sense to look for pitchers with good strikeout rates, low walk rates, and nice ground-ball rates if you can find them — after all, you’re still trying to predict a good season from a pitcher — but as long as there are no big red flags, the default closer for the upcoming season is the same as it was for the past season. Think about Chris Perez as an example.
What does this sort of thing mean in-season?
Well, you might look at Scott Downs and Ernesto Frieri and realize that the right-hander is the better bet. You could say the same for James Russell and Shawn Camp, but if someone is trying to sell you Camp, or you are trying to determine how much free agency auction budget you want to put down for the Cubs’ current co-closer, remember that he’s on a bad team that doesn’t score runs and has a bad bullpen. His ceiling is lower than that of a Frieri, for instance.
You might say much of the same for the Athletics, who have flamethrowing lefty (and former first baseman) Sean Doolittle coming up behind flamethrowing righty (and control-challenged) Ryan Cook, but you’d have to add a wrinkle: Brian Fuentes is still the closer. So he’s the guy to own, really.
Mostly this all means that crazy stuff happens in small samples, so it’s not surprising the old cliches don’t really hold water. Your typical closer pitches 60+ innings, and anything can happen in 60+ innings. Look up Greg Aquino or Joe Borowski for a little fun while you’re trolling your wire for saves. And when things are a crapshoot like this, quantity makes more sense than putting all your ducks in one quality basket.