But there is one way the Mets could keep Rodriguez from finishing 55 games while plausibly arguing it is for baseball reasons. They could essentially redefine the role of a closer, from a pitcher who is used almost exclusively in save situations to one that is used in the most important situations in a game.

Bases loaded in a tie game in the seventh inning? Bring in the closer. Clinging to a one-run lead with the heart of the Phillies’ order due up in the eighth? Bring in the closer. And if Rodriguez’s pitch count is too high by the ninth inning? Bring in someone else.

It’s a radical idea, but not a novel one. Sabermetricians have long argued that by making closers ninth-inning specialists, teams fail to maximize their potential value. The key word is leverage, which measures the relative importance of each situation on a game’s outcome.

– Brian Costa, Wall Street Journal.

I’ve beaten this drum about a million times and Costa absolutely nails it here, so there’s really little to add. I think he may be off in his use of Leverage Index — he writes that 19 pitchers came into games in on-average higher-leverage situations than Rodriguez, and I’m not sure that’s the case. I will cop to not being that comfortable with LI stats, but it says here that gmLI is the pitcher’s average LI when he enters the game. And Rodriguez actually had the top gmLI on the Mets last year.

But I could be wrong, and it doesn’t matter anyway. Costa’s point is that a team’s best reliever should be used in the most important situations, and no one could argue that coming in with the bases empty and a three-run lead in the ninth is a tighter spot than bases loaded with no outs in a tie game in the seventh. Too often, managers use their closers in — and only in — save situations, and managing to a powerfully flawed stat just doesn’t make a lot of sense.

I will again cite this tremendous Forbes research from 2007. It says:

Those veteran closers making millions can thank Tony LaRussa, who ushered in the era of hyper-specialization as manager of the Oakland A’s in 1988. That year, LaRussa decided that his best reliever, Dennis Eckersley, would be used strictly to protect ninth-inning leads. Other clubs soon followed suit, and top relievers found themselves racking up more saves (the biggest stat driving their paychecks) while pitching fewer innings.

Yet in the 20 seasons since LaRussa’s brainstorm, teams holding late leads have won at about the same rate they did in the 20 seasons before. Since 1988, teams leading after eight innings have won at a .951 clip, according to Baseball-Reference.com and STATS Inc., compared to .948 from 1968 to 1987. That adds up to less than one win per season per team.

To be fair, relievers do seem to appreciate defined roles, which makes sense. Anyone would prefer to know what he’ll be asked to do at work that day before he shows up at the office. But perhaps Collins, with the help of the Mets’ front office, can come up with some sort of newly defined role, or — dare I suggest — tell every reliever that his role is defined as: “guy who gets people out when we ask him to.”