In baseball, as it’s played today, teams for the most part fit players to strategies, rather than the other way around. It’s not hard to understand why. In any competitive sport, roles have to be standardized, for reasons ranging from the small differences in talent between the best and worst competitors to the risk of injury making it a better idea to build teams around roles rather than players. A manager may come up with some novel way of using his best relief pitcher that falls outside the conventional closer role, but if the pitcher blows out his elbow, leaves the team for more money, or suddenly goes into a huge slump for no reason, the manager will be up a creek. Sticking him out there in the ninth inning every time the team has a three-run lead may not be the best possible use of him, but it does, in addition to other benefits, make him more easily replaced.

That this is so doesn’t make it any less of a shame that teams don’t think through more novel means of using their players to their fullest potential. One strength of the 1986 Mets, for instance, was Davey Johnson’s willingness to run platoons based on which pitcher was on the mound: Since Sid Fernandez was one of the best strikeout pitchers in the league and never gave up a ground ball, it didn’t much matter who was playing shortstop when he pitched, so Johnson took advantage of this by playing Howard Johnson and Kevin Mitchell, getting an extra bat into the lineup. One strength of the 1999 Mets was that Bobby Valentine, who had six indistinguishable starters, used all of them for a few months in a six-man rotation, allowing his old pitchers some extra rest and thus keeping them about as effective as possible.

These weren’t radical plays, but simple applications of common sense. Looking at the Mets’ rotation headed into the playoffs, though, I wonder if they might not be better off doing something truly unusual, even while understanding there are very good reasons why they won’t and shouldn’t do so.

Right now, the Mets’ rotation has some clear strengths and clear weaknesses. The main strength is that they have quite a few good pitchers.Tom Glavine and Orlando Hernandez in particular are looking strong headed into October, but Pedro Martinez can never be counted out, and John Maine, Steve Trachsel, and even Dave Williams have all proved effective in varying ways at various times over the last few weeks.

The weakness is that none of them are great, or close to it. Hernandez’s postseason track record and great pitching of late makes one think he might have one last dominant run left in him, but that’s dodgy. Past him, given Martinez’s injuries, it’s more or less a slate of pitchers who can be expected to be good and might be treated like batting practice pitchers if things go badly. The main reason for that is no one in the rotation has very good stuff; if Trachsel leaves the bullpen with a flat fastball and no feel for his breaking ball, he’s going to get cuffed.

There is a pretty simple solution to this, one that’s even been used in the past, which is a tandem starter system. The A’s and Rangers have used it in the minors with great results, and the Cardinals even briefly tried it in the 1990s. The idea is that for each game you have two starters; the first goes out and pitches three innings, then the second one comes in and does the same. Were the Mets to do this, they’d derive two main advantages: First, they’d be able to get something out of every effective starter they have, and second, each of those starters would be able to go fullbore knowing they wouldn’t be expected to go deep into the game.

Look at their statistical splits and you can see why this might prove pretty effective. Maine allows a .449 OPS through the first three innings, .886 in the middle three. Glavine’s split is .648/.826. Trachsel’s is .766/.892. Martinez, Hernandez, and Williams actually show the opposite pattern, but with various caveats: Martinez has recently shown a marked lack of endurance, so it’s clear he’s not going to improve as the game goes on; El Duque’s numbers are a bit distorted by a game where he took one for the team and gave up 11 runs in four innings, and Williams isn’t great no matter when he’s pitching.

Basically the logic is that if you have a bunch of average six-inning starters of roughly equal value, pair them in teams and let them become superior three-inning starters, giving opponents a variety of looks and keeping your eggs from ending up in a couple of baskets.

For practical reasons ranging from warm-up routines to ego to what would happen to a manager’s reputation if he tried something like this and failed, it’s simply not going to happen. The point isn’t that it should; it’s that as we think through what can and will happen in the playoffs over the next week, it’s worth keeping in mind that there are far more ways of getting the most out of a team’s talent than most managers ever try. Sometimes, those ways even stand a pretty good chance of paying off.