History has a strange way of repeating itself. On Tuesday, the Mets pitched their big gun, Johan Santana, and won. No surprise there: the Mets have gone 21–12 when their lefty ace starts, the equivalent of being a 103-win team over a full season. When anyone else starts, they are more of an 86-win team over a full season. The difference between those two versions of the Mets is their biggest impediment to keeping their hold on a postseason slot.

Unfortunately, with Santana unavailable to pitch until Sunday — manager Jerry Manuel has ruled out a return on short rest — it is the “everybody else” version of the team that will determine its destiny, particularly this Saturday against the Marlins, a date for which the Mets have yet to announce their starter.

With John Maine hurt and rookie left-hander Jon Niese hammered in two out of three major league starts, the Mets are looking for a competent starter for what will almost certainly be a mustwin game. That the Mets will be facing a tough starter that day in the surprising 15-game winner Ricky Nolasco, who has allowed only 11 runs in 26.2 innings against them this year (a 3.71 ERA) only adds to the urgency of the matter.

Last season, when the Mets were unable to arrest their epochal collapse, a similar paucity of plausible starters contributed to the making of history. There are many potential smoking guns among the pile of late-season losses that did the Mets in, but certainly the team’s 1–5 record in games started by 31-year-old scrap heap refugee Brian Lawrence were key. Two starts in particular stood out even at the time: his late-August loss to the Phillies and, with 13 games to go, his touching off a 12–4 loss to the Nationals on September 17. Lawrence lasted just 3.1 innings in that game, to be followed by every trash-time arm that Willie Randolph had in the bullpen. Win or lose, Saturday’s game may end up bearing a strong resemblance to that game of just over a year ago.

The main difference will be the cast of characters and the quality of the potential starters. Lawrence went into his final start in a Mets uniform with a 6.31 ERA and no reason to believe he could do any better. Manuel has identified the choices for Saturday’s start as Nelson Figueroa, Brandon Knight, and Niese. This is in no way a fourstar buffet of possibilities for Manuel, but at least, unlike Lawrence, each pitcher has had a solid start this year. Figueroa and Knight are strictly conservative, no-second-guess possibilities, a bet placed on mere competence based on age and experience, even if not all of the experience is good.

Niese is only 21 and has pitched just 10 games above Double-A, but his September 13th start against the Braves, in which he pitched eight innings of shutout ball while striking out seven, remains a catnip-like enticement to allow him another start. There is also the additional consideration of the Marlins’ hitting predilections. Though the Marlins rank second in the National League in home runs, with 203 (the Phillies lead with 208), they don’t do very well against lefties, hitting just .236 AVG/.317 OBA/.390 SLG against them versus .259/.328/.447 against righties. Niese is the only left-hander of the Mets’ three options.

Whoever the starter is, he will undoubtedly be on a short leash. If Niese were to falter — and his rough start against the Cubs, in which he was undone by a grand slam by pitcher Jason Marquis, a fluke that almost demands dismissal — Figueroa and Knight would still be available in the pen, where the former has had more success this season (a 4.09 ERA, albeit in just 11 innings) and Knight’s stuff plays a little better. The Mets may also have a recuperating Maine available for relief work. The team will be wellstocked with long relief.

In a scenario such as this one, in which any of the starting pitchers is a transient figure at best, the obligation of the manager is to correctly line up his fallback positions so as to give the team the best chance to win. While the hesitation to trust the untested Niese is understandable, the other options are not so advantageous as to erase the advantages granted to the Mets if he starts.

In the history of baseball, just a few seasons have come down to the manager’s selection of one pitcher to make a season-deciding start. John McGraw going with a spent Christy Mathewson against the Cubs in 1908. Joe McCarthy making the weird choice to start Denny Galehouse against the Indians in 1948’s playoff game. Connie Mack choosing the superannuated Howard Ehmke to pitch against the Cubs, and reaping a dominating start, in the 1930 World Series. Saturday may not go down in history the way those games have, but it could. If Manuel chooses poorly, it will just be another loss in two years of such losses. If he gambles and wins, the world will be his oyster.

Mr. Goldman writes the Pinstriped Bible for yesnetwork.com and is the author of “Forging Genius,” a biography of Casey Stengel.