The Mets Must Pass Seven Teams. They Play Two of Them.
Twenty of their last 33 games are against clubs whose losses do them no good at all. A comeback needs games you can take from the people ahead of you; the schedule decides whether you have any.
The question: which of the Mets’ remaining games actually matter?
The answer: one series. Four games against Philadelphia, September 17 to 20, at home. Sweeping it is worth more than sweeping every other series put together.
Here is the surprising part. We rate a stretch run by how hard it is. Hardness is the wrong axis. The Mets’ schedule is not too hard; it is aimed at the wrong people.
What you get from it: a better question to ask about any team chasing from behind.
A win is worth one game. Sometimes it is worth two.
Games back is a simple thing. Count half a game for every win they have and you do not; count another half for every loss you have and they do not. Add those up and that is how far back you sit.
Now watch what a single win does. Beat somebody irrelevant and only your side of the ledger moves; you gain half a game. Beat a team you are chasing and both sides move at once — you go up, they come down — and you gain a full game.
So a head-to-head win against a rival is worth exactly twice an ordinary one. That is the whole idea. It is why a schedule can be kind or cruel in a way that has nothing to do with how good the opponents are.
Count who is in the way, then count who you play
As of August 23 the Mets were tenth in a race for three wild cards. Seven clubs stood between them and the last spot: San Diego, Arizona, Miami, St. Louis, Pittsburgh, Cincinnati and Washington.
They play two of them.
San Diego, Arizona, St. Louis, Pittsburgh and Cincinnati are done with the Mets for the year. Whatever those five do in September, the Mets can only watch.
Twenty of the remaining 33 games are against clubs whose results do not touch the wild card at all. Seventeen of those are against American League teams; three are against a Milwaukee club running away with its own division.
Records go stale in a week. A schedule is fixed in January.
The ProfessorSuppose they win 25 of 33
Give them a run nobody expects: 25–8, a finish of 83–79. We played the rest of the season out 250,000 times to see where that lands.
It lands at about 2.4%. But now change nothing except where those 25 wins fall, and the number moves by a factor of three and a half.
Sweep the ten games against clubs they must pass and it is 3.5%. Scatter the wins at random and it is 2.4%. Take only two of those ten and beat everyone else instead, and the same 25–8 season is worth 1.0%.
Same record. Same number of wins. Three and a half times the value, depending on who you took them from.
Which series to spend them on
So we swept each series in turn, holding the total at 25, and ran 200,000 seasons for every one. Two results are worth your attention.
The first: only Philadelphia matters. Sweeping those four is worth +0.6 points of playoff probability. Nothing else on the schedule clears half of that.
The second is the one we did not expect. Beating Washington is worth no more than beating the Yankees. The Nationals are a team the Mets must pass, and sweeping them measured no better than sweeping a club in the other league.
The reason is worth holding onto. Two things have to be true at once: you have to play them, and they have to be someone who would still be there at the end. Washington is nine games out itself and is not going to hold that final spot. Knocking down a team that was never going to be there removes nothing.
There is a third finding hiding underneath. Sweeping a series that changes nothing in the race came out slightly negative. Wins are a fixed budget in this exercise; three of them spent on the Yankees are three not spent on Philadelphia. A win you cannot aim is a win you partly waste.
And even then, it is not up to them
In the small fraction of seasons where 83–79 is enough, look at what the rest of the league had to do. Philadelphia had to play .375 the rest of the way. San Diego .394. Arizona and Cincinnati .424. Miami .438.
Five clubs have to fold at once. The final wild card has to come back to 83 wins, a number the Mets tie rather than clear.
What to take home: count opponents, not wins
The next time a team gets hot from behind, do not ask how many they can win. Ask how many of the people ahead of them they still play.
That number is fixed and public and nobody looks at it. It was set when the schedule came out, months before anyone knew who would need it.
The habit is general. Think of a salesperson chasing a rival, a party chasing a majority, a fund chasing a benchmark. In each of them a few contests move both sides at once; most move only yours. A brilliant run against people who are not in your way is just a nice month.
Watch for the Washington trap too. Being technically ahead of you is not enough; they also have to be someone who will still be there when it is over.
Notes & sources
Standings, schedules and results from the MLB Stats API, as of the morning of August 23, 2026. Records move; the schedule does not. Whatever the standings say when you read this, the Mets do not play San Diego, Arizona, St. Louis, Pittsburgh or Cincinnati again this season, and that was settled before the year began.
Method as in the companion piece on streak timing: team strength is a 50/50 blend of actual winning percentage and Pythagorean expectation, regressed toward .500 by 69 games; game probabilities use log5 with a home-field adjustment at a .535 league home win rate; every remaining game in both leagues is simulated, and ties are broken at random rather than by the real tiebreaker rules.
The series comparison forces the Mets to exactly 25 wins in all cases, sweeps the named series, and distributes the remaining wins at random across the other games, so every row is the same record differing only in placement. 200,000 seasons per row gives a 95% interval of about ±0.07 points. Philadelphia at +0.63 clears that comfortably; Miami at +0.09 is small but separates from the control series; Washington at −0.18 sits alongside the Yankees at −0.20 and Milwaukee at −0.21.
An earlier pass at 40,000 seasons per row put Washington third and showed every series as a gain. That ordering was noise. It is quoted here because it is the ordinary way this kind of table goes wrong: the differences being measured were smaller than the error on the measurement, and nothing in the output says so.
One limit. Forcing a sweep hands over results without making the team better, and a real September sweep of a rival might carry information this exercise throws away. The direction of that bias favours the Mets, and it does not come close to closing the gap.
Baseball is ranked by championship leverage: we simulate the rest of the season 60,000 times, then split those seasons by who won each of today's games. College football cannot be simulated that way, so it is ranked by how close to a coin flip the ratings make it — a lopsided game teaches you nothing. Tomorrow this block reports what happened to today's pick.
