Zero for Three in October

Atlanta and Philadelphia have met three times in the postseason. Philadelphia has won all three.

The 1993 National League Championship Series went six games. The 2022 Division Series went four. The 2023 Division Series went four. Atlanta lost every one of them. Sum the three rows on the Braves' own postseason results table and the games come out Atlanta 4, Philadelphia 10 — that games figure is a derivation, not a published record, and it is worth labeling as such before anyone leans on it.

Three series. Thirty-three years. None won.

That is the sentence this week arrives carrying, and the first honest thing to say about it is that three different teams produced it. Different front offices assembled them. Different rounds ended them. The first belongs to a competitive era that no longer exists in any meaningful sense. A record spanning three rosters and three decades is a real record and a poor forecast.

Because the regular season had a completely different opinion.

The split that argues the other way

Against clubs that finished .500 or better, Atlanta went 43-29. Philadelphia went 31-38.

Those are winning percentages of .597 and .449, a gap of 148 points across seventy-two games for Atlanta and sixty-nine for Philadelphia. Strip out the games against bad teams — the ones every contender banks and nobody remembers — and one of these clubs was still good and the other was a losing team. That sample is roughly forty-three percent of each schedule, and it says the same thing the standings say: Atlanta finished 94-68 and won the East, the lower branch of the range I printed on Sunday; Philadelphia finished 88-74 and got in as the sixth seed.

And it does not get to testify. A best-of-three does not resolve a 148-point gap. It resolves one week.

The shape of the nine

Atlanta won the season series 9-4 across thirteen meetings. The total has already been spent in this space — Dex printed it Sunday — so the useful thing left is the shape, and the shape is not flat.

In April, Atlanta went 5-1: three straight at Citizens Bank Park, then two of three at Truist Park. In September, Atlanta went 4-3: 2-2 in Philadelphia across four games, then two of three at home. Five and one, then four and three. Six games, then seven. Nine and four, reconciled.

A record that front-loads is weaker than its headline. The April version of this matchup and the September version were not the same matchup, and the one that finished nearest the present finished nearest to even. Whatever the 9-4 is worth, the part of it closest to Tuesday is worth the least.

There is a second cut in the same thirteen games, and it runs against the framing the whole week is built on. Atlanta went 5-2 against Philadelphia at Citizens Bank Park and 4-2 at Truist Park. The edge over this opponent was not a home edge; if anything it was marginally wider away from Truist Park. The series is being played entirely at home. That all three games are at Truist Park is not news here — I wrote it Friday — and it functions this morning as a condition rather than a headline.

The only comparable object is two games long

Atlanta has played exactly two best-of-three Wild Card Series and split them. In 2020 it beat Cincinnati 2-0, both games at Truist Park, by scores of 1-0 and 5-0, without allowing a run. In 2024 it lost 0-2 at Petco Park. One and one in series. Two and two in games. (The 2012 loss to St. Louis was the single-elimination Wild Card Game, a different object entirely, and it should not be folded in.)

The 2020 series is the only Wild Card Series Atlanta has ever hosted. It swept it and did not allow a run.

That is the entire body of evidence about this franchise hosting a best-of-three. Two games, six years old, belonging to a roster that has turned over since. It is the most encouraging number available this morning and the least usable one, and those two things are not in tension — they are the same fact stated twice.

Two very small samples, pointing opposite ways. Neither gets to decide anything, and the temptation to let one of them do the deciding is the thing to resist.

A schedule with a conditional in it

Game 1 is Tuesday at 2 p.m. ET on NBC and Peacock. Game 2 is Wednesday, same window, same network. Philadelphia comes to Truist Park under interim manager Don Mattingly; Atlanta hosts as the third seed. Atlanta's Game 1 starter is expected to be Chris Sale — that is reporting, not an announcement, and no probable pitcher was populated on any of the three game records as of this morning.

Game 3 is Thursday, also at Truist Park, and in MLB's own scheduling data it does not carry the same flag as the other two. Games 1 and 2 are marked Normal Game. Game 3 is marked If Necessary.

Then the league did something stranger. It published its Wild Card broadcast schedule with a set of alternative Thursdays attached, sorted by how many of the four series actually get that far. There is a block headed "FOR ANY THREE-GAME SCENARIOS ON THURSDAY, OCT. 1:". There is another headed "FOR ANY TWO-GAME SCENARIOS ON THURSDAY, OCT. 1:". And there is a last one, shorter than the others: "FOR ANY ONE-GAME SCENARIO ON THURSDAY, OCT. 1: If there is only one game, then it will be scheduled at 8 p.m. on NBC and Peacock."

Read that slowly. Whether this series has a third game depends on Atlanta and Philadelphia. What time that game starts depends on three other series that have nothing to do with either of them. The league has committed, in writing, to several mutually exclusive versions of a single day, and it will find out which one it published the same way everyone else does.

What both records are actually worth

Everything past this round — including the Division Series arithmetic I laid out Friday — is unreachable until this best-of-three is won, which is a sentence I have now written twice and still believe.

Here is where the two halves of this piece meet. Atlanta has a season's worth of evidence that it is the better club: 43-29 against winning teams, a 9-4 head-to-head, a division title. Atlanta also has an October record against this specific opponent that has never once agreed with any of it. Both are true this morning. Neither is a prediction, and the honest position is that a three-game series is not wide enough for either one to prove itself.

I declined on Saturday to convert a count of branches into a probability, on the grounds that the branches were not independent and the conversion would be inventing something. The same objection applies here, for the same reason. Forty-three and twenty-nine describes a population of seventy-two games. Zero-for-three describes three series drawn from thirty-three years. Tuesday is not a population. It is a Tuesday afternoon at two o'clock, with a Thursday behind it that may not happen, at a start time other people's results can still move, and both of those long records will be asked a question in a sample far too small to answer it.

That is not a complaint. It is the format. October has always been a bad instrument for measuring baseball teams, which is most of why we watch it.