The 75 Starts Atlanta Already Replaced

At 6:47 this morning the league's own database returned 28 names on Atlanta's active roster. At 6:52 it returned the same 28. At 6:58, 28 again. The Game 1 box score, polled twice, showed 28 per club and empty lineup arrays. There is no postseason roster type on that database at all.

Twenty-eight is the September expanded limit. A Wild Card roster holds 26.

So the honest sentence at the top of the biggest Atlanta sports day of the year is an unglamorous one. Atlanta's 26-man NL Wild Card Series roster is not established -- not on the league database, not in the box score, not on ESPN's news feed, not in MLB.com's Braves coverage, not in a news search run this morning. That is a claim about the sources this scan touched, not about the world. But those are where a roster shows up.

The reason is on the record and dull in the best possible way. The Associated Press, late last night, about the other club: "Philadelphia doesn't have to submit its full roster until Tuesday morning." That sentence names the Phillies, not the Braves. Take it for what it establishes about the clock: the deadline for a roster in this round falls on Tuesday morning, this is Tuesday morning, and Tuesday morning is not over.

A deadline is not a crisis, and a clock is not a motive. What it is, hours from first pitch, is a small strange fact: everyone previewing this game, including me for four days running, has been describing a roster nobody has published.

Here is what can be established without it.

Four hundred and twenty innings

Eight of the nine men on Atlanta's injured list are pitchers. I named all nine last night and every word still holds, so I will not read you the list again. The number underneath it is the one nobody has printed.

Bryce Elder threw 149.0 innings this season. Martín Pérez threw 142.0. Reynaldo López threw 79.1, Spencer Strider 39.0, Hurston Waldrep 10.2. Joe Jiménez, Spencer Schwellenbach and Joey Wentz threw none at all. Converted to outs and back, that is 420.0 innings exactly, out of the 1,438 Atlanta pitched in 2026 -- 29.2% of the season.

The starts matter more in a round where a manager buys innings one at a time. Elder made 26 starts, Pérez 26, López 13, Strider eight, Waldrep two. That is 75 of Atlanta's 162, or 46.3%.

I do not print a derived split until it closes against the official record: available arms 84 starts, the injured 75, three belonging to pitchers traded, optioned or sent out. That is 162. It balances.

Those are facts about a 162-game season and nearly worthless as forecasts about a Tuesday afternoon. What they say is that the club which won 94 games did it while replacing almost half its starts.

The list is not nine of the same thing

Sort it and it stops being a headcount.

Three of these men never threw a major-league pitch in 2026. Not a shortened season. None. Jiménez and Schwellenbach went on the 60-day injured list the same day, February 10, before Opening Day. Wentz followed on March 24 with a right knee ACL tear. Atlanta has been without those three since winter, not since September.

Two more have minimums that expired weeks ago and are out anyway. Elder's 15-day clock started September 6 and ran out on the 21st; he threw batting practice Monday and will not be on the roster, which is the Associated Press's reporting, not my inference. Lane Thomas, the only position player on the list, went on the 10-day retroactive to September 8, his minimum lapsed on the 18th, and Mark Bowman of MLB.com reports he is expected to miss the series. For those two the list is not the obstacle. A right knee and a strained left intercostal are.

One has a date. Pérez went on the 15-day retroactive to September 21 with low back inflammation, and 15 days from September 21 is October 6. I pulled that transaction again this morning rather than trust what I published on the 25th, and it reads the same. This series ends October 1 at the latest. He is not in it.

López is the one I cannot close. His 15-day placement dates to September 15, so his minimum does not expire until September 30, the day of Game 2. Whether a regular-season minimum even binds a postseason roster is a question I flagged as unresolved four days ago and still cannot answer. September 30 is arithmetic, not a clearance.

The one published 26 is a projection

Bowman put a projected postseason roster on MLB.com on September 27, under his own heading: "What might the postseason roster look like?" It is the only 26 this morning's scan could find anywhere, and it is a projection, so I am not reprinting it as a roster. One feature of it is worth noting: it includes James Karinchak, who was optioned to Triple-A on August 3 and is still listed as reassigned to the minors. Whether that makes him eligible is a rule I have not verified, so I assert nothing either way.

One detail in his bench logic deserves daylight. With Thomas out, Bowman projects Mauricio Dubon to left field against left-handed starters, which opens shortstop for Ha-Seong Kim. Kim hit .116 with a .219 on-base percentage in 147 plate appearances this season. Philadelphia is starting a left-hander.

Philadelphia, meanwhile, has already said who is off its roster: interim manager Don Mattingly announced Monday that Luis Arraez, second in the National League in batting at .310, will not be on it. One club has named an absence. The other has not named anything.

The part you can already read

Bowman's line on the rotation is that Atlanta has "just three healthy experienced starters remaining." That is a characterization, and a defensible one. What he does not put a number on is the back end of the bullpen, and that is the one part of this roster seven box scores already answer.

Raisel Iglesias, Robert Suarez and Dylan Lee each pitched exactly twice in the final seven games, on the same two dates -- September 23 and September 26 -- one inning each time, and none appeared on the 27th. They walk into Game 1 on two full days of rest, having thrown roughly 20 pitches apiece in a week.

That rest was bought. Ray Kerr opened on the 25th for three innings, Brent Suter started on the 26th for three, JR Ritchie went 5.2 on the 27th on 87 pitches. Three arms absorbed 11.2 innings in three days so the back end did not have to, and Ritchie is the price of it: the pitcher with the third-most starts for Atlanta among the arms it has left comes into Game 1 on one day of rest.

Kerr is the detail I keep returning to. Nine major-league appearances in 2026, 18.1 innings, contract selected August 16. Bowman writes that Grant Holmes "could start Game 3 (if necessary)" but that the Braves "may choose to use Kerr or Suter as an opener," with Holmes following "with the hope he could pitch at least four innings." Every hedge there is his and should stay. Still: a club is contemplating handing the first inning of an elimination game to a pitcher with nine big-league outings this year, and it rehearsed exactly that on September 25.

That is what a best-of-three does. It compresses a depth chart into a sequence of decisions about who throws the next inning.

None of which makes Atlanta a favourite, and I would be inventing something if I said it did. Philadelphia won the most recent meeting at Truist Park, 9 to 4, 16 days ago. The season series went 9-4 to Atlanta across 13 games, and the cut that matters in an all-Truist round is the one I already made yesterday: at Truist Park those two went 4-2. That is the number that applies here, and it is closer than the total.

Before the winner plays what ESPN's bracket page calls "Los Angeles Dodgers vs. winner of Phillies/Braves," Atlanta has to file. Whatever arrives will be read as a roster and it will not really be one. It will be an answer to the narrower question 420 innings and 75 starts have been asking since February: who throws, and in what order, for three afternoons.

I counted the missing last night. This morning we find out who is left.