The Fourth Inning Got an Owner. It Broke Anyway.

Forty-seven.

That is how many pitches Didier Fuentes threw at Dodger Stadium on Sunday night, and they are the pitches that decided a game Atlanta won 3-2. Fuentes covered the sixth inning, the seventh, the eighth, and two outs of the ninth. He walked nobody. He allowed one hit. That hit was a home run that made the final margin uncomfortable without making it different.

The series is tied 1-1.

I want to begin with Fuentes because I spent Saturday and Sunday building a ladder that would have ranked him near the bottom of it.

The inning was right. The arm was wrong.

Sunday morning I published a sentence I still believe: Atlanta has a repeatable three innings and no repeatable fourth. I called the fourth inning the seam. I wrote that the fourth had no owner.

Sunday night the fourth inning got an owner.

Ray Kerr started and gave Atlanta two and two-thirds innings on 42 pitches - two hits, one walk, four strikeouts, nothing across. Grant Holmes came in to finish the third and stayed out there for the fourth. In the bottom of the fourth, Freddie Freeman hit a ball into the right field seats off Holmes, and a 1-0 game became a 1-1 game.

So the seam held open. Same inning. Second consecutive game.

Here is the part that is mine to carry, and I would rather not bury it. Holmes is the arm I named. On Saturday I called him the next rung, the one that mattered, the only pitcher on the staff I credited with a starter's build and a workable pitch count, and I spent two pieces observing that the club had walked past him in Game 1. On Sunday the club did not walk past him. It handed him the precise inning I had identified as the problem.

The problem happened anyway.

That is twice in two days that my ladder has named the wrong arm. Saturday the denominator was short - I said two arms were unspent when the honest count against a thirteen-man staff was four. That was an arithmetic error and I corrected it the same way I am correcting this one. But Sunday was not arithmetic. Sunday the whole organizing idea was wrong, and that is the larger miss, because a short denominator can be fixed with a longer look at the roster feed and a wrong idea cannot be fixed with anything except a better idea.

What a ladder cannot see

The ladder ranked arms by what each had already spent. It is a reasonable way to think about a staff in a week this compressed, and it has now failed twice against the same opponent.

Consider what it would have said about Fuentes. He pitched in all three Wild Card games against Philadelphia, on three consecutive days. By the only logic the ladder knows - the logic of accumulated work - Fuentes belonged among the last arms Atlanta would reach for on Sunday night. He is exactly the pitcher a spreadsheet protects.

He threw more innings than any Atlanta pitcher in the game. More than the starter. More, by a full inning, than anyone Atlanta used in Game 1 except Dylan Dodd.

This is, statistically speaking, not what I projected.

What I think I got wrong is subtler than the arm I picked. I treated availability as a property of a pitcher - a thing he has or has spent, countable, reducible to a number of pitches. The club appears to treat it as a property of a matchup, and those are not the same claim. A ladder sorts pitchers. A manager sorts innings, and the two lists do not have to agree.

I am not going to tell you what Atlanta meant by any of this. Nobody from the club has said anything on the record about Sunday's sequencing that I have found, and I held that line Sunday morning for the reason I hold it now: several readings fit, and a confident guess dressed as reporting is worse than an honest blank.

Six arms, 138 pitches

The shape of the night, stated once.

Kerr 42 pitches. Holmes 17. Dylan Lee 14, which were enough to make him the winning pitcher. Robert Suarez 15, covering the fifth. Fuentes 47. Raisel Iglesias three pitches for the final out and the save. Six pitchers, 138 pitches.

Game 1 took four pitchers and 144 pitches. So Atlanta used two more arms and seven fewer pitches, and the difference between those two afternoons is not workload. It is distribution.

One more figure for the ledger, because this is the column I have been keeping all week. Atlanta threw 404 pitches across the three Wild Card games and 144 in Game 1. Add Sunday's 138 and the total is 686 pitches, five games in six days, and the staff has one day of rest before doing this again in Georgia.

And a note on the starter, because it belongs in the record. Ray Kerr made exactly one start in the 2026 regular season. One. He appeared in nine games, threw 18.1 innings, walked two men all year and carried a 0.60 WHIP, and the role that produced those numbers was not this one. He had also opened the October 1 clincher on 47 pitches, which is the same number Fuentes threw on Sunday night. That means nothing whatsoever, and I checked it twice anyway, because a number that repeats inside one week is usually me making an error. Atlanta's answer to a division series game it could not afford to lose was a reliever making the second start of his season, and he gave them two and two-thirds scoreless innings. The numbers tell a story, but not the one you would expect.

Not solved. Routed around.

So what actually happened on Sunday night?

Atlanta did not fix the fourth inning. It is important to be precise, because the win invites the opposite reading. The fourth inning is where the only earned damage of the night occurred, for the second game running, and it occurred through the arm best positioned to prevent it. If the fourth were a solvable problem, Sunday was the night it got solved, and it did not.

What Atlanta did instead was make the fourth inning smaller. Holmes got one out in it. Lee got the next two. Suarez took the fifth clean. And then Fuentes absorbed four innings that, in the model I published Sunday morning, were supposed to be the hardest part of the night to staff.

That is not a repaired seam. That is a club that has stopped trying to find a fourth starter inside a bullpen and started asking one reliever to be three of them.

The obvious thing about whether that holds: it is a tactic with a short shelf life. Fuentes has now pitched in four of Atlanta's five postseason games. The arm that solved Sunday is the arm Sunday spent, and Game 3 is Tuesday evening with a single day off in between.

One sentence about the offense, which I leave to a colleague. Atlanta scored three runs without a home run, the first runs by either club in this series produced any other way, and what it meant for the pitching is that 138 pitches spent the night protecting a one-run and then a two-run margin.

The column with nothing in it

I keep coming back to a column I started on Saturday: the arms that had not yet thrown. On Saturday morning it held four names. By Saturday evening three of them had pitched, and two of those three had allowed every run Los Angeles scored.

After Sunday, the list of Atlanta pitchers who have not thrown a pitch in this Division Series is shorter again, and the most interesting name on it is still Chris Sale - 14-9 with a 2.16 ERA and 200 strikeouts across 162.2 innings in the regular season, 116 postseason pitches spent, and not a single one of them in this series. Tyler Mahle is still listed Day-To-Day. And JR Ritchie is still the only man on the thirteen-pitcher staff without a pitch in this postseason - this October, not a career, because I got the broader version of that wrong on Sunday.

Baseball has a long memory for the arms a club finds in October. Atlanta came to Los Angeles needing one win and got it from a reliever on his fourth appearance in six days, an inning after the arm everybody wanted gave up the tying run.

Tuesday is Game 3 at Truist Park. Wednesday is Game 4, and Wednesday is no longer conditional. Atlanta spent Sunday night earning a second home game and 47 pitches it will not get back, and I have spent two days being right about the inning and wrong about everything else in it.