Steve Cohen Has Sold Before. The Data Mostly Agrees with Him.
The question facing every front office whose team has fallen well under .500 by Memorial Day is the same question Cohen faced two years ago and the same question Hannibal’s opponents kept getting wrong: when the present campaign is lost, do you spend more men trying to save it or do you preserve what you have for next year? In the modern major-league version, the decision compresses into a few weeks in late July and early August around the trade deadline, and it gets framed in tabloid English as buyers vs. sellers. The cleaner framing is older. A sell is a Fabian retreat. A hold is the army that marched to Cannae.
That framing is rhetoric. The data is more cautious about it, as the data usually is. To get a real answer to the sell-or-hold question, this newsletter pulled every clear example since 2010 of a major-league team that arrived at the late-July trade deadline well out of contention — sub-.450 winning percentage or ten-plus games back of a playoff position — and either traded away multiple significant veterans (the sellers) or kept its roster more or less intact (the holders). The sample is small. The signal, accordingly, is also small. Both facts deserve to be said in plain English.
The Sellers: Eight Cases, 2010–2024
| Year | Team | Deadline Record | Major Pieces Sold | Year+1 W% |
|---|---|---|---|---|
| 2012 | Marlins | 47–55 | Hanley Ramirez, Aníbal Sánchez, Edward Mujica | .383 |
| 2017 | White Sox | 38–49 | Jose Quintana, Todd Frazier, David Robertson | .383 |
| 2018 | Orioles | 28–73 | Manny Machado, Zach Britton, Kevin Gausman, Schoop | .333 |
| 2021 | Cubs | 48–55 | Bryant, Rizzo, Báez, Kimbrel, Chafin, Tepera | .457 |
| 2022 | Reds | 44–57 | Luis Castillo, Tyler Mahle, Brandon Drury, Pham | .506 |
| 2023 | Mets | 50–55 | Scherzer (TEX), Verlander (HOU), Robertson, Canha, Pham | .549 |
| 2023 | Nationals | 46–59 | Jeimer Candelario, several relievers | .438 |
| 2024 | White Sox | 27–81 | Fedde, Kopech, Pham, Jiménez, DeJong | .370 |
The unweighted Year+1 winning percentage of the eight sellers is .427. If you strip out the 2023 Mets, the most dramatic single bounce in the sample, it falls to .410. Neither number is a playoff record, but both are an improvement over what most of those teams could have expected without the trade returns. The 2023 Mets are the standout case — a fourteen-game improvement to 89–73 in 2024, an NLCS appearance, and a directly visible payoff in players acquired (Luisangel Acuña, Drew Gilbert, Ryan Clifford). That payoff is, for the question this issue is built around, the most important data point in the table.
The Holders: Six Cases
| Year | Team | Deadline Record | What They Did | Year+1 W% |
|---|---|---|---|---|
| 2018 | White Sox | ~40–60 | Held core through mid-rebuild | .447 |
| 2022 | Pirates | ~40–60 | Refused to sell Reynolds, Bednar | .469 |
| 2022 | Royals | ~40–60 | Sold Benintendi (rental) only; held the core | .346 |
| 2023 | Rockies | 42–64 | Partial sell of rentals; kept young core | .377 |
| 2023 | Athletics | ~32–72 | Already gutted; minimal additional moves | .426 |
| 2024 | Rockies | ~45–60 | Refused offers on McMahon; held veterans | .265 |
The holders averaged .388 the following year. The 2024 Rockies are the catastrophic example on the other side of the ledger — they held Ryan McMahon and several other veterans through the 2024 deadline and proceeded, in 2025, to lose 119 games and finish at .265, among the worst seasons in modern baseball. The Pirates, who held in 2022 and have continued to hold most of their movable veterans since, have produced .469 and .469 in the two years that followed; competitive in spurts, never a playoff team. The pattern across the holders is roughly what you would expect: refusing to retool tends to produce more of the season you just refused to retool out of.
The Picture, Plotted
The Honest Read
Three things have to be said about that six-game gap before it gets used to settle the Mets’ July decision. First, the sellers in this sample started their deadline at a marginally better record than the holders — about a forty-point edge in deadline winning percentage — which means part of the next-year improvement is mean reversion from a slightly less-bad floor rather than a pure trade-return effect. Second, the sample is fourteen teams across fifteen years, which is enough to suggest a direction but not enough to nail down a magnitude. Third, the largest single bounce in the entire sample is the 2023 Mets — the same franchise this issue is built around — and a sample whose headline is “the case you are interested in is the outlier” is a sample to read carefully.
What survives all three caveats is the directional finding. Sellers, on average, finish next year ahead of holders. The advantage is modest but it points the same way every time you cut the data — with or without the Mets, by next-year W% or by playoff probability, at one-year horizon or extended into the second year as the prospect returns mature. The more useful version of the question, then, is not which group does better; it is what makes the sellers who did best different from the ones who did worst.
Why the 2023 Mets Bounced and the 2018 Orioles Did Not
The 2023 Mets and the 2018 Orioles both sold heavily into a deadline they entered out of contention. The Mets recovered in one year and made the NLCS. The Orioles bottomed out at 47–115 the following year and did not return to the postseason for five seasons. The reasonable explanation is roster age. The Mets sold rental veterans (Scherzer was 39, Verlander 40, Robertson 38) and the remaining core — Lindor, Alonso, Nimmo, Diaz — was either in or just past their primes. The Orioles sold their best young player (Machado) and the surrounding core was older and worse than Baltimore’s scouts believed. A sell rewards the team that has something left after the sell is finished. It does not rescue the team that has nothing left underneath.
The applied version for 2026: the Mets’ current core is structurally closer to the 2023 version than the 2018 Orioles, which is the part of the diagnosis that the data favors. The salary commitments to Lindor, Alonso, Nimmo, and Soto extend through the late 2020s. A sale of rental veterans — whatever short-term contracts the Mets carry into July — would be subtracted from an underlying roster that, on health and on its preseason projection, is a borderline playoff team. The Fabian retreat preserves the legions that are worth preserving. The Orioles, in the 2018 analogy, did not have legions left.
What Year Two Looks Like
One year is too short a horizon to grade most trade-deadline returns. Prospects acquired in a July deal typically arrive in the majors two or three years later, sometimes four. The Reds’ 2022 sale produced Spencer Steer and Christian Encarnación-Strand, both of whom became contributors in 2023 and 2024. The Cubs’ 2021 sale produced Pete Crow-Armstrong, who broke out in 2024. The Mets’ 2023 sale produced Drew Gilbert and Ryan Clifford, both still developing in the system. The honest version of this analysis would run on five-year horizons, not one-year, and the sellers’ advantage would almost certainly widen. The one-year gap of +.039 is a floor, not a ceiling.
“The Fabian general is unpopular for as long as the senators believe they could have won the campaign anyway. He becomes correct retroactively, every time, after Cannae.”
— The Sports Page, on the trade deadline and the Roman dictatorThe Cohen Decision, Stated Plainly
The Mets’ current roster, as of mid-June, is closer to the 2023 version than the 2018 Orioles, with an injured list that has historically been the team’s structural problem and a payroll that gives the front office options most teams do not have. Eight weeks remain until the August 3 deadline; Lindor and Alvarez are on rehab assignments and could return before the All-Star Break, which is the single piece of information that moves the calculus most. The data in this issue does not decide for Steve Cohen what he should do at the deadline. It tells him three things. Selling improved Year+1 winning percentage in seven of the past fifteen years for which there is a clear sell vs. hold comparison. The largest single improvement was achieved by his own team two years ago. Holding has, in two of the most recent six holder cases, produced a season substantially worse than the one the franchise was trying to save. The decision is still his. The base rate is what it is.
A note on the data: deadline records are approximations drawn from the closest weekly standings to each year’s late-July trade deadline; year+1 winning percentages are full-season figures from Baseball-Reference. The sample excludes teams that were genuinely competitive at the deadline (above .450 or fewer than ten games out) and teams whose deadline moves were a single rental departure. Edge cases are discussed in the inline notes. The 2026 MLB trade deadline is Monday, August 3, at 6pm ET — a calendar shift the league has been quietly making toward early August in recent years. A reader with a candidate the newsletter overlooked — especially among the 2010–2016 vintage — is invited to send it.
