Nationals vs. Mets: Alvarez Takes the Ball, Mets' Rotation Unsettled
Nationals Park | Friday, Sept. 25, 2026 | 6:45 p.m. ET
The New York Mets arrive in the nation's capital at 73-86 with a run differential of minus-35 and a tragic number of one. A loss Friday — or the right result elsewhere — mathematically ends their season. The Washington Nationals, 75-84, are playing out the string, but they own the better offense (5.08 runs per game to New York's 4.29), the better run differential (plus-9), and the better starting pitcher on the card.
Yet the Mets are -134 favorites on the road.
That tension is the entire handicapping story of this game, and it's why Friday's card has to be read carefully rather than bet reflexively.
The Total Moved a Full Run — That's the Story
| Market | Open | Current |
|---|---|---|
| Total | 8.5 | 7.5 |
| Run line | WSH +1.5 | WSH +1.5 |
| Moneyline | +114 / -134 | +114 / -134 |
A full run off an 8.5 opener is a serious market statement, particularly with the moneyline frozen and the run line parked at 1.5. Somebody with conviction got down hard on the under.
Here's the squeeze. Earl's model still projects 8.21 total runs, which sits 0.71 above the current number, so the model carries an Over lean. But the model's own expected-value calculation on that Over returns -8.27. That is not a contradiction — it's the model telling you the edge it identified at 8.5 has been arbitraged out of the number entirely by the move to 7.5. When a total drops a full run and your fair number lands only two-thirds of a run above the new line, you're no longer betting a projection. You're betting the sharpest money in the market got this wrong.
Treat the Over as a lean, not a play.
Andrew Alvarez Is Not a 2-6 Pitcher
| Andrew Alvarez (L) | 2026 |
|---|---|
| Record | 2-6 |
| ERA / WHIP | 3.17 / 1.26 |
| Innings | 108.0 |
| K/9 / BB/9 | 8.92 / 4.00 |
| H/9 | 7.33 |
| Home runs allowed | 5 |
The win-loss record is the least useful number on this page. Alvarez has thrown 108 innings at a 3.17 ERA with a 1.26 WHIP, and the underlying profile is genuinely strong: 8.92 strikeouts per nine, 7.33 hits per nine, and — the headliner — five home runs allowed all season, a 0.42 HR/9 rate. He's doing that in a park with a 105 home-run factor, 337 feet down the left-field line and 335 to right.
The flaw is real: 4.00 walks per nine, 48 in 108 innings, a 2.23 strikeout-to-walk ratio. That's a pitch-count tax, and it will bite against a lineup that works counts.
The record, though, is a run-support artifact. Washington's offense has produced 8, 6, 5, 2 and 1 runs across Alvarez's last five starts — a 4.4 average that collapses to 1.5 across the two most recent outings. You don't win many games when your club scores three runs across two starts.
He also averages 6.75 innings per start (108 IP / 16 starts). In a game where Washington's bullpen has six pitchers on the injured list, that workload rate matters more than usual.
The Run Differential Disconnect
Strip away the records and the profiles point the opposite direction from the price.
| WSH | NYM | |
|---|---|---|
| Record | 75-84 | 73-86 |
| Run differential | +9 | -35 |
| Runs scored/G | 5.08 | 4.29 |
| Runs allowed/G | 5.02 | 4.51 |
| Home / Road | 37-41 home | 37-41 road |
A 44-run gap in differential favoring a +114 home underdog is not a rounding error. The Nationals score nearly eight-tenths of a run more per game than New York does, and they've been the better team by run differential across 159 games.
What the market is pricing instead is situational: a sub-.500 home team, a TBD opposing starter that could be an upgrade, and a Mets club that — despite its record — is 37-41 on the road, actually better away from Citi Field (36-45 at home) than in it.
That's the honest counterargument. It's also thinner than the price suggests, which is why the model's moneyline edge is so small you could miss it.
What Earl's Model Actually Says
| Market | Model Pick | Projection | Expected Value |
|---|---|---|---|
| Run line | WSH +1.5 | NYM 4.5 – WSH 3.7 | -9.04 |
| Total | Over 7.5 | 8.21 runs | -8.27 |
| Moneyline | NYM | NYM by 0.82 | +0.41 |
Earl's model says the Mets win this game 4.5 to 3.7 — a margin of 0.82 runs, which falls comfortably inside the 1.5-run run line. So the model's formal pick is Washington +1.5, and mathematically that's coherent: a projected margin under a run means the favorite rarely covers.
But look at the expected value. The run line is priced at -9.04 and the total at -8.27. Both are negative. Translating a coin-flip margin into a bet at a juiced price is how you lose money slowly, and the model is telling you exactly that. A 0.82-run projection against a 1.5-run line is the definition of no separation.
The only side with positive expected value is the Mets moneyline at +0.41 — and +0.41 is a whisper, not a shout. It's a thin, marginal edge on a game the model essentially grades as a toss-up.
Bullpens, the TBD, and the Injury Ledger
Washington's injury report is the more damaging of the two. Six pitchers are on the IL — PJ Poulin, Connelly Early, Ken Waldichuk, Mitchell Parker, Orlando Ribalta and Matt Waldron — five of them on the 60-day list. Catcher Drew Millas is also on the 60-day. That is a bullpen stripped to its skeleton behind a starter who, fortunately, has been pitching deep into games.
New York's ledger shows five pitchers (Reed Garrett, Tylor Megill, Daniel Duarte, Christian Scott, Robert Stock) plus first baseman Jared Young (7-day) and DH Jorge Polanco (60-day). Our research brief separately flags an ankle injury that forced Arraez out of a game in Philadelphia, with his availability for this series listed as a question mark — and notably, that name does not appear on the official IL ledger above. That gap between the beat report and the transaction log is worth watching at lineup release.
The Mets' starter is listed as TBD, which is the single largest unknown in this game. If it becomes a bullpen game, New York is asking a relief corps already missing four arms to cover nine innings — in a game it may need to win to keep its season alive.
The Card
- Moneyline — New York Mets (-134): The model's only positive-EV side at +0.41. This is a lean, not a conviction play. Smallest stake, or pass entirely if you demand real separation.
- Run line — Washington +1.5: Model's pick, but -9.04 EV. The number is the problem, not the side. Pass.
- Total — Over 7.5: Model's pick, but -8.27 EV after a full-run market move. Pass.
Bottom Line
The public will see a road favorite at -134 against a 75-84 team and lay the Mets. They'll be right about the side and wrong about the price — and badly wrong if they compound it by taking New York on the run line.
Earl's model hands you a game it grades as a near coin flip, projects a margin of 0.82 runs, and flags exactly one market with a sliver of value. That's not a spot to force action. It's a spot to recognize that the sharpest bet on this board was made before you ever opened the app — the money that dragged this total from 8.5 to 7.5.