The Colorado Rockies visit Yankee Stadium on Wednesday, September 9, 2026, for the middle game of a three-game series, and the gap between these two clubs is wide on paper. Colorado enters at 55-89, fifth in the NL West and long since eliminated from playoff contention. New York sits at 82-62, second in the AL East and pushing for the first American League wild-card spot while holding off Boston. The listed starters are RHP Will Warren for New York against RHP Tomoyuki Sugano for Colorado.
The number
The pick is New York on the moneyline. Our model prices the Yankees at 82% to win. The market-implied probability is 70%. That leaves an edge of 12.5 points in our favor. Model confidence on this game reads 83.1. Those are the exact figures, and the entire distance between our number and the market's number is that 12-point spread in win probability — nothing more, nothing less.
Why the model leans this way
The core of this lean is the offensive mismatch. Our offense ratings put New York at 68.5 and Colorado at 43.5 — a 25-point separation that is among the largest in any matchup on the board. The park factor here is 1.04, marginally above neutral, which is one reason our model total sits at 10.5 against a market total of 8.5, an Over lean. When we project run-scoring environment north of the market and pair it with a large offensive edge for the home side, the moneyline moves in New York's direction.
Form points the same way, if modestly. The Yankees have won 1 straight; the Rockies have lost 2 straight. Colorado has been near the bottom of the standings all year and has played roughly 4-6 baseball over its last 10 games.
The pitching line is more balanced than the records suggest, and that matters for how we frame this. Sugano is 12-8 with a 5.19 ERA; Warren is 9-6 with a 4.16 ERA. Warren's most recent outing was 2.1 scoreless innings against the Padres, and he carried a 4.62 ERA across five August starts — solid, not dominant. On earned-run average alone, Warren has the cleaner profile, and that supports the model's confidence in New York rather than undercutting it. The lean is not a claim that either starter is elite; it is a read that the run-scoring environment and the offensive gap favor the home team by more than the market has priced.
The case against
This is a moneyline on a heavy home favorite, and the honest risks are real. First, Sugano has history with this opponent. He started four games against the Yankees last season with Baltimore and allowed 11 runs across those starts — a small sample, but not a clean one, and it is fair to flag a starter who has already faced this lineup multiple times. If he navigates the New York order the way a veteran with familiarity sometimes can, the projected offensive edge shrinks quickly.
Second, an Over lean cuts both ways for a favorite. Our model total of 10.5 implies a high-scoring game, and high-scoring games invite variance. A single bad inning from Warren — or a bullpen wobble in a game the Yankees have played on a West Coast trip that ended in a 3-3 stretch against the Angels and Padres — can turn a comfortable lead into a coin flip. Warren's 2.1-inning last outing also means his workload and length are worth watching.
Third, and most simply, Colorado is a major-league team with a professional lineup, and roughly one game in five ends with the underdog winning outright even when the favorite is correctly priced. Our number says the Rockies win 18% of the time. That is not a rounding error; that is close to one game in five. A short outing, a defensive miscue, or ordinary bullpen variance in a projected high-scoring environment are the everyday ways this pick loses. None of that would be a surprise, and none of it changes the discipline of playing the edge rather than the outcome.
Weighed together, the model's 82% against the market's 70% is the reason this appears on the board. The edge is 12.5 points. The confidence is 83.1. Those figures, and the honest 18% chance of a loss, are the full picture. Our complete record sits on a public, tamper-evident ledger; this pick joins it either way.
Predictions carry risk and can lose. Please make decisions responsibly. 21+.