The Rockies land in the Bronx on Tuesday night carrying one of the worst records in the sport, and the Yankees hand the ball to a pitcher in the middle of a season Yankee Stadium hasn't seen in a while. Colorado (55-88) at New York (81-62), first pitch 7:05 p.m. ET, is the kind of matchup where the numbers and the eye test agree — but agreement is not certainty, and it's worth walking through exactly why the model lands where it does.
On the mound: RHP Gabriel Hughes for Colorado against RHP Cam Schlittler for New York. Hughes enters at 0-6 with a 6.19 ERA and 42 strikeouts. Schlittler enters as an AL Cy Young candidate — 12-6, 2.04 ERA, 210 strikeouts across 171 2/3 innings, and a 0.94 WHIP. His September 2 start against the Angels (eight innings, one run, two hits, no walks, nine strikeouts) was his 19th outing of 2026 allowing one run or fewer, the most such starts in a single season in Yankees franchise history.
The number
Our model prices the Yankees at an 85% probability to win this game. The market-implied number is 72%. That leaves an edge of 13 points in favor of the pick — NYY on the moneyline. Model confidence sits at 85.
The edge is meaningful but not extreme, and the honest framing matters: an 85% line still means the pick loses roughly 15 times in 100 across a large sample. This is a favorite with a real chance to fall, not a formality.
Why the model leans this way
Start with the offensive gap. Our model rates the Yankees' offense at 60.1 and Colorado's at 21.1 — a wide separation that drives most of the projection. Layer in the pitching mismatch: Schlittler's 2.04 ERA and 0.94 WHIP against a Colorado club that ranks 30th in ERA (5.47) and 30th in WHIP (1.52). The Rockies also arrive worn down, having dropped 13 of their last 20, and their staff is thinned — Chase Dollander (right elbow sprain) and McCade Brown (shoulder) are both on the 60-day IL.
The totals picture reinforces the read rather than fighting it. Our model total is 8.0 against a market total of 8.5, a modest Under lean. A projected lower-scoring game generally favors the side with the superior arm, because run-suppression compresses variance in a favorite's direction. The park factor here is 1.04 — very slightly hitter-friendly, and already accounted for in the projection.
Form is close to a wash at the top line: both clubs enter having lost one straight. The difference is context. New York's single loss sits against an 81-62 record and a hold on the top AL Wild Card spot; Colorado's sits inside a stretch that made them the first team eliminated from playoff contention for the second consecutive season. The model reads those as very different one-game losses.
The case against
Baseball's variance is the first and largest threat to any moneyline favorite, and this is a single game. Colorado does not need to be better than New York to win once; it needs one crooked inning against Schlittler and a quiet night from the Yankees' lineup. Hughes' 6.19 ERA describes a body of work, not a ceiling — a struggling starter can still string together five or six competent innings, and if he does, a low-scoring game (recall the model's 8.0 total lean) actually narrows the margin the favorite has to work with.
Bullpen sequencing is another live risk. If the game stays close into the late innings — precisely what an Under-leaning script can produce — a single Colorado run against New York's relief corps can flip the outcome, regardless of how the starters pitched. Favorites priced near 85% still lose because the last three outs are the hardest, and a well-timed Rockies rally needs no scouting edge to happen.
There's also the plain fact that Colorado is a major-league roster with major-league hitters. A team ranked 21.1 on offense in our model can still put up a four-run night on the right evening; September lineups facing an eliminated club can also invite an off-script performance on the favorite's side. None of this changes the direction of the edge — the model still favors New York — but it's why the number is 85 and not higher. Roughly one game in seven, the side we've argued for here simply doesn't come through.
Weigh the 13-point edge honestly against that variance before deciding what it's worth to you.
Predictions carry risk; only stake what you can afford to lose, and if it stops being fun, call 1-800-GAMBLER. 21+.