Mariners vs. Astros — Sept. 23, 2026, T-Mobile Park (game 2 of 2)
The Houston Astros walked into T-Mobile Park on Tuesday and made a statement, blanking Seattle 7-0 in the opener of this brief two-game set. Now comes the finale — Wednesday, 10:10 p.m. ET — and while the scoreboard drama may not match the stakes narrative, the betting board absolutely does. Earl's model has a clear lean, and it runs against the team wearing the home whites.
Here's the lay of the land:
| Team | Record | Home | Away | RS/G | RA/G | Run Diff |
|---|---|---|---|---|---|---|
| Seattle Mariners | 73-84 | 40-36 | 33-48 | 4.04 | 4.44 | -63 |
| Houston Astros | 78-79 | 39-42 | 39-37 | 4.43 | 4.76 | -52 |
Houston's clubhouse message, per beat reporting out of the series, is simple: "it's go time." The Astros just surrendered the AL West lead and are publicly framing this as a must-win stretch. Seattle, eliminated from realistic contention weeks ago, is playing out the string — and the price doesn't reflect that gap in motivation.
Kirby's Traffic Problem
George Kirby (9-11, 4.24 ERA, 1.39 WHIP) is the name the public recognizes, and that recognition is doing a lot of work in this line. The underlying numbers tell a different story than the reputation.
Kirby has thrown 169.2 innings across 29 starts, allowed 191 hits — a hit rate of 10.13 per nine — walked 44 and struck out 143. His control is fine (2.33 BB/9), and his strikeout-to-walk ratio of 3.25 is respectable. But 7.59 K/9 against 10.13 H/9 is the profile of a pitcher who is putting the ball in play a lot, and in a season where he's also surrendered 21 home runs, that's a lot of traffic for a defense to erase.
Look at the recent run of team results in his starts: 7-2 over the Angels, then 5-6 and 2-6 against Oakland, 8-9 against Boston, 4-1 against Philadelphia. The Mariners went 2-3 in those five, but three of the five saw Seattle's opponent reach six runs or more. Whatever Kirby's personal line was in those games, the games themselves were high-variance, run-heavy affairs.
Pecko: Small Sample, Long Layoff
Houston counters with rookie right-hander Ethan Pecko (1-0, 3.77 ERA, 1.12 WHIP) — and here's where this matchup gets genuinely interesting.
Pecko's surface numbers are better than Kirby's across the board. He's allowed just 20 hits in 28.2 innings (6.28 H/9) and sports a WHIP 27 points lower than the veteran he's opposing. But the profile is not clean: 5.97 K/9 is below average, 3.77 BB/9 is elevated, and his strikeout-to-walk ratio of 1.58 reflects a pitcher who isn't missing bats and is giving away free passes. He's also allowed five home runs in under 29 innings.
Two things matter more than any of that. First, this is a four-start sample — we are not dealing with established data, we're dealing with a scouting report still being written. Second, his most recent start came September 5 against Arizona. By Wednesday night that's an 18-day gap between outings. That could be workload management, a rotation squeeze, or a deliberate rest period — but from a handicapping standpoint, it's an unknown. We are guessing at sharpness.
And here's the counterweight the sharp side has to acknowledge: the Astros went 3-1 in Pecko's four starts while he personally earned one win. That tells you run support and bullpen work carried the results, not dominance on the mound.
The Park and the Bullpens
T-Mobile Park plays to a 97 overall park factor with a 91 home-run factor, and the dimensions explain why: 380 to right-center, 401 to center, a cavernous power alley that has suppressed offense since 1999. That's a meaningful factor for a rookie who has allowed five home runs in a small sample — the park gives him margin for error that a place like Minute Maid would not.
Both bullpens, meanwhile, are battered. Seattle is without Matt Brash, Emerson Hancock, Nick Davila, Cole Wilcox and Logan Evans — five arms on the injured list. Houston is missing Steven Okert, Enyel De Los Santos, Mike Burrows and Brandon Walter. Neither manager has a fully armed relief corps, which raises the probability of a late-inning run or two and is why the model's total, while low, isn't suppressed even further.
Seattle's lineup depth is also thinned: Luke Raley, Will Wilson, Brendan Donovan, Brennen Davis and Colt Emerson all appear on the injury report. Houston's notable absence is Carlos Correa on the 60-day list.
Earl's Model Says: Back Houston
Our model's full card:
| Market | Pick | Model Projection | Expected Value |
|---|---|---|---|
| Moneyline | HOU +124 | HOU 3.9 – SEA 3.7 | +27.07 |
| Run Line | HOU +1.5 | Margin: -0.17 (away) | +21.42 |
| Total | Under 8.5 | 7.61 runs | -9.89 |
Start with the margin: -0.17 runs in favor of Houston. That is a coin flip on the field. The model is not telling you the Astros are meaningfully better in this specific matchup — it's telling you this game is a near-even proposition, and the market is pricing it otherwise at Houston +124.
That's the entire case for the moneyline. You are not buying conviction on the Astros' roster; you are buying a price edge on a toss-up. When the true line is even and you're getting plus-money, you take it. An EV of +27.07 is the strongest number on the board.
The run line at +1.5 is the higher-percentage construction of the same idea, and the model's +21.42 EV reflects it. You collect if Houston wins by any margin, or loses by a single run. In a game the model projects as a sub-run decision, that extra run and a half is worth real money.
The total is where discipline matters. Yes, 7.61 is 0.89 runs under the posted 8.5, and yes, the park, the pitching matchup and the retractable roof all point down. But the model's own EV on this market is negative at -9.89 — meaning the current price doesn't clear the bar. The under is the correct side; it is not the correct bet at 8.5. Don't force it. The two Houston markets are where this card lives.
Reading the Flat Line
Opening and current numbers are identical across every market: Mariners -146, Astros +124, run line -1.5, total 8.5. Nothing has moved.
That's informative. Either the opener was efficient enough that neither sharps nor public saw a reason to push, or Tuesday's 7-0 Houston win was already priced into the market before it happened. Either way, you are not fighting steam — you're getting the number the market settled on, which is exactly the environment where a model edge in the 20-plus EV range is most actionable.
Where the Public Is Wrong
The public sees "George Kirby" and a home favorite at -146 and assumes quality. The 2026 version of Kirby is a 4.24 ERA pitcher with a 1.39 WHIP and 191 hits allowed in 169.2 innings. Name recognition and home-field bias are inflating this price.
The public also underweights venue symmetry. Houston is 39-37 on the road; Seattle is 40-36 at home. These are functionally equivalent splits. There is no park advantage here to justify a 146-to-124 gap on the moneyline.
Where the public has a legitimate point: Houston allows 4.76 runs per game against Seattle's 4.44, and Pecko's sample is 28.2 innings of noise. That's real, and it's precisely why the model's projected margin is a fraction of a run rather than a full one. It's also why Houston +124 — not Houston at a short price — is the play.
Bottom Line
Earl's model lands on Houston +124 as the primary play (EV +27.07), with Houston +1.5 as the preferred run-line construction (EV +21.42) for anyone who wants the extra run in a game projected as a near-tie. The Under 8.5 is a directional lean backed by T-Mobile's 97 park factor and a 7.61 projected total, but with negative EV at the current number, it belongs in the "small or pass" column, not on your main card.
The motivated team is the road team. The price says otherwise. That's the edge.