Tigers vs. Rockies — Sept. 13, 2026 (series finale, Comerica Park)
Earl's model says the biggest number on this page is a lie. Not the total, not the moneyline — the 225-run gap between these two clubs in run differential. Detroit is +63 on the season. Colorado is -162. On paper, that's a mismatch. In the model, Sunday's game at Comerica Park projects as Tigers 4.7, Rockies 4.0 — a 0.78-run margin, which is the same thing as saying this is a coin flip that happens to be leaning toward the home dugout.
That tension is the entire handicap. And it's why our model's sharpest edge on the board isn't the side anyone is talking about.
What Earl's Model Sees
| Market | Model Pick | Model Number | EV |
|---|---|---|---|
| Run line | Colorado +1.5 | Home margin 0.78 | +6.15% |
| Total | Over 8.5 | 8.69 runs | -6.62% |
| Moneyline | Detroit | — | -11.56% |
Read that table the way a pro reads it. The model's only genuine value play is Colorado +1.5. The other two picks are directional leans the model would not pay a premium to hold — a projected total just 0.19 runs above the number, and a favorite priced at a number the model can't defend.
So how does a team that's been outscored by roughly 1.1 runs per game all season get within 1.5 runs here? Two reasons, and neither of them is "the Rockies are secretly good."
The Tigers Are Starting a Short-Start Arm
Jackson Jobe's season line is modest, but the number that matters most for a run line is innings: 28.1 IP across 6 starts, an average of 4.7 innings. He's 1-2 with a 4.45 ERA, a 1.41 WHIP, and a 2.17 K/BB — competent for a young arm, and nowhere near the kind of length a -1.5 favorite needs from its starter.
Look at his last five starts and a pattern jumps:
| Date | Opp | DET | Opp |
|---|---|---|---|
| 9/6 | CLE | 2 | 3 |
| 8/31 | MIN | 1 | 11 |
| 8/25 | TB | 4 | 1 |
| 8/19 | PIT | 3 | 4 |
| 8/14 | CWS | 5 | 9 |
Detroit has scored 15 runs across five Jobe starts — 3.0 per game. That's the quiet engine behind the +1.5 edge. Laying a run and a half requires an offense that can build a two-run cushion; when your starter's average game script is a 3-4 final, you're asking for a save situation instead of a comfortable win. The Tigers are 38-36 at home, a .514 clip — nothing about that home record suggests a blowout profile.
Gabriel Hughes: The 0-7 Record Is Not the Story
The public will see 0-7 and a 5.81 ERA and assume the Rockies are dead. The model doesn't. Hughes' 1.38 WHIP is actually better than Jobe's 1.41, and his 9.1 hits per nine and 3.3 walks per nine are livable rates. He's been sabotaged by sequencing (five home runs in 57.1 innings) and by short outings that leave a thin bullpen exposed — he averages 5.2 innings per start.
His run support has been bizarre, and it cuts against the lazy read:
| Date | Opp | COL | Opp |
|---|---|---|---|
| 9/8 | NYY | 3 | 5 |
| 9/1 | BAL | 4 | 2 |
| 8/27 | WSH | 1 | 7 |
| 8/22 | CLE | 3 | 4 |
| 8/16 | SF | 13 | 7 |
Colorado has scored 24 runs in his five most recent starts — 4.8 per game. That's more support than Jobe has received from a vastly better lineup. Hughes hasn't won because he's been the guy tagged in the middle innings, not because the offense abandons him. For a +1.5 ticket, that's precisely what you want: a pitcher who keeps a game in range and a lineup that has been producing behind him.
The Injury Ledgers Are Both Ugly — and That Favors the Dog
Detroit's IL is a roster hole, not a list. Three center fielders (Matt Vierling, Parker Meadows, James Outman), two corner outfielders (Kerry Carpenter, Wenceel Pérez), shortstop Trey Sweeney, and six pitchers including Reese Olson, Justin Verlander, Will Vest, Jack Flaherty, Burch Smith, and Bailey Horn.
That matters to Sunday specifically. Jobe is a 4.7-inning pitcher. Behind him, the Tigers are covering roughly four innings with a pitching staff missing six arms. Colorado, meanwhile, has its own attrition — Willi Castro, Kris Bryant, Tyler Freeman, plus Kyle Freeland, Sean Sullivan, Jeff Criswell, Jose Quintana, Chase Dollander, McCade Brown, Pierson Ohl, and Welinton Herrera.
Two depleted staffs, two starters averaging under 5.5 innings, and a park with a 104 home run factor is why the model spits out 8.69 runs against an 8.5 line. The Over is the correct side, but at -6.62% EV it's a lean, not a bet. If 8.5 is the number, the model says the number is right.
The Line Movement Is the Tell
Detroit opened at -168. It's now -184. That's 16 cents of juice added to a team that is 69-79 overall, has scored 4.45 runs per game, and has scored 3.0 per game behind Sunday's starter. Colorado drifted from +142 to +154 on a 25-52 road record.
That's the public laying the favorite against the worst record in baseball — the single most predictable reflexive bet on the board. It's also the wrong way to express Detroit's edge. At -184, you need to win 64.8% of the time to break even. A model that projects a 0.78-run margin is telling you Detroit is the better team, and still telling you not to pay that price. The model's moneyline EV of -11.56% is the loudest number on this page.
Meanwhile, the +1.5 for Colorado — with a run differential that says blowout and a model that says one-run game — carries a +6.15% EV, the only actionable edge here.
The Betting Card
- Primary: Colorado Rockies +1.5 (EV +6.15%). The model's highest-conviction play and the one that survives the line movement rather than fighting it.
- Lean only: Over 8.5. Two sub-5.5-inning starters, two thin bullpens, a 104 park HR factor. But at -6.62% EV, this is a small-stake lean, not a play. The number is fair.
- Pass: Detroit moneyline. The model's side, at a price the model can't support. Take the win probability, not the tax.
- If you're on Detroit anyway, the run line is the wrong vehicle. You're asking a 4.7-inning starter to protect a two-run cushion with six arms on the IL.
The Bottom Line
The Rockies are bad. That's not in dispute — -162 in run differential and a 55-93 record don't lie. But "bad team" and "bad bet" are different concepts, and the market has spent all week conflating them. Detroit's 225-run advantage in differential is a season-long fact that has almost nothing to do with what happens over nine innings between Jackson Jobe and Gabriel Hughes.
Earl's model sees a 0.78-run game, a total that's correctly priced, and a favorite the public has juiced past the point of value. The edge isn't on the better team. It's on the run and a half — and that's where the money should go.