Sports Edge Model — every game called, every call graded, in public

San Francisco Giants at Pittsburgh Pirates

MLB · Thursday, September 3, 2026 · 12:37 PM

The model's call
Pittsburgh Pirates to win
Model confidence 60.3%
Books consensus 60.3%
Identical on purpose: the model applies no adjustment to the books’ de-vigged consensus until a candidate signal proves itself in shadow audit — none has yet. Prediction markets below are a comparison, not an input.

Closing moneyline by book

The last price each book posted before first pitch — the prices this call could actually have been bet at. Best on each side is highlighted. In-game odds are deliberately excluded: once the game starts, the price tracks the score, not the matchup.

Book San Francisco GiantsPittsburgh Pirates
BetMGM +140 -170
BetRivers +145 -175
DraftKings +141 -171
ESPN BET +149 -169
Fanatics +140 -170
FanDuel +140 -166
Hard Rock Bet +135 -165
Caesars +154 -164

Prediction markets vs the books

Every venue sells the same thing — a dollar if that side wins — at a different price. Implied cost per venue below; amber = the cheapest way to buy that side at the close.

Venue San Francisco GiantsPittsburgh Pirates
Books consensus fair, de-vigged 39.7% 60.3%
Best book price incl. vig 39.4% 62.1%
Kalshi contract price 39.5% 60.5%
Polymarket contract price 39.5% 60.5%
Markets average 39.5% 60.5%

Contract prices exclude exchange trading fees (Kalshi ~1–2¢ midrange), which narrow small gaps. A venue meaningfully below the books consensus is either a bargain or an early opinion — the graded record decides which.

Trade the event contracts
The prediction-market prices on our game pages are tradeable markets.
Referral links — we may earn a reward if you sign up. 18+. Event contracts involve risk; this is not investment advice.

Back to today's board · The full record