The cleanest mispricing on the board this week isn't between two platforms. It's sitting on a single page.
BetMGM's Bears-Seahawks preview lists Seattle as a -3 favorite with a moneyline of -175, which implies roughly 63.6% win probability once you strip the vig. Scroll down the same page and BetMGM's own simulation model gives the Seahawks only 51.3% confidence to win the game outright. That's a 12-point gap between what the house is charging you to bet Seattle and what the house's own projection system thinks Seattle's actual odds of winning are.
This is the kind of split that usually hides between platforms — Polymarket pricing one way, DraftKings pricing another, with a liquidity gap or a stale injury report explaining the difference. Here it's happening inside one shop. The market-facing line (-175) is built to balance two-sided action and absorb public money, which has been pouring onto Seattle after a statement win over Kansas City; the model number (51.3%) is a cleaner read on team strength without the vig or the crowd lean baked in. When a book's posted price runs that far ahead of its own stated model, it's usually a signal that the number is being set for bettor behavior, not for truth. That's useful information on its own, independent of any second platform.
The Spread Tells the Same Story, Quieter
The against-the-spread number from the same BetMGM page backs this up. The model gives Seattle just 50.1% confidence to cover its own -3 line — essentially a coin flip, the kind of number that should produce a spread priced near pick'em with standard vig on both sides, not a line that's collecting -175 on the moneyline side. A -3 spread with the house's own model sitting at 50.1% to cover isn't wrong, exactly — public teams that just blew out the Chiefs draw public money regardless of model edge — but it tells you where the inflation is. Fade the Bears number if you're inclined, but know you're betting into a line shaped by sentiment, not by the shop's stated projection.
Worth noting: DraftKings' Week 8 picks column covers this same Bears-Seahawks matchup but stops short of publishing a numeric line, leaning instead on narrative — "I'm not expecting the dominant performance we just saw against the Chiefs, but they'll make the Bears..." That's not a knock on DraftKings; it's a reminder that when you're hunting for cross-book deltas, you need two sides actually posting prices. Right now only BetMGM has given you both the market price and the internal model number on this game, which is exactly why the gap is visible there and nowhere else this week.
Where the Data Runs Thin
The bigger prediction markets aren't cooperating this week either. Polymarket's Chiefs-Seahawks contract shows real volume — $6.0 million traded — but that game already happened; Seattle beat Kansas City 20-3, per DraftKings' recap, so there's no live price to arbitrage against. The Notre Dame-Navy market is split across four separate contracts with volumes ranging from roughly $4,400 to nearly $82,000, which tells you the margin-of-victory markets are getting real action even if the win-probability price itself isn't visible in what's publicly rendered. That's a liquidity signal, not a pricing signal — useful for knowing where attention is concentrating, not for building a trade.
On the NBA side, FanDuel's model has the Lakers at -2.5 (-106) off a 4-1 start and a +9.6 scoring margin, with the line explicitly hedged on Luka Doncic's back. That's a single-book price shaped by a known injury variable, not a cross-platform gap — worth watching if Doncic's status firms up and the number moves before other books catch it.
The actionable piece this week is narrower than usual: BetMGM's own internal math says its Seahawks price is running hot. Watch whether that gap closes by kickoff Monday night, or whether the public money keeps pushing it the other way.