One model, 14 games, three spots that cleared the bar on edge and price. Columbus moneyline tops a board heavy on overs.
Fourteen games got run through the model today. Three cleared the gate — 50% confidence on the side, 2% expected value minimum, both conditions or nothing. Only one hit the top tier.
Columbus Blue Jackets moneyline at -102 in Toronto is the play, and it's the only one graded BET rather than LEAN. The model has Columbus winning this one 52.9% of the time. The price only needs 50.5% to break even, so the gap — 4.8% EV — is real room, not a coin flip dressed up as a lock.
The factors lean Columbus's way across the board, though none of them is a landslide on its own. Home injuries sit at -0.64, a weight the model counts against Toronto. Home high-danger rate is off at -0.15, another drag on the Leafs, while Columbus's own high-danger number comes in at +0.15 in their favor. Toronto's special teams mark is down at -0.125 as well. Stack those together and you get a model that likes the road side more than the market does, at a number that's basically a coin flip itself.
Two totals cleared as LEANs, both overs, both at 6.5, both with less margin than the Columbus play but still enough to clear the 2% floor.
Pittsburgh at Nashville over 6.5 at -102 has the model at 52.2% against a required 50.5%, good for 3.4% EV. Washington at Boston over 6.5 at -115 needs more lifting — 53.5% required — and the model's 54.6% projection clears it by a thinner 2.1%. Both are real plays by the model's own math, just not plays we'd size the same as the Columbus line.
Expected value here is measured against the real, bet-it-now price — not some theoretical vig-free number. That distinction isn't cosmetic. A play can sit a hair off fair value and still lose money once the book's standard 2-2.5 points of margin on an NHL moneyline gets baked in. Clearing 2% EV at the actual price is a higher bar than it sounds.
The model flags its own gaps on the top play: goalie starters aren't confirmed yet, the model disagrees with the book's puck-line favorite, and there's no observed GSAx data for either goaltender involved. That's not a reason to ignore the number, but it's a reason to hold it loosely.
This is one model's read on a small sample of games, not a demonstrated long-term edge and not financial advice. Three plays cleared out of fourteen boards today — a normal day, not a thin one, just a day where most of the slate didn't offer enough price to bother.