Two low totals clear the model's bar tonight, with Edmonton-Detroit grading out as the stronger of the pair.
Three games got run through the model today. Two cleared the bar. That bar isn't soft — a side has to grade at 50% or better and still return at least 2% expected value at the actual number on offer, not some theoretical fair line. When both conditions line up, you get a play. When they don't, you get nothing, and we're not going to dress up a game we don't like just to hit a quota.
Tonight both plays that cleared happen to be totals, and both happen to be unders at plus-money. Red Wings at Oilers under 6.5 at plus-110 is the stronger of the two by expected value. Maple Leafs at Flames under 6.5 at plus-100 is right behind it, actually grading higher on model confidence even if the dollar return is a touch smaller.
This is the top play on the board, and it's rated STRONG. The model has the under at 54.6% to hit, against a price that only needs 47.6% to break even. That's a 14.7% expected-value edge at plus-110 — the kind of gap that's worth paying attention to even before you get into the underlying factors.
On the injury side, the model has Edmonton dinged more than Detroit: home injuries sit at -0.48, which hurts the Oilers, while away injuries at -0.26 hurts the Red Wings — a smaller drag, but a drag. High-danger rate cuts a different way. Edmonton's figure of +0.15 helps the home side, while Detroit's -0.15 hurts the away side. Two factors working against each team, two working in their favor depending on which column you're reading — nothing here points cleanly in one direction for who wins. What it does support is the total: a roster facing more health trouble and uneven high-danger numbers on both sides is consistent with a model leaning toward fewer goals, not more.
Worth flagging before you bet it: the goalie starters aren't confirmed, and there's no observed GSAx on either netminder yet. That's a real gap in the picture, not a footnote.
This one actually carries the higher model probability of the two plays — 56.0% against a price that needs just 50% — but the expected value comes in a bit lower at 12.1% because the number is flat plus-100 instead of plus-110. Tier ELITE on confidence, but EV is EV, and that's why it slots second rather than first.
No factor breakdown was supplied for this one, so we're not going to manufacture a narrative around injuries or shot quality that the model didn't hand us. Take it for what it is: a total the model likes clearly, at a price that clears the hurdle.
Expected value here is measured against the real number you'd actually get paid at, not some theoretical no-vig line. That distinction matters — a play can sit a hair off the true fair price and still lose money once the book's margin, typically somewhere around 2 to 2.5 points on an NHL moneyline, gets baked in. Clearing the house's cut is the whole game.
Both plays reflect one model's output on a limited sample tonight. This isn't a demonstrated long-run edge and it isn't financial advice — it's what the numbers say right now, goalie uncertainty and all.