Projected shots on goal for every skater on the slate, with the probability of going over or under 1.5, 2.5 and 3.5. An XGBoost regression projects the mean from his recent shot volume and ice time and the opponent's shots allowed; his own game-to-game spread sets the width.
The over/under lines on this page are the model's own ladder, not a sportsbook's — compare the projection to the number your book is hanging. Every badged pick is graded against the box score the next morning on the performance page.
Projections run three times a day once the regular season opens, for every skater expected to play. Get the board in your inbox the morning it opens.
Last week's graded record — wins and losses — the plays for the week ahead, and a subscriber discount code. One email, no daily noise, unsubscribe in a click.
Shots on goal follow ice time. A top-line winger on the first power-play unit gets four or five looks a night; a third-pair defenseman gets one. The model starts there: shots and time on ice over the skater's last five and ten games, his season averages, and his recent goals and hits. It adds home or away, days of rest, back-to-backs, and the opponent's shots and goals allowed per game over their last ten.
That gives a projected mean. The width comes from the skater's own spread over his last ten games, floored at 60% of his season spread, and a normal distribution gives the over and under probability at each line. A skater needs ten games played before he gets a projection.
Each card shows the projection, the over/under split at the line nearest to it — roughly where a book would hang its main number — and a bar for each of his last ten games, green where he cleared that line. A 3.1 projection with seven green bars is a different bet from a 3.1 with bars that alternate, even when the probability is the same. Open "Every line" for the full ladder.
Ignore the lopsided rungs. A fourth-line defenseman will always show a high probability under 3.5; that is arithmetic, not an opinion. The useful signal is a projection that sits well away from the number your book is hanging.
The top two shots-on-goal picks of the day at 65% or better are graded against the box score the next morning, win or lose, and the running record sits on the performance page. Last season produced a few dozen graded SOG picks — too few to quote a hit rate honestly — so this page does not quote one. The 2026-27 record starts with the opener.
The model projects a mean from the skater's shots and ice time over his last five and ten games, his season averages, rest, back-to-backs and the opponent's shots allowed. His recent game-to-game spread sets the width, and a normal distribution gives the over and under probability at 1.5, 2.5 and 3.5.
Ice time and power-play role. A player who logs 20 minutes and sits on the first power-play unit has a far higher floor than one playing 13 minutes at even strength. Opponent style matters next: teams that allow a lot of shots raise every skater's projection against them.
No. The lines on this page are the model's own ladder at 1.5, 2.5 and 3.5 shots. No sportsbook line is archived for this market, so compare the projection with the number your book is hanging. A projection well above or below the book line is the signal, not the probability alone.
Because the ladder includes lines far from their projection. A defenseman projected for 0.8 shots will always show a very high probability under 2.5 or 3.5; no book offers that bet at a price worth taking. Read the probability at the line nearest the projection instead.
Shots on goal, goalie saves, goals, assists, points and hits, plus live in-game signals. MLB, NFL and NBA included.
Start your free 3-day trial