NHL Shots on Goal Props: How SOG Lines Are Set and How to Read Them
Shots on goal is the highest-volume prop in hockey and the one most driven by something you can look up before puck drop: ice time. Here is how the number is built, what three seasons of box scores say about each line, and where a model can and cannot help.
Published September 2026 · 9 min read
1. How Sportsbooks Set a Shots on Goal Line
A shots on goal prop starts from one number: the skater's recent shots per game. Books take a rolling average over roughly the last 5 to 10 games, adjust it for the minutes the player is expected to get, the power-play unit he is on and how many shots the opponent concedes, then round to the nearest half shot.
That rounding is most of the market. Hockey shot counts are small integers, so nearly every line in the book lands on 1.5, 2.5 or 3.5. A player projecting 2.6 shots and a player projecting 3.4 shots can be hung at the same 2.5, and the difference between them shows up in the juice rather than in the line. Reading the price is how you recover the projection the book did not print.
2. What Each Line Actually Hits
Across 141,677 skater games in the 2023-24, 2024-25 and 2025-26 seasons, this is how often a skater cleared each line:
| Line | Every skater game | Skaters averaging 2.5+ |
|---|---|---|
| Over 1.5 | 44.0% | 78.7% |
| Over 2.5 | 22.9% | 56.8% |
| Over 3.5 | 10.9% | 35.9% |
The two columns are the whole lesson. Over every skater who dressed, the 2.5 over hits 22.9% — a number that would make the market look unmissable from the under side. Restricted to the 316 player-seasons who actually average 2.5 shots or more, the same over hits 56.8%. Books do not post a 2.5 on a fourth-line winger, so the first column is not the population you are betting into. Any base rate quoted without the population attached is worth nothing.
Position is the next filter. Shots follow role, and role follows position:
| Position | Shots per game | Clears 2.5 |
|---|---|---|
| Right wing | 1.88 | 29.7% |
| Left wing | 1.76 | 26.6% |
| Centre | 1.67 | 24.8% |
| Defence | 1.31 | 16.0% |
3. Ice Time Is the Input, Not Form
The instinct is to bet shots on goal off scoring form. The data says to bet it off minutes. A shot is a volume event: a player who is on the ice for 21 minutes with first-unit power-play time will generate more attempts than a player on 13 minutes, regardless of who has been scoring.
That is why the things worth checking before a shots prop are the ones that move minutes — line promotions after an injury, a return from the press box, a new power-play unit, and the second night of a back-to-back, when coaches shorten star shifts. None of these show up in a scoring streak, and all of them move the shot count.
4. The Opponent Side of the Number
The other half of a shots line is who is trying to stop them. Teams differ substantially in shots conceded per game, and a skater drawing a team that concedes 33 shots is in a different game from one drawing a team that concedes 26. Shot suppression is also more stable across a season than goal prevention, because it depends on structure rather than on a goaltender's week.
The same fact is what makes the goalie saves prop work the way it does: the opponent's shot volume is the dominant term on both sides of the puck.
5. How Our Model Prices a Shots Line
Our SOG projection is an XGBoost regression on 18 features, trained on the same box-score archive the numbers above come from. The inputs are the ones the sections above argue for:
- Shot form: shots per game over the last 5 and last 10, the standard deviation of the last 10, and the season average and spread.
- Role: time on ice over the last 5, last 10 and the season, power-play goal rate, position, and games played.
- Context: home or away, rest days, a back-to-back flag, and the opponent's shots and goals allowed over their last 10.
The regression returns an expected shot count, and the probability for each line comes from a normal distribution around it whose width is the player's own recent variance, floored at 0.6 of his season spread so that a player on a quiet run does not get an artificially confident number. Held-out error is about one shot per game (MAE 1.01), with an R² of 0.20 — meaning the model explains a fifth of the game-to-game variance in shot counts, and the rest is the part of hockey nobody projects.
6. What We Will Not Claim
Our NHL player props have 40 graded picks on the shots market. That is a sample, not a record, and we will not print a hit rate off it — the same rule that keeps us from quoting our goalie or assists numbers until they have a season behind them. What we do publish, and grade in public, are the NHL team markets: 64.4% on the puck line over 261 picks and 60.3% on the moneyline over 58.
A projection that is right to within one shot per game is useful context for a line. It is not, on its own, evidence of an edge over the price, and the honest version of that sentence is the one we would want to read. Why most prediction models fail covers the rest of that argument.
7. Frequently Asked Questions
How do sportsbooks set NHL shots on goal lines?
Books start from the skater’s recent shots per game over roughly his last 5 to 10 games, adjust for ice time, power-play role and the opponent’s shot suppression, then round to the nearest half shot. Almost every line lands on 1.5, 2.5 or 3.5, so the rounding itself decides which side carries the juice.
How often does the over 2.5 shots on goal hit?
Across 141,677 skater games over three NHL seasons, a skater cleared 2.5 shots in 22.9% of them. That base rate is misleading on its own: among the regulars who average 2.5 or more shots, the same over hits 56.8%. Books only post the line on players in the second group.
Do defencemen get shots on goal props?
Sometimes, and usually at 1.5 rather than 2.5. Defencemen average 1.31 shots per game against 1.88 for right wings, and clear 2.5 in only 16.0% of games. Power-play quarterbacks are the exception, because point shots from the top unit are a volume role rather than a scoring one.
Does a back-to-back affect shots on goal props?
It affects them through ice time rather than through tiredness. On the second night a coach is more likely to shorten a star’s shifts and rest power-play minutes, and shot volume follows minutes. Our model carries rest days and a back-to-back flag as separate inputs for exactly this reason.
Are shots on goal props easier to predict than goals?
They are more stable, which is not the same as easier to beat. Shots are a volume statistic driven by role and ice time, so they repeat; goals depend on finishing, which does not. A model can project shots to about one shot of error and still not clear the price the book charges.
See tonight's shots on goal projections for every skater
Projected shot counts and over/under probabilities at 1.5, 2.5 and 3.5, rebuilt three times a day from ice time, role and the opponent's shot volume.
Start your free 3-day trialThe week ahead, once a week
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.