Prop LabPuck line and moneyline pricing
The NHL Puck Line +1.5, Explained With Nine Seasons of Margins
The puck line is always ±1.5, but the real margin of a hockey game is not. We modelled every second of regulation, priced 2,579 games and tested 15 rules written in advance. The market is close to right, with one small, steady lean.
By Edgehalla ResearchPublished 8 min read
What the puck line is
The puck line is hockey’s point spread, and it almost never changes. The favourite is −1.5 and the underdog is +1.5. To cover −1.5, the favourite must win by two or more. The underdog covers +1.5 if it wins, or if it loses by exactly one.
Because the line is fixed, the price does the work. A strong favourite might be +120 on −1.5, while the underdog is −145 on +1.5. The question the price answers is simple: how often will the favourite win by two or more?
A common thought is that some underdogs are “really” +0.8 and others +1.9, but the line still says +1.5. That is true. The real margin changes from game to game. So we asked whether the market’s ±1.5 price keeps up with it.
How hockey margins really fall
Over 2022-23 to 2025-26, 39.9% of games were decided by exactly one goal. Only 17.8% were one-goal wins in regulation. The other 22.1% went to overtime or a shootout.
That second number matters most. An overtime or shootout result is always a one-goal margin. So an OT loss always covers +1.5, and an OT win never covers −1.5. About one game in five is settled for the puck line before overtime even starts.
Final margins, with and without empty-net goals
Share of games by final goal margin
Without empty-net goals, 55 percent of games would be one-goal games. Empty nets push many of them to two or three goals.
- Final margin
- Without empty-net goals
Empty nets are the puck line’s hidden engine
Take empty-net goals away and 55% of games would be one-goal games. With them, it is 40%. Empty nets turn about 16–17% of all games from “won by one” into “won by two or more.” In 2024-25 and 2025-26, 28% of all two-goal-plus wins were created by an empty-net goal.
They also explain an odd pattern. Three-goal wins (23.6%) are more common than two-goal wins (19.9%). An empty-netter on a two-goal lead makes it three.
Empty nets are getting more common. The league averaged 0.28 empty-net goals per game in 2017-18 and 0.40 in 2024-25. Coaches pull earlier now: the median pull when down one moved from 1:40 left to 2:00 left. You can see every team’s pull timing on our goalie pull and empty-net page.
A one-goal lead with 3:00 left
A one-goal lead with three minutes left becomes a two-goal win 42 percent of the time.
That 42% is the whole puck-line story for close games, and it is fairly stable across teams. Team empty-net goal rates barely repeat (split-half r 0.19), so “this team always scores into the empty net” is mostly noise.
The season matters more than any angle
The share of games won by two or more swings from season to season more than chance allows. Its standard deviation is 2.0 points, against 1.4 points from pure luck.
Games won by two or more goals, by season
Share of all regular-season games
The share of two-goal-plus wins ranged from 56.8 to 61.9 percent; 2025-26 was the lowest in nine seasons.
This is the trap in most puck-line systems. Any rule that backs the underdog +1.5 looked bad in 2024-25 and great in 2025-26. The rules were not smart or dumb. The seasons were different.
Our margin model, and how the market compares
We built a margin model (we call it M1) that plays out regulation second by second. It knows the score, the clock and whether the trailing team has pulled its goalie. Its score effects, pull timing and empty-net scoring rates were fitted on 2017-18 to 2023-24.
For each game, we tune M1 so it matches the market’s own moneyline and total. Then we ask what it says the puck line should cost. That makes it a fair test: same inputs as the market, just a different view of margins.
Shape matters. M1 gives a 22–23% chance of a regulation tie, which matches the real rate of about 22%. A plain Poisson (Skellam) model gives only 16–17%. That gap is the classic reason simple models overprice the favourite −1.5.
How often does the favourite win by 2+?
Market no-vig price vs our margin model vs what happened, by moneyline band
The margin model sits on the actual rate in four of five bands; the market sits two to four points above it.
- Market (no-vig)
- Margin model (M1)
- Actual
Pooled over 2,579 games, the market said the favourite would win by two or more 37.9% of the time. M1 said 35.5%. The actual rate was 36.0%.
The gap is the same almost everywhere: 2.1 to 2.7 points for home and road favourites, back-to-backs, backup goalies, and totals of 5.5, 6.0 and 6.5. It looks like a flat shade on the favourite −1.5, the plus-money bet casual bettors like. It is not a blind spot in one corner.
But the vig on the +1.5 side is about the same size as that shade. Betting every underdog +1.5 returned −1.0% over the two seasons: −4.4% in 2024-25 and +2.4% in 2025-26.
Is the market efficient? Mostly, yes
On accuracy (Brier score, lower is better), it was a tie. The market won narrowly in 2024-25 (0.23847 vs 0.23879). M1 won narrowly in 2025-26 (0.21448 vs 0.21574), and that confidence interval only just excluded zero.
The market is also internally consistent. When the moneyline moved, the puck line moved 0.83 points per point; M1 says it should move 0.85. Goalie news, back-to-backs and injuries flow through to the puck line. It does not lag.
Fifteen rules, written before we looked
Before buying any odds, we wrote down 15 betting rules and a pass bar (a false-discovery q of 0.10 or lower). We tested them on 2024-25, froze them, then graded them once on 2025-26. None passed in the first season, so by our own rules none is an edge.
| Rule | 2024-25 (n, ROI) | 2025-26 (n, ROI) | Verdict |
|---|---|---|---|
| A1a: margin model 3+ pts above market → underdog +1.5 | 302, +3.9% | 179, +4.7% | Promising |
| Every underdog +1.5 | 1,296, −4.4% | 1,283, +2.4% | Myth (flips with the season) |
| Every favourite −1.5 | 1,296, −0.9% | 1,283, −16.4% | Myth |
| Heavy favourite (65%+) −1.5 | 291, −11.4% | 145, −11.6% | Myth (bad both years) |
| Aggressive-pull underdog coach → favourite −1.5 | 423, +7.4% | 396, −15.3% | Myth |
| Total 6.5+ → favourite −1.5 | 309, −1.7% | 459, −6.9% | Myth |
The one rule worth watching is A1a. When M1 rated the underdog +1.5 at least 3 points better than the market’s no-vig price, the bet returned +3.9% and then +4.7%. Pooled, that is +4.2% on 481 bets, winning 59.0% (95% range 54.6–63.3%) against a 56.6% break-even.
It also made money in 2024-25, a season that favoured favourites. That is the main reason it is on the watch list. But its q-values were 0.31 and 0.23, far from proof. Two seasons is a small sample for a 2-point effect, and A1a’s +4% could be half luck.
How to use this
- Think in margins, not winners. About one game in five goes to OT or a shootout, and every one of those covers +1.5.
- Check the goalie and the pull. Empty nets create 28% of two-goal-plus wins. A late goalie pull is the single biggest swing on −1.5. See starting goalies today.
- Shop prices. The best price across 11 books added 2–3 points of ROI to every rule. That is bigger than most of the edges themselves.
- Treat A1a as a watch list. On October 11, 2026 it flagged Vancouver +1.5 and Philadelphia +1.5. Those are tracked calls, not picks, and no live result exists yet.
Related reading: back-to-back betting covers the one other rule we are watching (A8), and the Prop Lab hub has every other hockey test.
Frequently asked questions
What does +1.5 mean on the NHL puck line?
The underdog gets a goal and a half. A +1.5 bet wins if the underdog wins outright or loses by exactly one goal. The favourite −1.5 needs a win by two or more goals.
Does an overtime loss cover the +1.5 puck line?
Yes, always. An overtime or shootout result is always a one-goal margin, so the losing team covers +1.5 and the winner never covers −1.5. From 2022-23 to 2025-26, 22.1% of games went to overtime or a shootout.
How often do NHL games end with a one-goal margin?
About 40% of the time. From 2022-23 to 2025-26, 39.9% of games were decided by one goal: 17.8% in regulation and 22.1% in overtime or a shootout. Without empty-net goals, it would be about 55%.
Is it better to bet the puck line or the moneyline?
Neither is free money. In our two-season test the market priced the favourite −1.5 about 2 points rich, but the vig on +1.5 was about as large, so betting every underdog +1.5 lost 1.0%. Pick the bet type by price, not by habit.
How do empty-net goals affect the puck line?
A lot. Empty-net goals turn about 16–17% of all games from a one-goal finish into a two-goal win, and they created 28% of all two-goal-plus wins in 2024-25 and 2025-26. A one-goal lead with three minutes left becomes a two-goal win 42% of the time.
Why is the favourite −1.5 usually plus money?
Because the favourite wins by two or more only about 36% of the time on average. Even favourites priced at 70% or more to win covered −1.5 only about half the time in our data (50.3%, 177 games).
More in Puck line and moneyline pricing
Sources and method
- The Odds API (historical NHL moneylines, puck lines and totals)
- NHL stats (NHL.com)
- NHL API reference (community documentation)
- Skellam distribution
- Brier score
- False discovery rate and the Benjamini–Hochberg procedure
- Wilson score interval
- Responsible Gambling Council (Canada)
- ConnexOntario (1-866-531-2600)
- National Council on Problem Gambling (1-800-GAMBLER)
Research and analysis, not betting advice. Past results do not guarantee future returns; bet only what you can afford to lose.