Shot models
Scoring chances and high-danger chances: the zone method
Scoring chances are the transparent cousin of expected goals. There is no model: a fixed set of zones and two adjustments decide whether a shot was dangerous. That makes them easy to check by hand and comparable with other sites that use the same definition.
Updated
How it's calculated
value = zone points + (1 if rush or rebound) − (1 if blocked)
- zone points
- 3 in the net-front slot, 2 in the wider slot, 1 anywhere else in the offensive zone
- SC
- value ≥ 2
- HD
- value ≥ 3 (MD is exactly 2; LD is 1 or less inside the offensive zone)
| Zone | Points | Area |
|---|---|---|
| Net-front slot | 3 | 69 ≤ x ≤ 89 and |y| ≤ 9 (goal line out to the faceoff dots) |
| High slot | 2 | 44 ≤ x < 69 and |y| ≤ 9 |
| Between the dots | 2 | 55 ≤ x < 69 and 9 < |y| ≤ 22 |
| Low triangles | 2 | 69 ≤ x ≤ 89 and 9 < |y| ≤ 9 + 13 × (89 − x) / 20 |
| Rest of the offensive zone | 1 | x > 25, including behind the net |
| Neutral and defensive zone | — | x ≤ 25: excluded |
- Rebound: an attempt within 3 s of an earlier attempt by the same team, with no stoppage in between.
- Rush: NST's rule, an attempt within 4 s of any event in the shooting team's neutral or defensive zone. See the transition page for why we also publish a broader transition definition.
- Zones are computed from our own normalized coordinates, not the feed's zone code, which is recorded from the blocker's point of view on blocked shots.
Danger zones and the six example attempts
Example data: invented for illustration
Offensive half rink with the high-danger net-front slot, the medium-danger wider slot and the low-danger remainder, plus three example shot locations.
Worked example
Six attempts, scored by hand
Includes example dataThree shot locations, each taken twice: once plainly, once with a bonus or penalty. The rink above shows the zones.
- Net-front shot at (80, 3): slot, 3 points. High danger.
- Net-front rebound at the same spot: 3 + 1 = 4 points. Still high danger; the bonus doesn't change the band here.
- Shot between the dots at (60, 15): 2 points. A scoring chance, medium danger.
- The same shot, blocked: 2 − 1 = 1 point. Not a scoring chance.
- Point shot at (38, −28): 1 point. Low danger.
- The same point shot off the rush: 1 + 1 = 2 points. Now a scoring chance.
Four of the six attempts are scoring chances and two are high danger. The locations are made up; the rules are the real ones.
Reliability
Scoring-chance counts behave like xG counts with a coarser scale. We have not published a split-half r for them yet; our reference for how quickly on-ice shot quality stabilises is xGF% (r 0.29 at about 7 xG per half).
| Measure | Split-half r | Sample | Verdict |
|---|---|---|---|
| Team SCF / HDCF per gameA match test, not a reliability test. | — | Target: ≥ 95% exact match with NST on 20 private team-games | Not measured yet |
| On-ice 5v5 xGF% (reference) | 0.29 | About 7 xG per half-sample | Moderate |
Split-half r: we split each sample in two (for example odd and even games) and correlate the two halves across players or teams. Near 0 means the stat is mostly noise; near 1 means it is a stable trait.
Limitations
- The zones were measured from an image, so an attempt within about a foot of a boundary may be classed differently here than on NST.
- A zone method treats every slot shot alike. A tip and a clean look from the same spot score the same 3 points; xG separates them.
- NST's rush rule is almost empty on modern data (1.7% of 5v5 attempts in our 2025-26 sample), so the rush bonus rarely applies. We kept it for comparability.
- Blocked attempts are located where the block happened, which can move them into or out of a zone.
FAQ
What counts as a scoring chance in the NHL?
Under the zone method we use, an offensive-zone attempt is a scoring chance if its location points (3 in the net-front slot, 2 in the wider slot, 1 elsewhere), plus 1 for a rush or rebound, minus 1 if blocked, add up to 2 or more.
What is a high-danger chance?
An attempt worth 3 or more points. Almost every unblocked attempt from the net-front slot qualifies, as does a rush or rebound attempt from the wider slot.
Why do your scoring chances differ slightly from Natural Stat Trick?
We implement the same War-on-Ice definition, but the zone boundaries were measured from NST's published image and may differ by about a foot. Shift-boundary timing can also move an attempt between players by one event.
Should I use scoring chances or xG?
xG is more precise because it uses shot type and the previous event, not just location. Scoring chances are simpler to check by hand. When they disagree over a small sample, trust neither too much.