Signature stats
Draw Equity: what faceoffs are worth in goals
Faceoff percentage is the most quoted and least explained stat in hockey. A centre who wins 55% of his draws sounds good, but how many goals is that? Draw Equity answers in goals, using what winning a draw in each zone is actually worth.
Updated · Method version sequences v1
How it's calculated
Draw Equity = Σ over draws (win − 0.5) × V(zone)
- win
- 1 for a win, 0 for a loss
- V(zone)
- net xG in the 20 s after a win minus after a loss, from the league: about 0.019 for offensive- and defensive-zone draws, 0.007 in the neutral zone (5v5)
Scoring each draw against a coin flip means a win earns +0.5 × V and a loss costs −0.5 × V. A centre at exactly 50% has zero Draw Equity no matter how many draws he takes.
- V is the difference in net xG (for minus against, cut at a whistle) in the 20 s after the draw between won and lost draws, measured for the league.
- We recompute V every season. The published numbers come from 200 games of 2025-26.
- We report Draw Equity in goals and per 100 draws, at 5v5. Power-play draws are valued separately.
Net xG in the 20 s after a 5v5 faceoff win
Research sample: 200 games from 2025-26 (2025020500–699)
An offensive- or defensive-zone win is worth 0.019 net xG, a neutral-zone win 0.007.
Worked example
A 56% centre over a season
Includes example dataA centre takes 400 offensive- and defensive-zone draws at 5v5 and wins 224 (56%). He takes 150 neutral-zone draws and wins 83 (55%).
- End zones: (224 − 176) × 0.5 × 0.019 = 48 × 0.0095 = 0.456 goals.
- Neutral zone: (83 − 67) × 0.5 × 0.007 = 16 × 0.0035 = 0.056 goals.
- Total: 0.456 + 0.056 = 0.51 goals, or about 0.09 goals per 100 draws.
- For scale: in our 200-game sample, the leaders (Toews, Strome and Kopitar) were each around +0.36 to +0.38 xG.
About half a goal over a season: real, but small next to a centre's shooting or play-driving. The draw counts are illustrative; the zone values are the league's.
Reliability
Winning draws is a skill. What happens after the draw is not, so all the value in Draw Equity comes from the win rate.
| Measure | Split-half r | Sample | Verdict |
|---|---|---|---|
| FO%About 0.8 over a full season. | 0.31 | 67 draws per half | Moderate |
| Post-draw execution (xG after wins vs league)Not published. | 0.00 | Split-half, 200 games | Noise |
| Zone value V | — | League, recomputed each season | Stable |
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
- V uses a 20 s window. Some of a win's value (a clean zone exit, a line change) lasts longer and is not counted.
- Draw Equity credits the centre for the whole win. Wingers and defencemen help on scrums, which the feed does not record.
- We have not yet tested faceoff records for arena scorer bias. Other scorer-recorded events (hits, giveaways) showed strong bias before 2024-25.
FAQ
How much is a faceoff win worth?
At 5v5, an offensive- or defensive-zone win is worth about +0.019 net expected goals over the next 20 seconds compared with a loss, and a neutral-zone win about +0.007. So 100 extra end-zone wins are worth roughly 1.9 goals.
Is faceoff percentage a skill?
Yes. In our sample FO% had a split-half correlation of 0.31 after just 67 draws per half, rising to about 0.8 over a full season. It is one of the most repeatable individual stats.
Is winning the draw and then scoring a skill?
No. Whether a centre's team generates more than the league average after his wins had a split-half correlation of 0.00, so we treat it as noise and don't publish it.
Why score draws against 50%?
Because a coin-flip centre adds nothing over replacement. Scoring each draw as (win − 0.5) makes Draw Equity zero at 50% and symmetric for winners and losers.