Playmaking

Step 01 · Shot quality

What his linemates’ shots are worth with him out there

The on-ice components pay all five skaters the same for every chance. Playmaking asks who made the chances better. One regression explains the expected-goal value of every unblocked attempt since 2010-11 by the shooter (his career and season level), every teammate and every opponent on the ice, and the game state. A teammate’s coefficient is how much more dangerous the shooters’ attempts are with him on the ice than their own records predict. It is shrunk hard (a ridge penalty of 1,000 attempts), so only a long, consistent pattern survives.

McDavid’s even-strength coefficient is +0.00623 xG per attempt, rank 1 of 2811; on the power play +0.00955. The even-strength leaders:

1. Connor McDavid +0.0062
2. Leon Draisaitl +0.0055
3. Artemi Panarin +0.0055
4. Nick Schmaltz +0.0052
5. Nathan MacKinnon +0.0047

Step 02 · Price it

Extra xG into net goals

On every attempt a teammate took while he was on the ice, he is credited his coefficient times the number of skaters who shared the on-ice credit for it (5 at 5-on-5, fewer at 4-on-4 and 3-on-3), times the price of an on-ice expected goal. The price is measured from the on-ice components themselves: one more expected goal for while a skater is on the ice is worth 0.2158 net goals to him at even strength and 0.1695 on the power play.

Step 03 · Assists

His share of his linemates’ finishing

The second half of Playmaking is the part of a pass the shot location cannot see: the goalie pulled across, the shot released before the defence set. Half of every even-strength and power-play shot’s goals above expected goes to the shooter’s teammates on the ice, split 4 : 2 : 1 between the primary assist, the secondary and everyone else; half of every shot’s expected-goal cost is split equally, so a passer whose linemates keep getting stopped pays too.

Take one goal: Andrew Mangiapane scored at P1 16:11 of game 2025020006 on a 0.100-xG shot, with Mattias Ekholm (on ice), Connor McDavid (primary assist), Trent Frederic (on ice), Evan Bouchard (on ice). Half of its 0.900 goals above expected goes to the teammates in proportion to their weights (total 7):

Over the season he was on the ice for 805 teammate shots on goal and 108 goals (48 primary assists, 32 secondary, 28 without a point):

Step 04 · Zero-sum

Where the credit comes from

None of this is new value. The shot-quality credit is taken back from the shared on-ice columns: on every attempt each skater on the ice (shooter included) pays the price times the sum of his teammates’ coefficients, out of EV offense or the power play. The assist credit is taken from the shooter’s Finishing. Every team’s total is exactly what it was. For McDavid:

Playmaking and each debit are then re-centred on the league per game played, so the average player is zero in each column (and the four constants cancel):

Step 05 · Position

Measured against the average forward

After every other step, each component is re-centred once more, separately for forwards and for defencemen, so the average forward in 2025-26 is exactly zero in every column. The games-weighted mean playmaking of all forwards that season was +0.00691 net goals per game, so 82 games of it come off:

Step 06 · Result

Per 84 games

Ratings are quoted per 84 games, one full schedule, so a part season and a full one sit on the same scale.

✓ Recomputed +13.72; his 2025-26 row on the player card and the Players table shows +13.7. They match.

Step 07 · Context

How unusual is this?

Skaters with 40 or more games in 2025-26 (612 players). Connor McDavid ranks 1 of 612.

How much of a season like this is skill? The model estimates, from how much players’ own rates bounce between seasons versus how much players differ, that a career rate of playmaking is half signal and half noise after 28 games. Season rows like this one are shown exactly as they happened; careers and the card’s Regressed view are pulled toward average by that amount.