Bands show position percentiles; axis labels show current rolling values.
Dashed outline: peer average in the fixed reference. Outside = above average; inside = below average. All axes point outward for stronger results.
Player value: 2.37 Peer range: 1.68 – 4.20 xGF/60
MeanTom Wilson
Lower Higher (better)
Player 2.37 · mean 2.54. Higher raw values rank higher.
Rolling profile against a fixed reference group
Fixed rolling-profile reference · 430 players. Values use the same history and smoothing as the radar.
History, sample sizes & calculation
Higher radar percentiles mean stronger recorded results against the fixed reference, including older seasons. These are separate from this season’s league percentile. Each player has a different amount of covered history. Weighted minutes and attempts give recent games more influence. “Prior” is the share contributed by a frozen league baseline: it is higher when a player has less evidence, and differs by metric because each uses its own smoothing strength. It is not a confidence score.
On-ice offense · xGF/60
61.4 weighted minutes · 77% prior
On-ice defense · xGA/60
61.4 weighted minutes · 87% prior
Chance generation · ixG/60
61.4 weighted minutes · 77% prior
Shot quality · ixG/attempt
7.9 weighted attempts · 76% prior
Finishing · goals above expected/60
7.9 weighted attempts · 83% prior
History through Oct 24, 2023. Recent games count more than older games. History carries across seasons and stays unchanged while inactive. On-ice results are shared; finishing describes results rather than established talent.
Recent scoring against his own season
Recent scoring against his own season
Loading recent games…
Loading splits…
Loading linemates…
Loading playoff games…
Loading game log…
Loading career ledger…
1 award records
Stanley Cup · 2017–18
Every metric name opens its definition; the full list lives in Metrics.
Traditional statistics cover all situations. Supporting five-on-five analyses use completed games, raw adjustment and exposure from xG model v4. The two samples can differ.
An em dash means unavailable, not zero. A partial season remains partial. Where a panel reports evidence coverage, its rates use only the covered games. History never substitutes a different model, and a missing appearance interrupts a rolling model window.
Shot maps use the same shot model as Game Center and show which model version scored recent shots beside the rink. Older seasons can use earlier model versions. High danger means a model scoring probability of at least 0.20 on an unblocked attempt; for goalies, medium danger is 0.08 to 0.20 and low is below 0.08.
Special-team buckets include the wider recorded man-advantage and man-down states, read from the player’s own team; for a goalie, penalty kill means facing a power play. Their ranks need 100 recorded minutes and at least 30 qualified peers; below 100 minutes, penalty-kill percentiles are provisional. Playoff values have no ranks or percentiles. Team share is player ice time divided by team ice time in that situation, over the games the player dressed. Zone starts describe faceoff-started five-on-five shifts and exclude on-the-fly starts.
Penalties count events, including one for a double minor; bench and served-by roles are excluded, and victimless infractions have no drawn role. EPV added sums changes in the probability that the possessing team scores next; it is not standings points or an isolated player grade. Rolling windows never cross seasons or season types, and a missed game is not a zero-performance appearance. Career rates use summed counts and exposure, never an average of season rates; a missing contributing season withholds its career metric, and Performance begins in 2023–24.
Playing-time minimum met. Rank withheld because the app’s scoring data are incomplete. The app has checked scoring results for 39 of 46 appearances (84.8%); season standings require at least 90%. Scoring data cover less than 90% of recorded workload or could not be verified.
vs 113 qualified forwards
Percentiles appear once this player qualifies for season standings. The score breakdown is under How this is calculated.
How this is calculated
Score breakdown
Component
Weight
Scaled
Contribution
Offense
25%
0.420
+0.105
Defense
25%
0.126
+0.032
Chance creation
20%
0.206
+0.041
Finishing
10%
0.502
+0.050
Assisting
20%
0.151
+0.030
Performance rate
0.258
Per 60 rated minutes, over the scored appearances. Each component is scaled separately, then weighted; the contributions add to the performance rate at full precision, so rounded figures may differ.
Scoring
Season grade is the qualified peer percentile on the same letter bands as Game Grade — A+ ≥88, A ≥78, A− ≥68, B+ ≥58, B ≥50, B− ≥42, C+ ≥35, C ≥28, C− ≥22, D ≥15, F below 15 — and needs qualification plus at least 30 eligible peers. The letter describes this season’s performance, not permanent ability.
Comparisons
Rank and component percentiles compare this season’s qualified forwards on unrounded rates over every checked appearance; Game Grade compares one appearance with frozen historical games.
Component percentiles are never averaged into the score. Recent history is optional, from up to three completed prior seasons; small samples are provisional.
Coverage
Qualification: 16 covered appearances, 268.1 rated minutes, and 90% appearance and workload coverage. Minimums advance with league progress; valid low-sample results stay visible.
Rated minutes exclude play with a goalie pulled: 12.5 min excluded; official TOI 712.2 min.
Coverage: partial. Updated .
Method
Five weighted components add to the performance rate. Five-on-five offense and defense account for recorded teammates, opponents and game state, so shared on-ice results are not isolated individual ability. Finishing is scored against the shot model’s historical-shooter expectation and assisting counts primary assists plus half of secondary assists; both use the strength baseline, chance creation its positional baseline.
Legacy rating
Legacy season score: 87.5 / 100. The fixed exposure baseline is 50; this is separate from the season peer grade.
Most often with Pierre-Luc Dubois: 631 five-on-five minutes, 55.2% xG share together
Most often with Pierre-Luc Dubois: 631 five-on-five minutes, 55.2% xG share together
Five-on-five · time together from shift charts (1,093:36 of his 5v5 ice time) · xG share together and apart
Teammate
Pos
TOI together
xG share together
GF–GA
CF%
Him without
Teammate without
Pierre-Luc Dubois
C
631:28
55.2%
31–24
51.4%
50.3%
54.4%
Connor McMichael
L
513:08
54.2%
23–20
49.8%
52.5%
49.9%
John Carlson
D
437:05
57.1%
19–12
53.3%
50.7%
53.3%
Dylan Strome
C
385:45
52.6%
16–11
50.2%
53.7%
46.5%
Jakob Chychrun
D
383:00
51.0%
13–15
51.4%
54.6%
49.5%
Rasmus Sandin
D
368:07
52.3%
13–12
48.7%
53.9%
52.1%
Trevor van Riemsdyk
D
321:20
45.6%
12–17
46.1%
56.7%
52.4%
Martin Fehérváry
D
310:09
58.8%
16–7
52.9%
50.9%
52.3%
Matt Roy
D
292:38
52.2%
12–6
49.7%
53.8%
49.8%
Aliaksei Protas
L
274:14
51.5%
9–8
54.2%
53.9%
51.6%
Alex Ovechkin
L
195:06
54.9%
11–7
47.8%
53.1%
46.5%
Andrew Mangiapane
L
55:26
57.4%
1–1
48.8%
53.2%
53.5%
Without columns are five-on-five xG share apart: him on the ice without that teammate, and the teammate without him (with this club). Bars are centred on an even 50%.
2024–25 playoffs
Official playoff totals · no peer ranks, since playoff samples are too small for a fair comparison
Games
10
20:01 per game
Goals
3
1 on the power play
Assists
4
official
Points
7
3 power-play points
Plus-minus
+3
even strength and shorthanded
Shots
19
15.8% shooting
Playoff games
Official totals; expected-goal columns are five-on-five.
Elite90+Strong70–89Average30–69Weak10–29Poor<10No rank yetBar length is the peer percentile, so longer is always better · tick at the 50th
Player value: — Peer range: 0.76 – 1.99 / 60
Mean
Lower Higher (better)
Player — · mean 1.28. Higher raw values rank higher. Complete recorded inputs are unavailable.
Recorded five-on-five goals plus assists per 60 minutes.
Complete recorded inputs are unavailable.
How to read these comparisons
Player values are recorded season results. Peer averages are arithmetic means of qualified players in the selected group; each row uses the same peers for its average and rank. Ranks use unrounded values, with tied players sharing a rank and the next rank skipped. Lower xG against ranks better. Bar colours are the five performance tiers; context rows are blue and not graded. Differences from the average describe the season, not isolated player impact.
Goals above expected is the raw goals-minus-xG rate, without the shrinkage used in the older finishing comparison. On-ice measures include teammates, opponents and deployment. These recorded metrics describe different quantities from the adjusted Season score components beside them.
Skaters need 20 appearances and 300 five-on-five minutes; quality and finishing also need 75 unblocked attempts. Early in a season these minimums scale down with games played so far. At least 30 peers must qualify. Playoff ranks are withheld. A reviewed axis whose own evidence fails is withheld and shown with its reason rather than ranked. These workload rules do not express statistical confidence.
Position uses the NHL listed primary position available now; it does not establish historical position or actual deployment. Broad groups are forwards, defence and goalies. This version uses the current model’s covered games only.
Season vs reference
Per game · all situations
Metric2024–25ReferenceDiff
0.410.24+0.16
0.400.23+0.17
0.800.47+0.33
2.092.28-0.20
Reference: 2023–24 · 74 games.
Each game against his own season
Points in each game, coloured by how good the game wasTrailing 10 games, per gameSeason 0.80
