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: 2.80 Peer range: 1.68 – 4.20 xGF/60
MeanJason Robertson
Lower Higher (better)
Player 2.80 · 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
67.9 weighted minutes · 75% prior
On-ice defense · xGA/60
67.9 weighted minutes · 85% prior
Chance generation · ixG/60
67.9 weighted minutes · 75% prior
Shot quality · ixG/attempt
12.9 weighted attempts · 66% prior
Finishing · goals above expected/60
12.9 weighted attempts · 76% 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…
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.
Five-on-five · time together from shift charts (1,222:04 of his 5v5 ice time) · xG share together and apart
Teammate
Pos
TOI together
xG share together
GF–GA
CF%
Him without
Teammate without
Joe Pavelski
C
1,025:12
60.0%
36–31
55.5%
63.9%
44.4%
Roope Hintz
C
847:28
60.6%
27–28
55.4%
60.5%
47.7%
Miro Heiskanen
D
600:25
61.9%
13–12
58.8%
59.5%
52.8%
Ryan Suter
D
403:17
62.9%
11–7
58.2%
59.5%
50.5%
Colin Miller
D
380:34
57.5%
12–15
55.2%
62.2%
52.0%
Jani Hakanpää
D
312:59
59.1%
11–6
54.1%
61.2%
45.0%
Esa Lindell
D
305:37
58.1%
13–9
53.9%
61.5%
45.9%
Nils Lundkvist
D
283:43
60.5%
10–8
54.0%
60.6%
53.4%
Tyler Seguin
C
184:31
66.5%
6–5
62.6%
59.6%
50.9%
Wyatt Johnston
C
105:24
57.5%
2–0
51.9%
60.9%
52.1%
Joel Hanley
D
79:26
67.4%
3–5
52.8%
60.1%
49.8%
Jamie Benn
L
63:51
74.4%
0–2
75.6%
60.3%
50.7%
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%.
2022–23 playoffs
Official playoff totals · no peer ranks, since playoff samples are too small for a fair comparison
Games
19
18:46 per game
Goals
7
4 on the power play
Assists
11
official
Points
18
10 power-play points
Plus-minus
0
even strength and shorthanded
Shots
57
12.3% shooting
Playoff games
Official totals; expected-goal columns are five-on-five.
Elite90+Strong70–89Average30–69Weak10–29Poor<10Bar length is the peer percentile, so longer is always better · tick at the 50th
Player value: 2.83 Peer range: 0.58 – 3.13 / 60
MeanJason Robertson
Lower Higher (better)
Player 2.83 · mean 1.69. Higher raw values rank higher.
Recorded five-on-five goals plus assists per 60 minutes.
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
Metric2022–23ReferenceDiff
0.560.55+0.01
0.770.51+0.25
1.331.07+0.26
3.822.97+0.84
Reference: 2021–22 · 74 games.
Each game against his own season
Points in each game, coloured by how good the game wasTrailing 10 games, per gameSeason 1.33
