Bands show position percentiles; axis labels show current rolling values.
Dashed outline: average forward in the fixed reference. Outside = above average; inside = below average. All axes point outward for stronger results.
Player value: 3.09 Peer range: 1.68 – 4.20 xGF/60
MeanSeth Jarvis
Lower Higher (better)
Player 3.09 · 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
1377.6 weighted minutes · 13% prior
On-ice defense · xGA/60
1377.6 weighted minutes · 23% prior
Chance generation · ixG/60
1377.6 weighted minutes · 13% prior
Shot quality · ixG/attempt
290.5 weighted attempts · 8% prior
Finishing · goals above expected/60
290.5 weighted attempts · 12% prior
History through Apr 11, 2026. 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.
Seth Jarvis
333 attempts37.46 xG34 G
Goal34Shot on goal196Missed shot103Blocked shot0Larger dot = higher xG
Shooting by zone
Zone
Shots
xG
Goals
S%
Inner slot
117
17.7
13
11.1%
High slot
110
12.5
13
11.8%
Point
30
1.6
1
3.3%
Right circle
29
0.8
1
3.4%
Outer slot
23
2.6
5
21.7%
Left circle
21
2.0
1
4.8%
Below goal line
3
0.1
0
0.0%
Hover or focus a row to light up those attempts on the rink.
Finishing evidence. Scoring 10.2% on chances vs 11.2% expected (−1.0 points), with the largest shortfall in the inner slot.
Shot types333 attempts
Snap37%
Wrist29%
Backhand11%
Tip-in10%
Slap9%
Deflected2%
Other2%
High-danger chances
Expected
20.5
Goals
18
Conversion
12.6%
Model 14.3%
143 high-danger attempts · Model is the expected conversion on those attempts.
Recent form
Loading recent games…
Season vs reference
Loading season comparison…
Loading playoff games…
Loading game log…
Loading career ledger…
1 award records
Stanley Cup · 2025–26
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.
Ice time, share of team ice time, scoring and scoring rate, with where shifts start.
Scope
TOI/GP
Team share
G
A
P
P/60
Five-on-five
12:56
26.6%
17
12
29
1.89
Power play
3:07
63.6%
5
16
21
5.68
Penalty kill
1:36
34.4%
4
2
6
3.17
Other
1:13
—
6
4
10
6.91
All situations
18:53
31.0%
32
34
66
2.95
Where the 66 points come from · All situations
Goals32 · 1.43/60
Primary assists19 · 0.85/60
Secondary assists15 · 0.67/60
22.7% of points are secondary assists. Peer average 22.7%. Primary points per 60 (2.28).
Zone starts 68.8% offensive of OZ + DZ · 90th percentile
Offensive 41.7%Neutral 39.3%Defensive 19.0%
Shifts
21.7 a game · 0:52 average
Zone starts: five-on-five · Shifts: all situations
Season metrics
2025–26 · five-on-five · against qualified peers
Production
Player
Percentile
Rank
1.89
39 / 109
1.11
10 / 109
0.33
90 / 109
Shooting & finishing
Player
Percentile
Rank
1.15
1 / 109
9.34
4 / 109
16.14
20 / 109
0.082
16 / 91
-0.04
53 / 91
Above averageNear averageBelow averageBars and shaded cells show peer percentile, so more is always better · tick at the 50th
Player value: 1.89 Peer range: 0.30 – 3.71 / 60
MeanSeth Jarvis
Lower Higher (better)
Player 1.89 · 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. Green/red differences show favorable/unfavorable values versus average, not isolated player impact. Brass marks top-ten ranks.
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.
Recent form
Trailing ten appearances. Shaded games have fewer than ten in the window.
Trailing 10 games · Points / gamePoints in each gameSeason points / game
1.00 P / GP
Apr 11, 2026@ UTA
Trailing 10 · Points / game
1.00
Points in this game
0
Season points / game
0.93
10 appearances in window10 with five-on-five model coverage
Elite two-way play and chance creation with excellent playmaking — he just isn't converting his looks, sitting at 0.04 goals below expected per 60 (53rd of 91).
5 NHL seasons: 375 games, 279 points
5 NHL seasons: 375 games, 279 points
Totals cover all situations; rates are five-on-five.
Expected goals against while the player is on the ice (fewer is better)
Chance creation
Expected goals from the player’s own shots
Finishing
Goals minus expected goals on the player’s own shots
Assisting
Primary assists plus half of secondary assists
Each is measured against expectation, per 60 rated minutes: plus is above, minus below.
How this is calculated
Score breakdown
Component
Weight
Scaled
Contribution
Offense
25%
0.424
+0.106
Defense
25%
0.242
+0.061
Chance creation
20%
0.704
+0.141
Finishing
10%
−0.092
−0.009
Assisting
20%
0.143
+0.029
Performance rate
0.327
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: 29 covered appearances, 492.0 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: 47.7 min excluded; official TOI 1340.5 min.
Coverage: complete. 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: 89.5 / 100. The fixed exposure baseline is 50; this is separate from the season peer grade.
