Arturs Silovs — 2025–26 stats and season rank | IceSheet
Arturs Silovs
Pittsburgh Penguins·#37·Goalie · Catches left
Age 25 · 6′4″ · 219 lb · Riga, LVA
Loading season standing…
Games39T-30th in NHL
Starts38T-30th in NHL
Save %.88849th in NHL
GAA3.0749th in NHL
Shutouts2T-16th in NHL
Shots faced1020…
/
Season comparisons
Loading peer comparisons and model evidence…
Loading model evidence…
Shot locations
Loading canonical shot events…
Usage
Loading usage…
Recent form
Loading recent games…
Loading playoff games…
Loading game evidence…
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.
T-31st in NHL
Arturs Silovs faced
1594 attempts118.98 xG128 G
Goal128Shot on goal914Missed shot552Blocked shot0Larger dot = higher xG
Shooting by zone
Zone
Shots
xG
GA
SV%
High slot
453
44.9
34
92.5%
Point
443
10.1
12
97.3%
Inner slot
361
48.2
65
82.0%
Left circle
119
3.6
2
98.3%
Outer slot
115
8.6
9
92.2%
Right circle
86
2.7
5
94.2%
Below goal line
17
0.9
1
94.1%
Hover or focus a row to light up those attempts on the rink.
Stopping evidence. Allowed a goal on 8.0% of attempts faced vs 7.5% expected (0.6 points more), with the most goals above expectation from the inner slot.
Shot types1594 attempts
Wrist58%
Snap11%
Slap11%
Tip-in11%
Backhand7%
Deflected1%
Other1%
High-danger chances
Expected against
57.6
Goals against
75
Save %
84.8%
Model 88.3%
492 high-danger attempts · Model is the expected conversion on those attempts.
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.
Stopping and rebounds measure results. Stoppages describe style; shot difficulty and workload describe what the goalie faced. A larger shape is not a better season.
Among G peers · 57 qualified players · Descriptive profile · performance, style and context
Player value: -0.32 Peer range: -0.89 – 1.32 goals saved per 100 SOG
MeanArturs Silovs
Lower Higher (better)
Player -0.32 · mean 0.20. Higher raw values rank higher.
Adjusted goals saved above expected per 100 verified five-on-five shots on goal. Higher is better.
How to read these comparisons
Each axis ranks the player against qualified peers. The dashed reference is the percentile of their arithmetic mean, which need not be the 50th percentile. Ties share a midrank.
This shape combines performance, playing style and the difficulty and volume of shots faced. Farther from the center means more of that named quantity. A larger shape is not an overall grade. Post-save stoppages are a recorded event proxy, not goalie-attributed freezes.
Goalies need 15 covered appearances, 500 verified five-on-five shots on goal and 90% shot and appearance coverage. Rebound suppression needs 600 eligible saves across 15 appearances. Post-save stoppages need 600 observable saves across 15 appearances, at least 90% horizon coverage and 98% reason coverage. At least 30 goalies must qualify per axis. 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
All situations
Metric2025–26ReferenceDiff
.888.861+.027
per 60 minutes3.073.65-0.58
per game26.1523.80+2.35
Reference: 2024–25 · 10 games.
Recent-history profile
Loading recent-history profile…
Goalie history radar
Stopping and rebound results across recent NHL seasons, compared with a fixed goalie reference. Includes previous seasons; separate from the selected-season standing above.
Through Oct 16, 2026. Radar and percentiles use the same checked history; season ranking qualification does not hide this profile.
Bands show position percentiles; axis labels show current rolling values.
Dashed outline: average goalie in the fixed reference. Outside = above average; inside = below average. All axes point outward for stronger results.
Player value: -0.40 Peer range: -0.60 – 1.49 residual per 100
MeanArturs Silovs
Lower Higher (better)
Player -0.40 · mean 0.49. Higher raw values rank higher.
Rolling profile against a fixed reference group
Fixed rolling-profile reference · 52 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.
Overall stopping · residual per 100
615.8 weighted verified SOG · 62% prior
Mid-danger stopping · residual per 100
221.2 weighted verified SOG · 53% prior
Low-danger stopping · residual per 100
347.3 weighted verified SOG · 85% prior
Rebound suppression · residual per 100
539.2 weighted eligible saves · 88% prior
Verified 5v5 stopping and rebound profile. High-danger stopping is withheld; these overlapping axes are not added together.
History through Oct 16, 2026. Recent games count more than older games. History carries across seasons and stays unchanged while inactive.
Poor: -1.29 goals saved below expected per 100 five-on-five shots (36th of 41).
Season grade
F
#36 of 41 qualified G13th percentile
Goals saved / 100-1.29039/39 GP
Poor: -1.29 goals saved below expected per 100 five-on-five shots (36th of 41)
Poor: -1.29 goals saved below expected per 100 five-on-five shots (36th of 41)
Qualified · 2025–26 goalies
Goals saved above expected
-7.82
Five-on-five
Expected goals against
79.18
87 allowed at five-on-five
Save percentage
.893
812 five-on-five shots
Goals saved above expected
-3.34
All situations
Saves by shot danger · five-on-five
Danger
Shots
Share
Goals against
Save %
High danger
207
25%
41
.802
Medium danger
202
25%
27
.866
Low danger
403
50%
19
.953
Against qualified goalies · five-on-five
Shot stoppingAdjusted goals saved above expected per 100 verified five-on-five shots on goal. Higher is better.-0.32 goals saved per 100 SOG
13th percentile among 57 qualified goalies
Rebound opportunities preventedAdjusted quick follow-up attempts prevented versus expected per 100 eligible saves. Team defense also affects this result.0.81 follow-ups prevented per 100 saves
59th percentile among 40 qualified goalies
How this is calculated
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 goalies on unrounded rates over every checked appearance; Game Grade compares one appearance with frozen historical games.
Coverage
Qualification: 13 covered appearances, 656.0 verified 5v5 shots, and 90% appearance and workload coverage. Minimums advance with league progress; valid low-sample results stay visible.
Zero-shot relief appearances do not count toward qualification until eligible five-on-five ice time is verified.
Coverage: complete. Updated .
Method
Stopping uses the independent verified five-on-five shot-on-goal model. Danger and rebound measures have separate gates and are not added to this score.
-10.46 total goals saved above expected.
Legacy rating
5v5 legacy season score: 18.7 / 100. The fixed exposure baseline is 50; this is separate from the season peer grade.
