Petr Mrazek — 2017–18 stats and season rank | IceSheet
Petr Mrazek
Anaheim Ducks·#34·Goalie · Catches left
Age 34 · 6′2″ · 192 lb · Ostrava, CZE
No checked season results are available for this selection.
Games3933rd in NHL
Starts33T-34th in NHL
Save %.90256th in NHL
GAA3.0351st in NHL
Shutouts4T-9th in NHL
Shots faced108035th in NHL
/
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.
Goals saved against expectation, and where the shots came from
Goals saved against expectation, and where the shots came from
Petr Mrazek faced
1530 attempts106.40 xG112 G
Goal112Shot on goal982Missed shot436Blocked shot0Larger dot = higher xG
Shooting by zone
Zone
Shots
xG
GA
SV%
High slot
465
43.9
35
92.5%
Point
461
9.8
11
97.6%
Inner slot
297
39.5
50
83.2%
Right circle
104
2.6
1
99.0%
Outer slot
102
7.4
7
93.1%
Left circle
94
2.9
8
91.5%
Below goal line
7
0.2
0
100.0%
Hover or focus a row to light up those attempts on the rink.
Stopping evidence. Allowed a goal on 7.3% of attempts faced against 7.0% expected, tracking the model across 7 zones.
Shot types1530 attempts
Wrist46%
Slap18%
Snap18%
Backhand8%
Tip-in6%
Deflected3%
Other1%
High-danger chances
Expected against
47.1
Goals against
57
Save %
85.9%
Model 88.3%
403 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.
This season’s goalie profile has not been published yet. It appears once the season’s numbers have been reviewed.
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.
Goalies need 15 covered appearances, 500 verified five-on-five shots on goal and 90% shot and appearance coverage; high danger also needs 50 high-danger shots, mid and low danger need 100 tier shots, and rebound suppression needs 600 eligible saves across 15 appearances with eligible saves. 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
Metric2017–18ReferenceDiff
.902.901+.001
per 60 minutes3.033.04-0.01
per game27.6929.24-1.55
Reference: 2016–17 · 50 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 Jan 5, 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: -0.25 Peer range: -0.60 – 1.49 residual per 100
MeanPetr Mrazek
Lower Higher (better)
Player -0.25 · 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
586.1 weighted verified SOG · 63% prior
Mid-danger stopping · residual per 100
202.8 weighted verified SOG · 55% prior
Low-danger stopping · residual per 100
338.3 weighted verified SOG · 86% prior
Rebound suppression · residual per 100
517.7 weighted eligible saves · 89% prior
Verified 5v5 stopping and rebound profile. High-danger stopping is withheld; these overlapping axes are not added together.
History through Jan 5, 2026. Recent games count more than older games. History carries across seasons and stays unchanged while inactive.
