clock menu more-arrow no yes

Filed under:

Petr The Mraznificent

New, comments
James Guillory-USA TODAY Sports

He may be just 23-years old and have played in only 67 games, but Detroit Red Wings goaltender Petr Mrazek is quickly making a name for himself. After stealing the net for last season's playoff run, Mrazek started the season splitting time with Jimmy Howard. However, after a couple of poor performances by Howard, Mrazek seems to have officially taken a stranglehold on the goalie timeshare. Mrazek has started six of the last seven games while Howard has been pulled in each of his last two starts. Just how good has Petr Mrazek been and is his performance sustainable for years to come?

Explaining Goaltending Metrics

Using our conventional metrics, Petr Mrazek 2.20 goals against average ranks 12th and his .928 SV% ranks 7th. However, these two numbers, while familiar to almost all, don't really capture how well Mrazek has played. Goals against average isn't a great metric to use because it's so heavily reliant on a team's system. Chris Osgood has a career goals against average of 2.49 compared to 2.54 for Patrick Roy largely because Detroit's system was more sound defensively.

While save percentage is better than goals against average, it's still not great because it doesn't account for the quality of shot faced. If two goaltenders both faced 30 shots and saved 27, their save percentage will be 90%. However, if goaltender A faced 20 shots from the slot and goaltender B faced 5 shots from the slot, you could argue that the level of difficulty was higher for goaltender A and that his saves were "better". Thus, adjusted SV% was created to adjust for shot difficulty faced. Adjusted SV% can be found on War-On-Ice, but Greg Balloch at InGoal Magazine had a fantastic, easy to read breakdown of the metric. You'll most often find this statistic reported at 5v5 to account for the fact that goalies will see varying amount of time shorthanded, on the powerplay, and at 4v4.

Additionally, save percentage encompasses all types of shots, from center ice dump-ins, to one-timers in the slot. High Danger shots are the dark green shots in the image below.

Previous articles have shown that there is a correlation between goaltender pay grade and high-danger SV%. Additionally, high-danger SV% is subject to less variation when compared to low-danger and medium-danger shots on a year-by-year basis. Looking at the 60 goaltenders who have faced at least 1000 High-Danger shots (all strengths) since 2007, the average HDSV% is 82.3%. The average HDSV% of the top-10 goalies is 84.1% with a range of 83.62% to 86.4%. This information is also available on War-On-Ice.

In addition to adjusted SV%, Rob Vollman of ESPN has developed three additional stats to evaluate goaltenders. The first is referred to as Really Bad Start%. A Really Bad Start is defined as a start made by a goaltender where he had a save percentage <85%. On the flip side, Vollman also discusses Quality Starts and Quality Start%. A Quality Start is defined as a start where a goaltender posts a save percentage greater than the league-average SV% or a save percentage >88.5% in a game where the goalie sees 20 or fewer shots against. The third statistic Vollman developed is known as Goals Saved Above Average or GSAA. Here's an example that allows you to understand how this statistic works:

Braden Holtby's SV% is .933 SV% this season and he's faced 930 shots. League average SV% is .916 SV%. If you applied the league average SV% to the 930 shots Holtby's faced, you would expect the league average goaltender to give up 77.75 goals ((930-(930 x .916)). Holtby has only given up 62 goals. So he's saved 77.75-62 = 15.75 goals above average.

These three statistics are available on Hockey-Reference.

Goals Saved Above Average does have some limitations. First, it's a running statistic, meaning that goalies that face more shots and play more will have higher goals saved above average. Additionally, this metric does not account for shot quality. To combat this, Nick Mercadante has created adjusted Goals Saved Above Average/60 or adjGSAA/60 or "Mercad"/60. He has written a fantastic piece explaining this concept, but essentially, by making GSAA a rate statistic, it removes the advantage that goalies who played more had as well as adjusted for shot quality.

Mrazek's Amazing Season

Now that we have a decent background on some of the statistical tests available to us, let's talk about Mrazek this season. Here's a table of some of Mrazek's relevant statistics along with his NHL rank:

Statistic Value NHL Rank NHL Rank Qualification
Really Bad Starts% 0% 1st

≥20 Games Started

Quality Start% 69.6% 4th

≥20 Games Started

5v5 Mercad/60 +0.572 3rd

≥300 5v5 Shots Against

5v5 SV% 94.77% 3rd

≥300 5v5 Shots Against

5v5 Adj SV% 94.58% 3rd

≥300 5v5 Shots Against

5v5 HDSV% 88.20% 4th

≥100 5v5 High-Danger Shots Against

As you can see from the numbers, Mrazek has been simply fantastic this season. He's given the Wings a chance in every game as he's the only goaltender who has started more than 11 games without posting a really bad start. Not only that, but he's posted the 4th best quality start percentage. Basically, he's not had a really bad outing and has given the Wings a quality effort in just under 70% of his games.

From his Mercad/60 we see that Mrazek is saving more than 0.5 goals above average per 60 minutes. Basically every 120 minutes of 5v5 play (roughly 2.5 games) he's saving one more goal than the average netminder. In my discussions with Nick, he's mentioned that goaltenders that are able to sustain a +0.150/60 mark over the season are considered as having excellent seasons. I expect Mrazek to regress at some point, but right now we should appreciate the level he's playing at.

In terms of his 5v5 SV% and 5v5 adjusted SV%, he's in the top-3 in the NHL, behind only James Reimer and Connor Hellebuyck, two goaltenders who have both played fewer than 20 games. It becomes even more impressive when you recognize that Mrazek is facing 29.80 5v5 shots against/60, 7th most in the NHL (min 1000 5v5 mins). Last but not least, Mrazek's 5v5 high-danger SV% is elite, trailing only Reimer, Jonathan Quick, and Cory Schneider. Simply put, Mrazek has had an elite start to the season, perhaps one of the best in recent memory for a Red Wings netminder.

Mrazek's Career Performance And Future Projection

As you can see, Mrazek is absolutely on fire right now. However, we should temper our expectations. Mrazek's started just 23 games this season and the Wings still have half the season remaining. In the eight years we have of 5v5 SV% and 5v5 HDSV% data, the highest 5v5 adjusted SV% for a single season (min 2000 5v5 mins) is 94.99% (Tim Thomas - 2010-2011) and the highest 5v5 HDSV% (min 2000 5v5 mins) is 88.54% (Brian Elliott - 2011-2012). So will Mrazek finish with his current marks of 94.58% and 88.20%? Doubtful. However, he does have some encouraging career trends.

Since making his first NHL start on February 7th, 2013, Mrazek leads the NHL in 5v5 adj SV% at 93.35% (minimum 2000 minutes played). Looking at 5v5 Mercad/60 from 2012-present, we see that Mrazek has been exceptional in his small sample size.

While we should be impressed by Mrazek's current position, this particular image also does an excellent job of illustrating why we can't call Mrazek a shoo-in for "elite" status just yet. In this chart, Mrazek has faced 1,304 5v5 shots and has a 5v5 Mercad/60 of 0.3002, which is the best mark on the chart. However, you can see that he's faced 3030 shots fewer than Tuukka Rask who maintains a 5v5 Mercad/60 of 0.2741, the 3rd best mark. With all of that being said, we've got far too small of a sample size to confidently say that Mrazek will be a top-tier, elite netminder given how goalies like Martin Jones showed early promise but have faltered of late. The real test for Mrazek will be to see if he can sustain his elite level or at the very least avoid a steep drop off.

All the signs thus far are encouraging that Mrazek may be able to join the "elite" netminder club. The biggest knock against him was his inconsistency as he had a Really Bad Start% of 18.9% and a Quality Start% of 54.1% in his first 37 starts. However, over his last 23 starts, his Really Bad Start% is 0% and his Quality Start% is 69.6%. He's shown the ability to bounce back from tough goals against and poor efforts. He's kept the Wings in every single one of his starts this season. He's been a winner at every level he's played at. While I can't 100% tag him with the "elite" moniker just yet, all signs point to Detroit having found their franchise netminder.

Stats from War-On-Ice and Hockey-Reference