A deeper look into S94 goaltending part 2 - let's talk about the consistency

[VattghernCZ 983]

Part 1: https://vhlforum.com/topic/150303-a-deeper-look-into-s94-goaltending-part-1-goals-saved-above-average/#comment-1034715

 

In last week’s media spot, I ranked all of our VHL goaltenders by Goals Saved Above Average, a piece of statistics that describes how many goals the goaltender saved or allowed compared to a goaltender with league average save percentage. The runaway winner was Lachlan Summers, preventing about 42 goals compared to the amount an “average” goaltender would have allowed, with about 13 goals saved more than the second-placed Fuukka Rask (who played significantly fewer games).This week, I'm going to take a look into two pieces of statistics that describe goaltenders’ consistency, and I'm going to add a not hockey-related one.

 

Note: in this article, I'm only going to focus on starting goaltenders, including Georgiy Costanzov, who had been the Bears’ starter in the first half of the season, before Fuukka Rask took over the reigns and finished the season with more games played than Costanzov. The backup goaltenders’ sample size is too small to be significant, in my opinion.

 

Quality Starts (QS)

 

Quality Starts is a piece of statistics developed by Rob Vollman in the Hockey Abstract. According to him, a quality start means a game goaltender starts* and finishes with a save percentage higher than the league average (which is .917 for S94, as calculated in Part 1), or with a save percentage of at least .885 in games with fewer than 20 shots against. It's often used as Quality Starts Percentage, which is the percentage of Quality Starts among all Games Started (GS), describing how often a goaltender has strong performances. And yes, this required me to go through game logs to figure out the amount of games started and the goaltenders' save percentage game-to-game. Although to be brutally honest, I "only" went through 439 of the total 576.

 

For reference, in the NHL, QS% > 60 is considered good, QS% < 50 is considered bad, and QS% 53 is about league average. I'm curious to see what the numbers will be in the VHL.

 

*If a goaltender plays the whole game, it counts as a GS. If a goaltender starts the game and is pulled, it counts as a GS. If a goaltender doesn't start a game, but enters the game later, it doesn't count as a GS, even though it counts as a game played.

 

1.  Fuukka Rask @Jubis: 72.4 QS%

2.  Lachlan Summers @kirbithan: 72.0 QS%

3.  Ash Sparks @DarkSpyro: 63.8 QS%

4.  Red Panda @Lemorse7: 59.3 QS%

5.  Joel Castle @animal74: 56.5 QS%

6.  David Slezák @ShawnGlade: 56.0 QS%

7.  Merome Dilson @MexicanCow123: 53.1 QS%

8.  Dalkr Vidarsson @KaleebtheMighty: 51.0 QS%

9.  Amir Redzic @Slav_Sloth: 50.0 QS%

10.  Herald Benson @Benson: 49.0 QS%

11.  Jonny Elgar @Jonny: 48.9 QS%

12.  Olober Syko @Spartan : 47.9 QS%

13.  Matthew McCagg @Jack kidd: 47.8 QS%

14.  Georgiy Costanzov @Mongoose87: 46.2 QS%

15.  WWWWWWWWWWWW WWWWWWWWWWWW @rory: 45.0 QS%

16.  Justin Lion @Emperor_Fun: 44.7 QS%

17.  Ryan Artyomov @Enorama: 43.5 QS%

 

The same two names who stood out in Goals Saved Above Average, Lachlan Summers and Fuukka Rask, are running away with QS% as well, giving their teams a chance to win significantly more often than not.

 

Really Bad Starts (RBS)

 

Another piece of statistics developed by the aforementioned Rob Vollman. It stands on the opposite side, as it describes how often does a goaltender have a bad game. These often result in the goaltender being pulled. As a really bad start, Vollman describes a game started with a save percentage below .850. It looks too arbitrary to me; I don't personally like it, but it's an “official” stat listed on Hockey Reference, so I'm going to include it here.

 

1.  Dalkr Vidarsson; 2.0 RBS%

2.  Fuukka Rask: 3.4 RBS%

3.  Lachlan Summers: 4.0 RBS%

4.  Merome Dilson: 4.1 RBS%

5.  Justin Lion: 4.3 RBS%

6.  Red Panda: 5.6 RBS%

7.  David Slezák: 6.0 RBS%

8.  Herald Benson: 6.1 RBS%

9.  Ash Sparks: 6.4 RBS%

10.  Amir Redzic: 8.3 RBS%

11.  WWWWWWWWWWWW WWWWWWWWWWWW: 10.0 RBS%

12.  Olober Syko: 10.4 RBS%

13.  Matthew McCagg: 10.9 RBS%

14.  Jonny Elgar: 12.8 RBS%

15.  Joel Castle: 13.0 RBS%

16.  Georgiy Costanzov: 15.4 RBS%

17.  Ryan Artyomov: 19.6 RBS%

 

Again, Fuuka Rask and Lachlan Summers are among the most reliable goaltenders, this time joined at the top by the likes of Dalkr Vidarsson, Merome Dilson or Justin Lion.

 

Standard Deviation

 

Okay, that's not a hockey stat, but a general statistics tool. While the “official” stats describe how often a goaltender is good, or how often they're not bad, a Standard Deviation only describes the tendency to deviate from the average. For this purpose, it's not important what exactly does the numeric value mean. Unless you have a higher education in maths or statistics, the number itself won't probably tell you much. But it's the comparison of these values that matters. Higher number means higher tendency to have uncharacteristically good or bad games, lower number means more consistent results. 

 

A goaltender with a high save percentage and high standard deviation is a star goaltender who has consistency issues. A goaltender with a lower save percentage and lower standard deviation is a lower-end goaltender who is less likely to surprise you either way.

 

Goaltenders ranked from the most consistent.

 

1.  Lachlan Summers: SD 0.040

2.  Red Panda: SD 0.042

3.  Merome Dilson: SD 0.043

4.  Dalkr Vidarsson: SD 0.044

5.  Ash Sparks: SD 0.047

6.  WWWWWWWWWWWW WWWWWWWWWWWW: SD 0.047

7.  Fuukka Rask: SD 0.048

8.  Amir Redzic: SD 0.050

9.  Matthew McCagg: SD 0.052

10.  Herald Benson: SD 0.056

11.  Jonny Elgar: SD 0.057

12.  Joel Castle: SD 0.058

13.  David Slezák: SD 0.058

14.  Olober Syko: SD 0.061

15.  Georgiy Costanzov: SD 0.066

16.  Ryan Artyomov: SD 0.074

17.  Justin Lion: SD 0.092

 

Is it even surprising, that on top of sporting the league's best save percentage, leading the league in Goals Saved Above Average and Quality Starts Percentage while having one of the lowest Really Bad Starts Percentages, Lachlan Summers was also the league's most consistent goaltender? Lachlan Summers’ S94 performance might have been one of the greatest goaltending masterclasses the VHL has ever seen.

 

It looks like the league actually knows what they're talking about, when voting awards winners.

 

A little over 1k words with the tags and everything, claiming 2×6 capped.

[VattghernCZ 983]

Claim 1/2 for W.E. 8/4

[VattghernCZ 983]

Claim 2/2 for W.E. 8/11

Edited August 4 by VattghernCZ
Too many numbers today

[rory 1,916]

2 minutes ago, VattghernCZ said:

 

It looks like the league actually knows what they're talking about, when voting awards winners.

xD

[kirbithan 1,712]

so. many. numbers. but also so interesting to see how all these goalers stack up!!

 

incredible work, vattghern! sorry not sorry for summers terrorizing you our entire careers.

[VattghernCZ 983]

2 minutes ago, kirbithan said:

so. many. numbers. but also so interesting to see how all these goalers stack up!!

 

incredible work, vattghern! sorry not sorry for summers terrorizing you our entire careers.

 

Lachlan's face will always have a spot on Kerr's dartboard!

[ShawnGlade 1,010]

Holy numbers.

 

For Part 3 I'd love to see some sort of comparison of SV% or GSAA compared to their teams scoring rates, to look at which goaltenders played a bigger part in their team's success (or lack thereof)

[Jubis 1,349]

Summers is a beast!!!

Glad Rask was able to slightly hang with him in this last season  

[sadie 904]

Awesome article, I love seeing data like this about the VHL! I can tell how much work went into it!

[Gaikoku-hito 2,089]

3 hours ago, VattghernCZ said:

Quality Starts is a piece of statistics developed by Rob Vollman in the Hockey Abstract

Actually not true as he was simple one of the first to publish data collection in which teams held in secret as teams actually all have different criteria in which to look at a goaltenders performance as in the article you are using league average and with a save percentage of at least .885 in games with fewer than 20 shots against but truth be told most teams use a flat .900 or higher plus league average to determine the stronger goaltenders in the league. But this still doesn`t account for Quality of shots against; which is also tracked versus Quality of goals allowed, aka high danger shots leading to either a goal or a save so a goaltender might have a low save percentage in a game but that is due to the fact that his team in front of him cannot play defense and allow a high number of point blank scoring chances.

 

Using the Quality of shots versus Quality of goals usually add a layer of if the goaltender is playing better than expect or worse by adding the two percentage together to create a personal average also know as the expected save%. So instead of using the flat .900sav% or .885sav%; you would use the actually save% against the goalies expected save%.

 

John Gibson (since it was one of the first I could find) had a .906 expected Sav% and his actual sav% was .911 (numbers from 2021). These numbers standing alone don`t tell the whole story as you need to compare these directly with is teammate (back-up) and the rest of the league!!

 

Great article overall!!

 

 

 

 

 

 

[VattghernCZ 983]

7 hours ago, Gaikoku-hito said:

Actually not true as he was simple one of the first to publish data collection in which teams held in secret as teams actually all have different criteria in which to look at a goaltenders performance as in the article you are using league average and with a save percentage of at least .885 in games with fewer than 20 shots against but truth be told most teams use a flat .900 or higher plus league average to determine the stronger goaltenders in the league. But this still doesn`t account for Quality of shots against; which is also tracked versus Quality of goals allowed, aka high danger shots leading to either a goal or a save so a goaltender might have a low save percentage in a game but that is due to the fact that his team in front of him cannot play defense and allow a high number of point blank scoring chances.

 

Using the Quality of shots versus Quality of goals usually add a layer of if the goaltender is playing better than expect or worse by adding the two percentage together to create a personal average also know as the expected save%. So instead of using the flat .900sav% or .885sav%; you would use the actually save% against the goalies expected save%.

 

John Gibson (since it was one of the first I could find) had a .906 expected Sav% and his actual sav% was .911 (numbers from 2021). These numbers standing alone don`t tell the whole story as you need to compare these directly with is teammate (back-up) and the rest of the league!!

 

Great article overall!!

 

 

 

 

 

 

 

Quality of shots and anything regarding expected goals is impossible to do without shot map tho. I'd very much like to do a part with HD/MD/LD Sv% and GSAx lol

Edited August 5 by VattghernCZ