Stoppable goals correlation with save percentages

[TomasHertlsRooster]

Okay, so, I just ran a quick regression to determine the impact of Stoppable Goals on Delta Save Percentage. Here's what I got:

For every 1 percentage by which stoppable goals percentage increases, delta save percentage decreases by 0.04%. Here is what this looks like on a graph:

Statistically speaking, the p value of this is 0.2108, which says that there is not a statistically significant correlation between the two.

However, I can't help but feel like it's no coincidence that Bleedred looks at goaltenders and says "a large percentage of the goals they allow are stoppable", and a metric calculated based off the location of shots these goaltenders face says "these goaltenders stop a smaller percentage of the shots against them than they should."

 

[Bleedred]

Okay, so, I just ran a quick regression to determine the impact of Stoppable Goals on Delta Save Percentage. Here's what I got:

For every 1 percentage by which stoppable goals percentage increases, delta save percentage decreases by 0.04%. Here is what this looks like on a graph:

Statistically speaking, the p value of this is 0.2108, which says that there is not a statistically significant correlation between the two.

However, I can't help but feel like it's no coincidence that Bleedred looks at goaltenders and says "a large percentage of the goals they allow are stoppable", and a metric calculated based off the location of shots these goaltenders face says "these goaltenders stop a smaller percentage of the shots against them than they should."

I just now posted this in the out of town scoreboard thread of the Devils forum. Very small sample of goalies used here, but all are between a .901% and .908% save percentage. 8 of the 13 are between 27% and 29% of goals allowed being stoppable. 11 of the 13 are between 27% and 31%.

I only did this for the goalies in the same neighborhood as Keith Kinkaid and the percentages are very close. With Marc Andre-Fleury and Phillipp Grubauer being the two guys that are under that range. Fleury being 25.5% and Grubauer at 22.5%. These are for goalies with between a .901% and a .908% save percentage and a minimum of 12 games played. It's really useless to include guys who have played fewer games than that, as these numbers were all over the place early on. I don't think it's a coincidence that they're all in the same neighborhood. Fleury and Grubauer have been a little bit better on the eye test though.

Keith Kinkaid: 27 games played, .902% save percentage, 77 goals allowed, 22 stoppable goals allowed. 28.5% of goals allowed stoppable.
Carey Price: 28 games played, .902% save percentage. 79 goals allowed, 23 stoppable goals allowed. 29% of goals allowed stoppable.
Martin Jones: 25 games played, .901% save percentage, 68 goals allowed, 20 stoppable goals allowed. 29% of goals allowed stoppable.
Jacob Markstrom: 25 games played, .904% save percentage, 76 goals allowed, 21 stoppable goals allowed. 27.5% of goals allowed stoppable.
Louis Domingue: 18 games played. .905% save percentage, 53 goals allowed, 15 stoppable goals allowed. 28.5% of goals allowed stoppable.
Sergei Bobrovsky: 24 games played, .905% save percentage, 65 goals allowed, 20 stoppable goals allowed. 31% of goals allowed stoppable.
Roberto Luongo: 15 games played, .906% save percentage, 40 goals allowed, 11 stoppable goals allowed. 27.5% of goals allowed stoppable.
Craig Anderson: 30 games played, .907% save percentage, 99 goals allowed, 27 stoppable goals allowed. 27.5% of goals allowed stoppable.
Connor Hellebuyck: 26 games played, .908% save percentage, 73 goals allowed, 22 stoppable goals allowed. 30% of goals allowed stoppable.
Antti Raanta: 12 games played, .906% save percentage, 33 goals allowed, 10 stoppable goals allowed. 30.5% of goals allowed stoppable.
Corey Crawford: 23 games played, .902% save percentage, 71 goals allowed, 19 stoppable. 27% of goals allowed stoppable.
Phillipp Grubauer: 14 games played, .904% save percentage, 40 goals allowed, 9 stoppable goals allowed. 22.5% of goals allowed stoppable.
Marc-Andre Fleury: 31 games played, .908% save percentage, 79 goals allowed, 20 stoppable goals allowed. 25.5% of goals allowed stoppable.

Now these are only for guys near Keith Kinkaid in save percentage and a minimum of 12 games played. All of these guys are between a .901% and .908% save percentage and 11 of them are between 27% and 31% of stoppable goals for percentage by my count. With 8 of the 13 being between 27% and 29% stoppable goals for percentage. So it's pretty close. Fleury and Grubauer are the two outliers in stoppable goals percentage in the same neighborhood save percentage wise.

 

[Bleedred]

bleedred, this is an awesome project. thank you for pushing the envelope. not sure i get the summary report of results though. the trouble with the stats as you state them for comparison is that it does not account for other variables (such as goalies who see proportionately more unstoppable goals).

wouldn't it be logical to report a "stoppable shot save percentage" that ignores unstoppable goals either as goals or shots and then restates a save percentage of just the stoppable shots. every goalie's save % then goes up, but some will go up more than others.

would also be very interesting to see how teams stack up on allowing unstoppable goals.

I appreciate your interest! Thanks!

As far as ''Stoppable shot save percentage'' I just can't do it, because it would be too time consuming. I'd have to watch every that's taken and every that's made. It's just not realistic, honestly. And I do have center ice, but it's impossible to watch every game, when 5 or 6, sometimes 7 or 8 are going on at the same time. So all I really have to go on for the games I don't actually watch are the highlight videos.

 

[TomasHertlsRooster]

I appreciate your interest! Thanks!

As far as ''Stoppable shot save percentage'' I just can't do it, because it would be too time consuming. I'd have to watch every that's taken and every that's made. It's just not realistic, honestly. And I do have center ice, but it's impossible to watch every game, when 5 or 6, sometimes 7 or 8 are going on at the same time. So all I really have to go on for the games I don't actually watch are the highlight videos.

Every save is a stoppable shot, correct? There are no “unstoppable shots” that get saved.

So, one could easily calculate “stoppable shot” percentage based on this data.

 

[Bleedred]

Every save is a stoppable shot, correct? There are no “unstoppable shots” that get saved.

So, one could easily calculate “stoppable shot” percentage based on this data.

Technically, I'm sure there are unstoppable shots that get saved, but like I said, that's a whole other project.

Or maybe I'm just not following you?

 

[TomasHertlsRooster]

Technically, I'm sure there are unstoppable shots that get saved, but like I said, that's a whole other project.

Or maybe I'm just not following you?

I mean, if you just pretend those don’t exist...

If a goalie allows 10 goals on 105 shots, but 5 of those were considered “not stoppable goals”, then we would subtract those goals/shots. Then, we would have 5 stoppable goals on 100 stoppable shots, or a “stoppable shot save percentage”.

 

[Bleedred]

I mean, if you just pretend those don’t exist...

If a goalie allows 10 goals on 105 shots, but 5 of those were considered “not stoppable goals”, then we would subtract those goals/shots. Then, we would have 5 stoppable goals on 100 stoppable shots, or a “stoppable shot save percentage”.

I could definitely do something like that, only with how much time I spend watching the highlights and updating the file (it takes me over an hour on nights where there’s 11 or more games) I think it might be best to tally those up at year’s end.

There’s still some guys that have allowed a lot of stoppable goals for how high their save percentage is and vice versa. Halak, Bishop, Koskinen and Rittich all come to mind. Though many of Rittich’s stoppable goals came early on. I haven’t added many to him lately. And on the lower end, Korpisalo and Saros are two names that come to mind that have a low stoppable goal count and low save percentages. Crawford had a very low stoppable goal percentage for a long time this season, but they’ve literally caught up to him a lot in the last 3 weeks or so.

Rinne, Andersen, Gibson all have a pretty low percentage in the lower 20 percents. Lundqvist did for a while, but it’s catching up a little to him. Varlamov’s was very low, but has added up a bit and like Lundqvist, his save percentage has fallen to coincide with it. Although both are still above average by quite a bit, in a year where we are seeing save percentages at an 11 or 12 year low. League average is only .907% this year.

Craig Anderson leads the league in these stoppable goals with 27, but not in percentage. He’s allowed the most goals, faced the most shots and I think he’s played the most games and minutes behind only Fleury, who recently passed him in games/minutes played. Allen, Smith and Talbot are all a bit over and into the 30’s as far as percentage goes. Not 30+ stoppable goals yet.

Scott Darling and Cory Schneider are around 45% right now, though in very limited starts/games played.

 

[Bleedred]

There’s probably several different things we could use and have some fun with after the results at season’s end

I definitely expect the numbers to a little more congruent with the save percentages for the guys who played 50-60 games as opposed to the guys that only played 25-30.

 

[krutovsdonut]

I appreciate your interest! Thanks!

As far as ''Stoppable shot save percentage'' I just can't do it, because it would be too time consuming. I'd have to watch every that's taken and every that's made. It's just not realistic, honestly. And I do have center ice, but it's impossible to watch every game, when 5 or 6, sometimes 7 or 8 are going on at the same time. So all I really have to go on for the games I don't actually watch are the highlight videos.

sorry i just meant to use your existing results and off the shelf data

i was hoping you could generate an adjusted save % for goalies as follows:

shots = total shots faced - unstoppable goals

saves = total saves

from this you get an adjusted save % that only measures goalie performance against shots that the goalie arguably could have stopped. as a result, every goalies' save percentage will go up, but some will go up more than others. theoretically, some good goalies on bad teams will see an improvement. what will be interesting is whether there are any significant risers, which would potentially indicate good goalies being hung out to dry by their teams.

 

[krutovsdonut]

I mean, if you just pretend those don’t exist...

If a goalie allows 10 goals on 105 shots, but 5 of those were considered “not stoppable goals”, then we would subtract those goals/shots. Then, we would have 5 stoppable goals on 100 stoppable shots, or a “stoppable shot save percentage”.

exactly!

 

[TomasHertlsRooster]

There’s probably several different things we could use and have some fun with after the results at season’s end

I definitely expect the numbers to a little more congruent with the save percentages for the guys who played 50-60 games as opposed to the guys that only played 25-30.

Yeah, I’m excited to see the correlation between Delta Save Percentage, plain save percentage, and expected save percentage by season’s end.

 

[krutovsdonut]

made couple of manual calculations using your post above to compare goalies and there is a noticable variation based on four examples. hellebuyck drops .005% compared to fleury on stoppable saves when they have the same raw sv%.

craig anderson .907 after 30 games
1070 shots - 72 unstoppable goals = 998 stoppable shots
971 saves/998 = .973 stoppable sv%

connor hellebuyck .908 after 26 games
795 shots - 51 unstoppable goals = 744 stoppable shots
722 saves/744= .970 stoppable sv%

marc-andre fleury .908 after 31 games
850 shots - 59 unstoppable goals = 791 stoppable shots
771 saves/791 = .975 stoppable save %

jacob markstrom - .904 after 25 games:
790 saves - 55 unstoppable goals = 735 stoppable shots
714 saves/735 = .971 stoppable sv%

 

[Bleedred]

made couple of manual calculations using your post above to compare goalies and there is a noticable variation based on four examples. hellebuyck drops .005% compared to fleury on stoppable saves when they have the same raw sv%.

craig anderson .907 after 30 games
1070 shots - 72 unstoppable goals = 998 stoppable shots
971 saves/998 = .973 stoppable sv%

connor hellebuyck .908 after 26 games
795 shots - 51 unstoppable goals = 744 stoppable shots
722 saves/744= .970 stoppable sv%

marc-andre fleury .908 after 31 games
850 shots - 59 unstoppable goals = 791 stoppable shots
771 saves/791 = .975 stoppable save %

jacob markstrom - .904 after 25 games:
790 saves - 55 unstoppable goals = 735 stoppable shots
714 saves/735 = .971 stoppable sv%

Fleury has been weird for me from what I counted. I think he's played better than other guys in his save percentage group this year, when it comes to the eye test.

This obviously won't be perfect, but I do find it funny that the guys that have been some of the worst in the league, have a higher percentage of stoppable goals than the guys that have been some of the best in the league. For the most part.

 

[krutovsdonut]

Fleury has been weird for me from what I counted. I think he's played better than other guys in his save percentage group this year, when it comes to the eye test.

This obviously won't be perfect, but I do find it funny that the guys that have been some of the worst in the league, have a higher percentage of stoppable goals than the guys that have been some of the best in the league. For the most part.

well the data supports your eye test.

if you get a chance, i would love to see a comparison of all goalies at mid season using that formula