Stoppable Goals Percentage

[542365]

Hey guys. There is a user by the name of Bleedred on the By the Numbers forum who is reviewing every goal scored throughout the year and determining if he considers the goal stoppable by the goalie. He gives his definition of a stoppable goal which you can read in his thread, but the idea is pretty obvious. Should a team reasonably expect the goalie to make that save. His goal is to see the correlation between what he considers to be stoppable goals and save percentage at the end of the year. I thought it would be an interesting exercise so I'm going to start where he left off and track Jake/Chad/Ville/Jordan/Whomever and see by the end of the year what percentage of goals scored against them were what I consider to be stoppable. Hopefully he continues doing every goalie so we can compare at the end to see who is the harsher judge

So far he has:
Jake Allen-.879 save percentage; 41 goals against, 14 stoppable(34% of his goals against were stoppable)
Chad Johnson-.872 save percentage; 6 goals against, 2 stoppable.(33% of his goals against were stoppable)

I haven't gone back and reviewed each goal just yet, but I will do that when I get home and see if I disagree on these numbers.

I really don't know what I'm hoping to accomplish with this, but I think it would be something interesting to discuss throughout the year.

So far I consider the following goals to be unstoppable:
Any clear cut breakaway. Yes you would like to see your goalie stop some of these(especially against lower line players), but if there is literally nobody back to help him out I won't fault the goalie.

Deflections near the goal mouth or radical changes of direction(like Staal's goal against Minnesota). Most difficult shot to stop, even if the goalie doesn't play it well(like how Jake was so deep in his net against Staal) I still won't fault a goalie most times. The deflection goal against Vegas that made it 2-2 I do fault the goalie. It was a low shot that got deflected slightly in a different direction and went five hole. It's not like it was going high and then got deflected down five hole. It was low the whole way and I think should have been stopped.

Side to side one timer snipes. If a guy goes bar down on a one timer from the faceoff dot I'm not going to hold that against the goalie. That's just great shooters making great shots. If they're beaten cleanly on a shot to the middle of the net I might still count it as stoppable.

Everything else is going to be a judgement call. If it's 50/50 I'm going to lean towards calling it unstoppable. I want at the end of the year to see our goalies' percentages on these types of shots and compare them with the rest of the league(hopefully Bleedred keeps it up) to see how many goals we might have expected to keep out with another goaltender.

It's not really meant to be some scientific study or anything, more of just a fun exercise that should generate some pretty intense discussions throughout the year.

EDIT TO ADD: I'm going to try to take context out of the equation entirely. A tying goal with 20 seconds left in the game counts the same as one that cuts our lead to 5-1 in the third. If it's stoppable, it's stoppable. The context of the shooter is going to be more challenging. Of course a shot by Ovi is going to be more difficult to stop than a shot by Craig Smith, but I'm going to try to just look at where the puck entered the net and the circumstances leading up to the shot to determine if it's stoppable or not.

I will also try to wait until the day after a game to review the video so my numbers won't be badly skewed by my emotions.

Click to read more...

 

[Stupendous Yappi]

It would also be nice to know the save percentage for 'unstoppable' shots, if you're going to bother. But its more work. It just would give a more rounded view of the value of the goalie.

 

[carter333167]

Hey guys. There is a user by the name of Bleedred on the By the Numbers forum who is reviewing every goal scored throughout the year and determining if he considers the goal stoppable by the goalie. He gives his definition of a stoppable goal which you can read in his thread, but the idea is pretty obvious. Should a team reasonably expect the goalie to make that save. His goal is to see the correlation between what he considers to be stoppable goals and save percentage at the end of the year. I thought it would be an interesting exercise so I'm going to start where he left off and track Jake/Chad/Ville/Jordan/Whomever and see by the end of the year what percentage of goals scored against them were what I consider to be stoppable. Hopefully he continues doing every goalie so we can compare at the end to see who is the harsher judge

So far he has:
Jake Allen-.879 save percentage; 41 goals against, 14 stoppable(34% of his goals against were stoppable)
Chad Johnson-.872 save percentage; 6 goals against, 2 stoppable.(33% of his goals against were stoppable)

I haven't gone back and reviewed each goal just yet, but I will do that when I get home and see if I disagree on these numbers.

I really don't know what I'm hoping to accomplish with this, but I think it would be something interesting to discuss throughout the year.

So far I consider the following goals to be unstoppable:
Any clear cut breakaway. Yes you would like to see your goalie stop some of these(especially against lower line players), but if there is literally nobody back to help him out I won't fault the goalie.

Deflections near the goal mouth or radical changes of direction(like Staal's goal against Minnesota). Most difficult shot to stop, even if the goalie doesn't play it well(like how Jake was so deep in his net against Staal) I still won't fault a goalie most times. The deflection goal against Vegas that made it 2-2 I do fault the goalie. It was a low shot that got deflected slightly in a different direction and went five hole. It's not like it was going high and then got deflected down five hole. It was low the whole way and I think should have been stopped.

Side to side one timer snipes. If a guy goes bar down on a one timer from the faceoff dot I'm not going to hold that against the goalie. That's just great shooters making great shots. If they're beaten cleanly on a shot to the middle of the net I might still count it as stoppable.

Everything else is going to be a judgement call. If it's 50/50 I'm going to lean towards calling it unstoppable. I want at the end of the year to see our goalies' percentages on these types of shots and compare them with the rest of the league(hopefully Bleedred keeps it up) to see how many goals we might have expected to keep out with another goaltender.

It's not really meant to be some scientific study or anything, more of just a fun exercise that should generate some pretty intense discussions throughout the year.

EDIT TO ADD: I'm going to try to take context out of the equation entirely. A tying goal with 20 seconds left in the game counts the same as one that cuts our lead to 5-1 in the third. If it's stoppable, it's stoppable. The context of the shooter is going to be more challenging. Of course a shot by Ovi is going to be more difficult to stop than a shot by Craig Smith, but I'm going to try to just look at where the puck entered the net and the circumstances leading up to the shot to determine if it's stoppable or not.

I will also try to wait until the day after a game to review the video so my numbers won't be badly skewed by my emotions.

That's pretty interesting. When I try to give a goalie the eye test, I focus really on positioning, how he transitions from position to position and whether shots are going through his body (i.e., 5 hole, under the arm) and how well the goalie is seeing and tracking the puck.

 

[Stealth JD]

All shots should be "stoppable". They'd only ever become unstoppable if the goalie put himself in such poor position in the first place to give up a wide open net. Somehow I feel as though Jake would benefit from anything that is considered an "unstoppable" goal, when it should be 100% the worst kind of goal to give up.

 

[542365]

It would also be nice to know the save percentage for 'unstoppable' shots, if you're going to bother. But its more work. It just would give a more rounded view of the value of the goalie.

You're right it would definitely give a more rounded view and would be more valuable to study, but I just don't have that kind of time. That would involve watching every shot on goal from every game rather than just reviewing the goals. I maybe could try to add up the amount of "big saves" the goalie makes in a season. That doesn't seem impossible, but it would require me to pay super close attention at all times, which admittedly I don't do

All shots should be "stoppable". They'd only ever become unstoppable if the goalie put himself in such poor position in the first place to give up a wide open net. Somehow I feel as though Jake would benefit from anything that is considered an "unstoppable" goal, when it should be 100% the worst kind of goal to give up.

I guess "stoppable" isn't the right term for it. Of course every goal is possible to stop other than empty netters. I guess it's more of an eye test version of expected save percentage. Which goals can you reasonably expect a goalie to make. Sure a goalie might be totally screened and guess right with his glove, but if it goes off the post and in after a wicked screen in front, I don't know that it would be reasonable to expect a goalie to stop it. Like Tarasenko's most recent goal against Chicago with Maroon right in Crawford's face. Tarasenko picks the far side post. I don't think that's a shot you can reasonably expect your goalie to stop. If Maroon isn't there then it becomes more of a judgement call since it was a perfect shot, but not from the greatest spot on the ice. I don't know. It's tough. I don't claim this to be some new advanced statistic or something. Really just more of a fun side project while I'm watching Blues games. Something has to keep me entertained since this team surely isn't.

 

[steinerecliner]

I think what should be viewed is how many are what a lot of people refer to as "soft" goals, which those people think those should been easy saves but they gone by. There are a lot of goals that doesn't just sit on the goalie but bad defensive side of things. You can't give up 40+ shots and expect every time the goalie to make all of them, that becomes a point that the goalie is just worn out because he doesn't have a chance to catch his breath because the puck is always in the zone and coming at him. Goalies don't get line changes and when you always having to take shots, it gets to any goalie. Johnson only made one start and came in twice, hard to judge him to much. The team has to play better defense to keep the shots down.

 

[542365]

So I went back and reviewed every goal against this year(depressing shit) and I guess I'm a bit tougher of a judge than Bleedred. My results:

Jake Allen-.879 save percentage; 17 of 41 goals stoppable(41% stoppable)
Chad Johnson-.872 save percentage; 3 out of 6 goals stoppable(50% stoppable)

I'm going to post my comments on every game and I'd like to hear the disagreements. I've found rebound goals to be the most difficult to judge. Yes the shot was very difficult to stop, but the entire reason there was a shot in that area was due to what I perceive to be a bad rebound given up. I also had a tough time with short side shots from close in. I understand that it's a scoring area, but being beat shot side, especially glove side, is stoppable in my opinion. I don't know. Again, it's tough and it's subjective so I'd like to hear your guys' thoughts on this. Just a heads up that this is pretty quick and dirty work on a Google doc. I'll clean it up in a spreadsheet eventually. I just wanted to get my starting numbers before tomorrow night's game.

10/4 Jets
1 Stoppable goal out of 5(5th goal)

10/6 Hawks
3 stoppable goals out of 5(1st, 3rd, and 4th)

10/11 Flames
0 of 3

10/13 Hawks
1 stoppable goals out of 4(1st)

10/14 Ducks(Johnson)
1 stoppable goal out of 3(3rd)

10/17 Habs
2 stoppable goals out of 3(1st and 3rd)

10/20 Leafs
0 of 1

10/22 Jets
3 stoppable goals out of 5(1st, 2nd, 4th)

10/25 Jackets
1 of 4 on Allen(2nd)
2 stoppable goals of 3 on Johnson(6th and 7th)

10/27 Hawks
3 stoppable goals of 3

11/1 Knights
2 stoppable goals out of 3(1st, 2nd)

11/3 Wild
1 stoppable goal of 5(4th)

I'll be using my own numbers from here on out and I'll post an update a day after every game.

 

[rumrokh]

Save percentages are ridiculously high when even most bad goalies are set. So it's basically whether or not the puck moved immediately beforehand (rebound, deflection, one-timer, or fairly quick settle-and-shot off a pass), if there was a massive screen, or if someone is shooting from the slot (breakaway). If the goalie was set and a guy just picked a corner with a crazy shot, props to him. If the goalie wasn't set, should he have been? Was he set, but went down early? If it's a shot from a dangerous area, but the goalie is set, does it go through him?

 

[Ranksu]

I was just wondering have we set Allen's bar a bit high? (level of performance)

Are waiting too much from Allen?

 

[Bjornar Moxnes]

Well there's already something known as expected goals against and expected save percentage.

 

[Majorityof1]

Well there's already something known as expected goals against and expected save percentage.

Expected goalie stats are based on the number of High vs low danger chances. The danger level is solely determined based on where the shot is taken from and ignores all other circumstances of the shot. If a defender misplays a 2-on-1, knocks the goalie over and gives the other person a wide-open shot, that is the same danger level as one of Maroon's jam plays into the goalies pads. But one save is impossible and one should be made. Likewise, a shot from the point is low-danger, regardless of whether the goalie has 15 guys in front of him or if the defender tips it in.

The analysis 542365 is doing, I assume, attempts to look at the actual circumstances of each goal, not just where the shot originated. So it is different than Expected Value stats. I personally am not sure if that type of anlysis has any value without comparing it league wide, and without a standard of what constitutes a "saveable" chance. But if the OP wants to do it, we are free to accept, ignore or counter his conclusions.

 

[BadgersandBlues]

For those of you who are interested, the Athletic had a really interesting interview with a guy named Steve Valiquette who used to be (still is?) on the Rangers' broadcast as a between periods analyst. He started a company that literally tracks every shot taken for the last couple of years and then scores the shot based on how dangerous it is. Link to the article:

Q&A: Steve Valiquette's radical approach to evaluating shots...

For those of you who don't have a sub, the premise of the article is how dangerous is any given shot for an NHL goalie against NHL competition. What he finds can be summed up as follows:

• If the goalie can see the shot coming for at least .5 seconds before it is released from the stick, it's a bad shot. This should be saved 97% of the time, pretty much no matter where the shot is coming from. "Clear sight" shots are bad, and generally should be saved.
• The best shots are double screens (your team's forward and the opposing team's d-man both in front of the goalie jostling for position) with a deflection.
• A shot that follows a pass across the "Royal Road" (Basically think if a Zamboni drove straight out to center ice from the goal mouth, the resulting fresh ice is considered the Royal Road) if a shot follows TWO passes across the Royal road, it's got the highest % chance of any shot.
• Breakaways are good, but statistically not the best scoring chance in the game.

It's a pretty interesting read. What's really interesting is on the topic of a breakaway, you have only about a 1 in 3 chance of scoring. He also says that that % holds true no matter the level of competition. So if an AHL goalie is going against an AHL shooter, it's 1 in 3. The only way to change that % is to have an AHL goalie against an NHL shooter.

Here's a link to a twitter feed showing how he "scores" the shots taken in a game, which should be available even if you don't sub to the Athletic:



Red shots are bad, Yellow are medium, and Green shots are good.

Anyway, thought I'd post it here, since we're talking about how to better quantify shots and goals against beyond simply stating, oh, this shot was close in, that's high danger.

 

[542365]

For those of you who are interested, the Athletic had a really interesting interview with a guy named Steve Valiquette who used to be (still is?) on the Rangers' broadcast as a between periods analyst. He started a company that literally tracks every shot taken for the last couple of years and then scores the shot based on how dangerous it is. Link to the article:

Q&A: Steve Valiquette's radical approach to evaluating shots...

For those of you who don't have a sub, the premise of the article is how dangerous is any given shot for an NHL goalie against NHL competition. What he finds can be summed up as follows:

• If the goalie can see the shot coming for at least .5 seconds before it is released from the stick, it's a bad shot. This should be saved 97% of the time, pretty much no matter where the shot is coming from. "Clear sight" shots are bad, and generally should be saved.
• The best shots are double screens (your team's forward and the opposing team's d-man both in front of the goalie jostling for position) with a deflection.
• A shot that follows a pass across the "Royal Road" (Basically think if a Zamboni drove straight out to center ice from the goal mouth, the resulting fresh ice is considered the Royal Road) if a shot follows TWO passes across the Royal road, it's got the highest % chance of any shot.
• Breakaways are good, but statistically not the best scoring chance in the game.

It's a pretty interesting read. What's really interesting is on the topic of a breakaway, you have only about a 1 in 3 chance of scoring. He also says that that % holds true no matter the level of competition. So if an AHL goalie is going against an AHL shooter, it's 1 in 3. The only way to change that % is to have an AHL goalie against an NHL shooter.

Here's a link to a twitter feed showing how he "scores" the shots taken in a game, which should be available even if you don't sub to the Athletic:



Red shots are bad, Yellow are medium, and Green shots are good.

Anyway, thought I'd post it here, since we're talking about how to better quantify shots and goals against beyond simply stating, oh, this shot was close in, that's high danger.

This is really good stuff.

I feel like I don't understand the "clear sight" shot though. I feel like a shot from the slot, even if the goalie can see it clearly, is saved much less than 97% of the time, more like 75% at best. I know that number encapsulates all clear sight shots from all over the ice, but it seems pretty rare for players to take a shot like that from just about anywhere but the slot unless they're just trying to generate a rebound. What am I missing here?

I really like the charts. This is exactly the type of information I'm looking for. NHL needs a statcast type system. I would never stop looking at the data.

 

[542365]

11/6 Canes(Johnson)
0 Stoppable Goals out of 1

Season total:
Allen-.879 save percentage; 17 of 41 goals stoppable(41% stoppable)
Johnson-.919 save percentage; 3 of 7 goals stoppable(43% stoppable)

 

[Brian39]

Good luck with this analysis. I participated in a project similar to this years ago and it turned into a disaster. My biggest takeaway is that it is really, really hard to quantify scoring chances with strict criteria. To paraphrase Justice Stewart's take on pornography, 'perhaps I could never succeed in defining pornography,l but I know it when I see it.' This is such a fluid game that it is really hard to come up with firm criteria to implement that actually works.

This was years ago and I think technology is helping us get there. Clear Sight Analytics is doing awesome work and that Athletic article is a must read if you're into hockey data.

Good luck if you keep this up. I hope you don't get as frustrated with the criteria as I did years ago.

 

[542365]

Don’t need to wait on this one

11/9 Sharks(Johnson)
Nothin’. Shutout!

Season total:
Jake Allen: .879 save percentage;17 of 41 stoppable goals(41% stoppable).
Johnson:.941 save percentage; 3 of 7 goals stoppable(43% stoppable).

 

[Ranksu]

To paraphrase Justice Stewart's take on pornography, 'perhaps I could never succeed in defining pornography,l but I know it when I see it.'

You must be fan of soft porn

 

[542365]

11/12 Wild(Johnson
1 stoppable goal out of 3(1st)

Season total:
Jake Allen: .879 save percentage;17 of 41 stoppable goals(41% stoppable)
Chad Johnson: .927 save percentage; 4 out of 8 goals stoppable(50% stoppable)

I went back and forth on the third goal. It's one I think you want your goalie to stop, but when your Dman gives the puck up right in the slot to their best player I don't think it's one you expect him to stop necessarily. I said at the beginning that if it's borderline I'm going with unstoppable. Sticking to that.

 

[ChuckLefley]

11/12 Wild(Johnson
1 stoppable goal out of 3(1st)

Season total:
Jake Allen: .879 save percentage;17 of 41 stoppable goals(41% stoppable)
Chad Johnson: .927 save percentage; 4 out of 8 goals stoppable(50% stoppable)

I went back and forth on the third goal. It's one I think you want your goalie to stop, but when your Dman gives the puck up right in the slot to their best player I don't think it's one you expect him to stop necessarily. I said at the beginning that if it's borderline I'm going with unstoppable. Sticking to that.

Did you do the same with every golden chance our D gave away with Allen in net?

 

[542365]

Did you do the same with every golden chance our D gave away with Allen in net?

Yes I did. Nick Foligno had a very similar goal against us earlier this year. Puck in the slot from 20 feet or so high short side. I called it unstoppable, though it's one I think you want your goalie to have.

 

[BadgersandBlues]

Out of curiosity, which goal did you feel was stoppable? I wouldn't ding him for any of those three.

 

[542365]

Out of curiosity, which goal did you feel was stoppable? I wouldn't ding him for any of those three.

First one. I thought that Johnson should’ve tracked the initial shot from Nino and redirected it into the corner of caught it in his glove. It was a very low velocity shot that he just lost sight of.

 

[542365]

I usually like to wait until the day after but this one is pretty easy in my book.

11/14 Hawks
0 of 1 goals stoppable.


Season total:
Jake Allen: .883 save percentage; 17 of 42 goals stoppable. (40% stoppable)
Chad Johnson: .927 save percentage; 4 out of 8 goals stoppable(50% stoppable).

 

[BadgersandBlues]

First one. I thought that Johnson should’ve tracked the initial shot from Nino and redirected it into the corner of caught it in his glove. It was a very low velocity shot that he just lost sight of.

I'd disagree personally. I think some of the hardest shots for goalies are the ones where a stick hits right as the guy is releasing his shot. In addition, the puck deflected off our d-man in front of the net too. I was actually amazed Johnson almost got his blocker on it at all.

Just my .02 though.

 

[542365]

11/16 Knights
0 of 1 goals stoppable.

Season total:
Jake Allen: .890 save percentage; 17 of 43 goals stoppable.(39.5% stoppable)
Chad Johnson: .927 save percentage; 4 out of 8 goals stoppable(50% stoppable)