Post-Game Talk: Point taken

[Missing smitty]

Your a math guy... a stats nerd told me the general accuracy of the models are limited to about 85%. That's a lot of noise among that signal.

If I were a GM, that might be useful enough to spot a trend in my team game... but it wouldnt' be enough for me to overrule my coaches eye test or send my pro-scouts home when it comes to making decisions about individual players.

The oilers goalie coach pushed for Jack Campbell. I'm not sure they should be trusting anything coming from him...

 

[CupofOil]

Skinner leaked in 2 goals and the Oilers stopped playing since. Sure Skinner played better in the game but he put the Oilers in a bad position in the first place

Leaked in 2 goals? The Wings best forward had a breakaway because Kulak was in lala land, how was that a "leaked in" goal? The first one was a deflection, bad break. I swear, some of you just invent reasons out of thin air to hate the guy, it's weird. Look at the 2nd goal Lyon gave up, now THAT was a leaky goal.
There's plenty of reasons this season to criticize Skinner, last night was not one of them.

 

[KeithIsActuallyBad]

Neither of those goals were the least bit leaky. Both were deflected.

In particular Skinner (and Emberson for that matter) played that breakaway great and were victims of bad luck... Emberson's stick check changed trajectory of the puck. Skinner extending his leg to cover low-right opens his five hole, and unfortunately that's where Emberson's stick check directed the puck. It is physically impossible for Skinner to cover both at the same time.

In the shootout, you might want to take a 2nd look at Larkin's 7-hole goal, but then again Skinner's saves on the Draisaitl PK is the only reason we saw OT in the first place.

Never mind that our shootout attempts were utterly pathetic. Nuge didn't even get a shot off and McDavid played himself into a spot where the puck could only go into one place.

That said, the shootout isn't really indicative of anything.

 

[Perfect_Drug]

There is a scorekeeper judging and recording each shot already. It’s already happening.

Toggling one more attribute would not be that difficult.

Yes there would be some bias subjectivity, but I’d rather have human judgement than no information at all and then conjour up (and blindly believe) info created by an algorithm.

Once we have real time puck tracking and teach AI how to discern a high difficulty save from a low difficulty save I may change my tune, but as it stands, having been explained the limitations of the current public models they are frankly useless. They can’t differentiate the most fundamental variable in save difficulty (did the goalie need to move quickly to get there or not).

If there was any money at all to be made from this (there isn't) I could do it.

In my line of work, I've created a camera rig that re-creates everything it sees into a cloud-point. We could do this to re-create everything on the ice in real-time.

Techno-babble aside:

Creating a system where the entire surface of the ice is re-created in a computer is very simple.
Where we can mathematically see all the "binary" calls (icing, puck in the net, offsides, glove covering puck in crease, puck touched above the crossbar).

Then we can teach it what penalties are (Cross Check, Boarding, High stick)

The accuracy is based on data (number of camera angles you can obtain) and computing power.
We can make the reffing as Orwellian as you want.
We can create or parse even higher-level statistical models from the data we got from this.


The more cameras you have, the more accurate everything becomes.
A setup like this is for scanning skin pores, and dielectric properties of surface scattering (Hockey wouldn't need THIS many cameras, unless we wanted to count the eyelashes on the players).

It's really f***ing easy.


The NHL just doesn't want to do something like this because they want to "manage" the games.

 

[Fixed to Ruin]

As I'm sure others have mentioned. Apparently some sickness going through the team right now. My wife and I are both super sick right now, and I can assure you I couldn't play hockey right now, lol. Take the point, hope people get better and carry on.

Hope you and your wife feel better soon!

 

[Tobias Kahun]

If there was any money at all to be made from this (there isn't) I could do it.

In my line of work, I've created a camera rig that re-creates everything it sees into a cloud-point. We could do this to re-create everything on the ice in real-time.

Techno-babble aside:

Creating a system where the entire surface of the ice is re-created in a computer is very simple.
Where we can mathematically see all the "binary" calls (icing, puck in the net, offsides, glove covering puck in crease, puck touched above the crossbar).

Then we can teach it what penalties are (Cross Check, Boarding, High stick)

The accuracy is based on data (number of camera angles you can obtain) and computing power.
We can make the reffing as Orwellian as you want.

The more cameras you have, the more accurate everything becomes.
A setup like this is for scanning skin pores, and dielectric properties of surface scattering (Hockey wouldn't need THIS many cameras, unless we wanted to count the eyelashes on the players).

View attachment 970217

It's really f***ing easy.


The NHL just doesn't want to do something like this because they want to "manage" the games.

To do this you’d almost have to do this with your own money and prove it works before the dinosaurs at the NHL saw the value.

 

[Perfect_Drug]

To do this you’d almost have to do this with your own money and prove it works before the dinosaurs at the NHL saw the value.

Haha yeah. A lot of people would love for me to work for free for them.


My old company I used to work for in LA spent $40million on their camera rig.

They re-created Amsterdam for their Call of Duty game using their 3D Scanner:


The difference is my old employer (Infinity Ward) pays incredibly well. The game they make generates 4-5x more revenue than the entire NHL.

Why the hell would anyone in this field want to work for free?

 

[foshizzle]

On every PP he looked like he was shot out of a cannon coming out of zone. Man, I think we're spoiled. I'll take a "McD disinterested tonight" game over any of the OOT superstars in the league who aside from Mackinnon on a lot of nights you barely notice.

This was a substellar McD game. But the post and the takes here are engaging in making it look worse than it was. Like I said I think he has the flu. That can certainly impact things. The effort was there as it nearly always is with McD.

In anycase imagine saying to McD what you posted. I'm not calling you out. But his passion for the game is unmatched.

To me nobody on Earth cares about hockey games more than Drai, McD, Mackinnon. They're the only ones consistently on that list.

I agree with you- that’s why I said I thought I’d never say that. I also said “he looked”- but I am far from knowing what he feels. I know he is the ultimate competitor and “looks” has zero meaning. I mean, people said Paul Coffey “looked” lazy, but he was one the greatest d men of all time.

Connor for sure gets the benefit of the doubt. He dragged this team from the mud. My first assumption was he has whatever has been going around on the team. 100% agree with you on his competitive nature.

 

[Perfect_Drug]

I agree with you- that’s why I said I thought I’d never say that. I also said “he looked”- but I am far from knowing what he feels. I know he is the ultimate competitor and “looks” has zero meaning. I mean, people said Paul Coffey “looked” lazy, but he was one the greatest d men of all time.

Connor for sure gets the benefit of the doubt. He dragged this team from the mud. My first assumption was he has whatever has been going around on the team. 100% agree with you on his competitive nature.

I'm sure the team has whatever flu I had this past week.

 

[iCanada]

You maybe don’t understand statistics as well as you think you do.

If it only takes one critical error to lead to a goal against and we assume that all humans are error prone…

Well then it stands to reason that the goalie would be committing 1/6th of the errors and the skaters were committing 5/6ths of the errors.

So the odds that you’d be right when picking the goalie would not be in your favor.

And then… if you take it a step further, you’d have to consider the fact that the goalie doesn’t even get a chance to make an error unless there was a shot on goal, and so maybe the shot on goal itself was caused by an error.

Boy… this is getting harder, now I have to parse whether it the skaters error or the goalies error that was the bigger issue. I suppose it depends, right?

Well I dunno, that’s too complicated to suss out… we need something easier.

Well I have noticed that every single time the goalie makes an error a red light goes on… I mean that’s pretty obvious right?

Oh, but you’re saying that maybe the error came earlier in the play? But that just takes us back to having to suss out relative blame and I just hate doing that… I liked the red light idea better.

Plus is just get so mad when we get scored on you know? It’s like garrrrrr!!!! And then that red light comes on and I’m like “I need to kill someone”!!!! You ever get that feeling?

Plus… maybe it’s just me, but have you noticed that the goalie is on for every single goal against? It’s pretty uncanny! Ooohhh… y’know im starting to hate that goalie guy, especially the one with the flippy hair… see him letting in goals all the time.

I guess you’re right after all… way more mistakes. Burn him! He’s a witch!!!

That's only the case if you consider a hockey team to be error prone the same way shitty Christmas lights wired in series are - where any fault or failure leads to everything breaking. Which to be frank, is clearly non-sense.

In reality - all errors are not created equally. Just by virtue of the game and the system some players can make errors in certain areas of the ice every time and the other team can never even have a chance of scoring - there is built in redundancy. I don't think it's exactly controversial to say that as much as you want to win the board battle behind the other teams net - realistically you could lose it every time and if the other 4 guys are back the other team might not even make it past the red line.

A more realistic reliability model would look like three lines wired in parallel to a single line that is your goaltender - one with two guys (ie forward two forecheckers; best 2 of 8), one with one guy (F3, best 1 of 4), and one with two guys (ie your D). Alternatively - your three forwards weird in parallel wired to your D in series which is then wired to your goaltender in series.

There's obvious problems with both of those models, but they're both far more accurate than a 6 guy series Daisy chain.

If you were to run reliability calcs based on statistically simulated exponential decay or even relatively simple back of the napkin probability calcs - your system would clearly show that the goaltender is the weak link in the reliability chain and it's not even close.

 

[Fourier]

Your a math guy... a stats nerd told me the general accuracy of the models are limited to about 85%. That's a lot of noise among that signal.

If I were a GM, that might be useful enough to spot a trend in my team game... but it wouldnt' be enough for me to overrule my coaches eye test or send my pro-scouts home when it comes to making decisions about individual players.

I think you and I may disagree here on the value of that 85% accuracy claim. I was actually very impressed by that number if the stat is used properly.

There is a reason why teams are hiring companies like Sportlogiq and in general investing significant resources in analytics. They won't replace experienced eyes but they can complement and potentially significantly augment them. The big advantage though is that you can gain information about many more things than you could from traditional methods.

Hockey is a complex sport. Breaking down film manually is immensely resource intensive. (I've done it in football which is actually a simpler game to analyze. ) You can have someone sit down and watch film on your guy and spend many many hours to come up with an opinion. But how do you do that efficiently for 1000+ players across the league. You want to not only be able to understand what your guy did but also how he compares in a plethora of similar situations to others around the league. That information is important in everything from personnel decisions through salary negotiations.

I think the other thing that needs to be addressed is what one wants to use data for. Something like xG is not really a stat designed to say anything about individual events. It is a macro stat. It seems to me that most of your focus is on analyzing individual events. The reality is that over many many events the uniqueness of an individual circumstance tends to disappear. There is a big difference between looking at one player in one game and that player over a season or multiple years and even more so when you look at units or teams in the aggregate.

 

[OilerTyler]

Leaked in 2 goals? The Wings best forward had a breakaway because Kulak was in lala land, how was that a "leaked in" goal? The first one was a deflection, bad break. I swear, some of you just invent reasons out of thin air to hate the guy, it's weird. Look at the 2nd goal Lyon gave up, now THAT was a leaky goal.
There's plenty of reasons this season to criticize Skinner, last night was not one of them.

I wonder how many of the 20 goals Skinner let in this month were "leaky" or "bad" according to some people on here. If I weren't so lazy I'd dig through the post history of some of his biggest detractors and find out. I'd guess they complained about at least 15, if not all 20. Seems like some people on here think he should have a 0.5 GAA and .980 save percentage.

The dude is putting up elite numbers as of late and the "Skinner is garbage" noise just keeps getting louder.

 

[KeithIsActuallyBad]

I wonder how many of the 20 goals Skinner let in this month were "leaky" or "bad" according to some people on here. If I weren't so lazy I'd dig through the post history of some of his biggest detractors and find out. I'd guess they complained about at least 15, if not all 20. Seems like some people on here think he should have a 0.5 GAA and .980 save percentage.

View attachment 970230

The dude is putting up elite numbers as of late and the "Skinner is garbage" noise just keeps getting louder.

Skinner is supposed to stop:

Breakaways
2 on 1s
Deflections
Screened shots
Cross ice passes
Rebounds in tight

Etc etc etc. Literally every goal against is his fault and his fault alone in some way.

I'd start asking myself why the Oilers have this uncanny ability to not score on average goalies. If you only allow 2 goals or less you'll win more often than not. If Skinner gives up 6 and we score 5 (see: The Panthers game) by all means roast him. But the very idea that Skinner needs to stop every single shot for us to win is unfathomable to me.

 

[OilerTyler]

Skinner is supposed to stop:

Breakaways
2 on 1s
Deflections
Screened shots
Cross ice passes
Rebounds in tight

Etc etc etc. Literally every goal against is his fault and his fault alone in some way.

I'd start asking myself why the Oilers have this uncanny ability to not score on average goalies. If you only allow 2 goals or less you'll win more often than not. If Skinner gives up 6 and we score 5 (see: The Panthers game) by all means roast him. But the very idea that Skinner needs to stop every single shot for us to win is unfathomable to me.

You can tell who the people are that don't watch any non-Oilers hockey games. They think that Skinner is the only goalie in the league letting in goals.

You'd probably blow their minds if you told them Bobrovsky and Shesterkin have more games with a save percentage in the .700s (or less) than Skinner does this year.

 

[Pip]

Overall for this game I’ll give them a pass considering the reports of the flu going through the room and we are in the dog days of the season against a team that was pretty hot. Didn’t like the lack of killer instinct after going up 2. Hard to evaluate Klingberg but he made some decent passes. Getting those outlet passes to stretch the other team is huge for the style they like to play.

We are probably close to a breaking point with having both Podkolzin and Arvidsson as Drai’s wingers. At some point you need more finishing playing with the best player in the league this season. At times this season it has felt like they were about to get going but it hasn’t materialized. If they don’t find a solution here that creates some consistent scoring on Drai’s line they’re going to be forced to go McDrai in the playoffs.

Not sure if it’s a popular opinion or not but of the two I’d probably keep Podkolzin on there. Much better forechecking presence, makes less mistakes with the puck and doesn’t try to do too much. I’d be open to a different look at both positions though.

 

[Tobias Kahun]

You can tell who the people are that don't watch any non-Oilers hockey games. They think that Skinner is the only goalie in the league letting in goals.

You'd probably blow their minds if you told them Bobrovsky and Shesterkin have more games with a save percentage in the .700s (or less) than Skinner does this year.

The same posters who defend every goal against skinner also don’t watch games because they think the oilers are the only team with defensive breakdowns at times.

 

[brentashton]

Agree with all of that, but I thought he seemed a little defferential on coverage... it seemed to impact the first goal against where he didn't engage the goalscorer and tie up his stick, despite being in perfect position to do so.

Ya I recall that play (remember, I’m also a big DRW fan, so I might be clouded in my opinions of this entire game ). It was guaranteed-win night in my house!!

His entire career, his offensive play has been his calling card and his defensive play has been “average”. I’m thinking Paul “the defence whisperer” Coffey will need some time with him. There still seems to be a player in there.

 

[Tobias Kahun]

I sort of guessed it

He was invisible, but with some bad defense. His skating and passing were very good. It was hard to see but some pundits made note of it. Soooo... imo he will capture his Dallas form. He has rust to shake and also has to learn and execute Knoblauch's system.

I think he’ll be a clear upgrade on stecher, I hope they keep looking for a solidified top 4 guy though.

 

[russ99]

I think if some of our fans just admit that because the Oiler skaters focus on denying entry and zone exits in lieu of a Dallas Stars/Minnesota Wild-like own zone without the puck prevention of quality scoring chances, this is putting undue pressure on our goalies with open looks on goal, and they're doing reasonably well, in spite of this.

And conversely, that any goalie the Oilers would acquire with limited capspace and lack of premium trade chips would do the same in the same situation.

There's no magic bullet, "Fuhr at the back" that will come in and save every chance listed above:

Skinner is supposed to stop:

Breakaways
2 on 1s
Deflections
Screened shots
Cross ice passes
Rebounds in tight

So maybe lets not pile on the goalie for every goal allowed that he's not Superman or in his prime Patrick Roy.

 

[tardigrade81]

Never mind that our shootout attempts were utterly pathetic. Nuge didn't even get a shot off and McDavid played himself into a spot where the puck could only go into one place.

That said, the shootout isn't really indicative of anything.

This. Shootout means nothing. Playoffs don't have them so it's meaningless

We are still first place and we played very well. Detroit has a solid team. I said all day yesterday I was worried about that game. We should have won, but thems the breaks. We didn't even let our foot off the gas either. We were pressing hard in the third. It just so happens that Detroit was too. It happens. Got a point at least

 

[KeithIsActuallyBad]

You can tell who the people are that don't watch any non-Oilers hockey games. They think that Skinner is the only goalie in the league letting in goals.

You'd probably blow their minds if you told them Bobrovsky and Shesterkin have more games with a save percentage in the .700s (or less) than Skinner does this year.

Skinner's not perfect. I would argue he's not even that great. He's serviceable. Sometimes it would be really nice if this team flexed its offensive muscle. I'm tired of the "We got goalie'd" excuse. How many 2 on 1s did we eff up in that OT yesterday? At least 3? The success rate for the Oilers on those is utterly abysmal because every single team knows we're going to force a cross ice pass that has little chance of scoring.

 

[FiveFourteenSixOne]

If there was any money at all to be made from this (there isn't) I could do it.

In my line of work, I've created a camera rig that re-creates everything it sees into a cloud-point. We could do this to re-create everything on the ice in real-time.

Techno-babble aside:

Creating a system where the entire surface of the ice is re-created in a computer is very simple.
Where we can mathematically see all the "binary" calls (icing, puck in the net, offsides, glove covering puck in crease, puck touched above the crossbar).

Then we can teach it what penalties are (Cross Check, Boarding, High stick)

The accuracy is based on data (number of camera angles you can obtain) and computing power.
We can make the reffing as Orwellian as you want.
We can create or parse even higher-level statistical models from the data we got from this.


The more cameras you have, the more accurate everything becomes.
A setup like this is for scanning skin pores, and dielectric properties of surface scattering (Hockey wouldn't need THIS many cameras, unless we wanted to count the eyelashes on the players).

View attachment 970217

It's really f***ing easy.


The NHL just doesn't want to do something like this because they want to "manage" the games.

 

[OilerTyler]

Skinner's not perfect. I would argue he's not even that great. He's serviceable. Sometimes it would be really nice if this team flexed its offensive muscle. I'm tired of the "We got goalie'd" excuse. How many 2 on 1s did we eff up in that OT yesterday? At least 3? The success rate for the Oilers on those is utterly abysmal because every single team knows we're going to force a cross ice pass that has little chance of scoring.

I've maintained all year that Skinner is an average goalie. With the way our pay structure is set up average is about as good as we can afford. The thing is, we're supposed to have the best offence in the league (we definitely pay them to be) so average goaltending should be enough to win us games on most nights. What annoys me is the nights when Skinner has an objectively good or great game and people still want to blame the loss on him because he didn't post a .970. That's not how this team is built to win games.

 

[Fourier]

Skinner's not perfect. I would argue he's not even that great. He's serviceable. Sometimes it would be really nice if this team flexed its offensive muscle. I'm tired of the "We got goalie'd" excuse. How many 2 on 1s did we eff up in that OT yesterday? At least 3? The success rate for the Oilers on those is utterly abysmal because every single team knows we're going to force a cross ice pass that has little chance of scoring.

With Skinner my question would always be...did he give the team a reasonable chance to win. I think it would be hard to say that last night he did not do so. The key question remains, for a team with aspirations to win it all are there still too many nights where the answer is no.

 

[KeithIsActuallyBad]

With Skinner my question would always be...did he give the team a reasonable chance to win. I think it would be hard to say that last night he did not do so. The key question remains, for a team with aspirations to win it all are there still too many nights where the answer is no.

.941 sv. Two goals against. The Oilers had every opportunity to put the game away and they didn't. Are we really touting Alex Lyon as some sort of new Patrick Roy?

I've maintained all year that Skinner is an average goalie. With the way our pay structure is set up average is about as good as we can afford. The thing is, we're supposed to have the best offence in the league (we definitely pay them to be) so average goaltending should be enough to win us games on most nights. What annoys me is the nights when Skinner has an objectively good or great game and people still want to blame the loss on him because he didn't post a .970. That's not how this team is built to win games.

I think back to that Vegas game where we lost 1-0 and very few people pointed out the fact that we didn't score as if it's somehow possible to win when you put zero pucks in the net.

Conversely when we beat LA 1-0 the narrative was "Well it's the Kings so Skinner had little to do with the win."