Advanced Stats- Expected Goal %

[PerreaultForever]

On 10/20/2024 at 4:44 PM, mjd1001 said:

On the other hand, Quinn, Benson, and Cozens are near the bottom of that list, making it look even worse for them.

Didn't take any advanced stats to know that line has been shite. 

[JoeSchmoe]

11 hours ago, PerreaultForever said:

Didn't take any advanced stats to know that line has been shite. 

True, but they're getting better.

With the eye test, it's easy to see who's really good and who's really bad.

What the eye test isn't as good at is seeing which players consistently make the seemingly innocuous plays that lead to scoring chances either for or against. That said, as the sample size increases.... actual GF% is much more important.

[LGR4GM]

 

 

[LGR4GM]

Team sucks

 

 

[JoeSchmoe]

29 minutes ago, LGR4GM said:

Team sucks

 

 

We need scoring... Who'd have thought?!?!?!

On that note, can we end the Dennis Gilbert experiment now?!??

[JoeSchmoe]

Here's an update after the west coast swing.

Maybe it's a function of playing with Dahlin, but Byram continues to rise. Bryson is falling fast.

Quinn and Cozens continue to get back to team mean. 

Our checking line is sinking. Is this matchup based or is this them not playing well. Lafferty being the worst of the bunch stats wise matches the eye test. As an aside, I haven't seen any of the blazing speed we were promised in the offseason. 

[Big Guava]

7 minutes ago, JoeSchmoe said:

Here's an update after the west coast swing.

Maybe it's a function of playing with Dahlin, but Byram continues to rise. Bryson is falling fast.

Quinn and Cozens continue to get back to team mean. 

Our checking line is sinking. Is this matchup based or is this them not playing well. Lafferty being the worst of the bunch stats wise matches the eye test. As an aside, I haven't seen any of the blazing speed we were promised in the offseason. 

The speed doesn't show up in creating chances on offense usually, it shows up in getting back on D and getting sticks on pucks in the D zone which they have been really good at as of late.

And on the penalty kills where they have been all over teams lately. Sabres have risen to 12th in the NHL in penalty killing from 30th earlier in the year.

This team is starting to come together and learn under Ruff. Winning isn't automatic like some think, it's a process and having to go thru learning periods and figuring things out when they are in certain situations.

Edited November 24 by Big Guava

[Pimlach]

How many teams have 4th lines that are NOT the bottom line in “expected goals”?  
 

I think Malenstyn is doing what was expected. Lafferty too.  
 

My eye test says that Abe-Kubel has been the least productive of the 4th liners.   Ruff has played him the least games by far (only 7 games).  He has no goals, no points, and only 4 SOGs.   Due to injuries, he did recently did get shifts on the higher lines and now all of a sudden he has higher ExGF, he actually is rated with the better scorers on the team.   That seems flawed.   

Fancy stats are nice but you cannot use them stand alone.   His “real” production is not remotely close to “expected”.   If he played more games this would likely change.  

 

 

Edited November 24 by Pimlach

[Big Guava]

1 minute ago, Pimlach said:

How many teams have 4th lines that are NOT the bottom line in “expected goals”?  

 

Yeah it's a really weird take, IMO.

 

Why would anyone expect their 4th line to be good at scoring? If they were they wouldn't be on the 4th line, it's not their job.

Edited November 24 by Big Guava

[JoeSchmoe]

14 minutes ago, Pimlach said:

How many teams have 4th lines that are NOT the bottom line in “expected goals”?  

 

It's expected goal %... Not expected goals. If they're playing against opposing 4th lines they have just as good an opportunity to be high as anyone. If they're out there shutting down the opposition 1st lines, then what you're saying is right, and they'll be low. That's why I asked about the matchups they're getting.

[Pimlach]

13 minutes ago, JoeSchmoe said:

It's expected goal %... Not expected goals. If they're playing against opposing 4th lines they have just as good an opportunity to be high as anyone. If they're out there shutting down the opposition 1st lines, then what you're saying is right, and they'll be low. That's why I asked about the matchups they're getting.

Isn’t expected goals percentage still a function of expected goals ?   Seems like it would be to me.  

Ruff usually plays the 4th line against scoring lines and they get defensive zone starts when he can get the match up.  Sometimes they line up against the other 4th line.   Strategy changes as the game evolves and depending on who gets the last change.  

Now, if the opponents 4th line is a defensive shut down line, and our fourth line is playing against them, would that make it even harder for our 4th line to score?   Seems like it would.  
 

If I’m playing against the Sabres, and I want to score a goal, then I want to play against Cozens and Quinn’s line, especially if Bryson is out there with them.  

[JoeSchmoe]

26 minutes ago, Pimlach said:

Isn’t expected goals percentage still a function of expected goals ?   Seems like it would be to me.  

It's a function of expected goals for vs expected goals against. 

If as you say, he puts them out there against the top lines, it makes sense that they're low. I've never noticed this to be the case, but I haven't really paid attention to it. That's why I was asking. 

[Big Guava]

2 hours ago, JoeSchmoe said:

It's expected goal %... Not expected goals. If they're playing against opposing 4th lines they have just as good an opportunity to be high as anyone. If they're out there shutting down the opposition 1st lines, then what you're saying is right, and they'll be low. That's why I asked about the matchups they're getting.

Most 4th lines aren't there to play against other teams 4th lines they are there to play against the other teams top lines since they are more defensive players usually with limited offensive potential.

Essentially they have a role to play on the team and offense isn't it. Anything you get from them in that way is a bonus.  Even then just spending 45-60 seconds in the offensive zone on a hard fore check without even getting a shot on goal is a win since it means the other teams top line isn't even getting in the O zone before they have to change.

Edited November 25 by Big Guava

[RochesterExpat]

2 hours ago, JoeSchmoe said:

It's expected goal %... Not expected goals. If they're playing against opposing 4th lines they have just as good an opportunity to be high as anyone. If they're out there shutting down the opposition 1st lines, then what you're saying is right, and they'll be low. That's why I asked about the matchups they're getting.

Krebs, Lafferty, NAK and Malenstyn have by far the fewest offensive zone starts on the team and it’s not even close. They’re split between defensive zone and changing on the fly. There is a clear top 6 and bottom 6 differentiation.

Weird other thing I noticed: among forwards, Malenstyn is second only to Tuch in 5v5 shots blocked, but at shots blocked per 60 he’s at 6.16 to Tuch’s 4.35. Third on the team per 60? Kulich at 2.89. The only other Sabre above 2 is Greenway at 2.11.

[ska-T Palmtown]

13 hours ago, JoeSchmoe said:

It's a function of expected goals for vs expected goals against. 

If as you say, he puts them out there against the top lines, it makes sense that they're low. I've never noticed this to be the case, but I haven't really paid attention to it. That's why I was asking. 

This missing piece is the actual goals. The Tage number tells us that when he is on the ice, he (his line) should get about 61% of the goals scored. When he is not on the ice, the rest of the team is only at 49%. But that is purely a function of where the shots are taken and whether the "average" NHL player would score on the shot.

If you have a high xGF% but a low actual GF - you (and your linemates) are either terrible shooters or are having bad puck luck. Conversely, if you have a low xGF%, but high actual GF - your goalie bails you out a lot and you bury the few chances you get.

One without the other does not paint a clear picture of actual performance. A line stacked with snipers (or D-men he get those seeing eye pucks through) should actually have a better GF% than their xGF, since xGF is based on the "average".

Based on that chart, if Gilbert's actual goals for % is higher than the expected, once could argue he is outperforming expectations, as low as those may be. Statistics without context can be made to say just about anything you want. There are very few players who single-handedly carry or sink their lines consistently.

[JoeSchmoe]

24 minutes ago, ska-T Chitown said:

This missing piece is the actual goals. The Tage number tells us that when he is on the ice, he (his line) should get about 61% of the goals scored. When he is not on the ice, the rest of the team is only at 49%. But that is purely a function of where the shots are taken and whether the "average" NHL player would score on the shot.

If you have a high xGF% but a low actual GF - you (and your linemates) are either terrible shooters or are having bad puck luck. Conversely, if you have a low xGF%, but high actual GF - your goalie bails you out a lot and you bury the few chances you get.

One without the other does not paint a clear picture of actual performance. A line stacked with snipers (or D-men he get those seeing eye pucks through) should actually have a better GF% than their xGF, since xGF is based on the "average".

Based on that chart, if Gilbert's actual goals for % is higher than the expected, once could argue he is outperforming expectations, as low as those may be. Statistics without context can be made to say just about anything you want. There are very few players who single-handedly carry or sink their lines consistently.

I've said most of what you said in this thread. It's the art vs science. Also, as more data becomes available, the flukiness of shots going in vs being saved tends to work itself out and as you said, the actual GF% is much more telling.

Gilbert for his part though has 1 shot attempt and 0 SOG in 6 games. We can do better.

[Broken Ankles]

11 hours ago, RochesterExpat said:

Krebs, Lafferty, NAK and Malenstyn have by far the fewest offensive zone starts on the team and it’s not even close. They’re split between defensive zone and changing on the fly. There is a clear top 6 and bottom 6 differentiation.

Weird other thing I noticed: among forwards, Malenstyn is second only to Tuch in 5v5 shots blocked, but at shots blocked per 60 he’s at 6.16 to Tuch’s 4.35. Third on the team per 60? Kulich at 2.89. The only other Sabre above 2 is Greenway at 2.11.

The Beck/NAK/Lafferty minutes are real low. That is driving those two stats you referenced.  They are obviously a defense first unit, so low offensive starts are limited.  But when you look at % of starts in the defensive zone they are not the highest.  Which means Ruff has consistently put third line players (like Krebs, McCloud, Greenway) out there more frequently than the fourth line (I assume late I games) to take faceoffs.  
 Similarly, Tuch dwarfs Beck in TOI for both 5v5 and PK, so the stat of blocked shots per 60 minutes makes perfect sense.  
  What I find interesting is that if you look at Malenstyns 2023 vs 2024 TOI he is playing less than he did last year in Washington.  And while the ratios for scoring/60 are about even, his giveaways and takeaways are both going in the wrong direction.  In terms of forwards, Tuch, McCloud, Greenway have more Pk time per game. Maybe a surprise to no one.  
  Maybe it’s too early to root cause but why does Ruff utilize the fourth line less than expected?  Specifically Malanstyn, who to your point is blocking shot at a high rate when on the ice? And winning faceoffs at a high rate.  I was able to watch him in Anaheim and he had one significant hit in the first period. After that the game became more physical and I would have expected more from Beck and got less. And he had the lowest 5v5 time in that game at 10 minutes.  Not saying the experiment is a failure but lots of hype around the fourth line this summer not living up to expectations early in the season. 
 

[LTS]

Without digging into it I would surmise that the 4th line is getting less than expected time (expected than they did last year or than a player did on another team last year?) because the Sabres find themselves behind in games.  When a team needs goals they generally shorten the bench and that puts the 4th line off the ice. In fact, Krebs often gets moved up and one of the 3rd liners gets dropped.

Too busy this morning to put more into it, but that's my hypothesis.

[JoeSchmoe]

Just going eye test wise, I've liked Krebs and Malenstyn well enough, but I haven't liked Lafferty.

I was really hoping putting some guys together with elite level speed would create scoring chances, but unfortunately this hasn't been the case.

Going back to the Bailey/Baptiste days, it never ceases to amaze me how guys with speed but no skill never seem to have the willingness to get the puck into dangerous spots. Girgensons too. They skate it into the opposing team's corner, keep it around the perimeter, lose the puck, and repeat next shift.

You can't score if you don't try.

Edited November 25 by JoeSchmoe

[Pimlach]

54 minutes ago, JoeSchmoe said:

I've said most of what you said in this thread. It's the art vs science. Also, as more data becomes available, the flukiness of shots going in vs being saved tends to work itself out and as you said, the actual GF% is much more telling.

Gilbert for his part though has 1 shot attempt and 0 SOG in 6 games. We can do better.

 I don’t think anyone associated with the Sabres needs this data to know what Gilbert’s role is and what his limitations are.     Is anyone surprised?   

Gilbert played in 7 games, and he averages ~11 minutes TOI.   The team was 5-2 in those games.  He got his games at a critical point too, when the team started to turn things around, and he did fine.  He is here for a physical presence, he is a low cost 8th D option, and he is not going to score many goals.  

Can we do better?   Of course.  Ryan Johnson is in Rochester.  So is Novikov.  Let them continue to develop in the AHL and let Gilbert be the 8D.  I like it set up the way it is and I have no problem with Gilbert at 8D at this point in the season.