Eric Johnson- Advanced Stats Nightmare

[Brawndo]

21 minutes ago, LGR4GM said:

The Leinonen pick sticks in my craw.

https://www.eliteprospects.com/player/525669/luca-del-bel-belluz

[Captain Caveman]

4 hours ago, inkman said:

Clifton signed for $10 million guaranteed over the next 3 seasons.  He’s not get waived.  

Is anyone claiming him?  If not he can play in Rocheseter...

[PerreaultForever]

Our D sucks, no question. As does the goaltending on many occasions. 

For years now people have argued about it, and which D is the worst and who has to go and some of them do go (and often play just fine on other teams) but just once, I'd like to see the discussion of what the forwards do NOT do which is the biggest factor in all of this. 

Every single D man who comes here plays worse than they did on their previous team. Advance stat that. 

[Pimlach]

12 minutes ago, PerreaultForever said:

Our D sucks, no question. As does the goaltending on many occasions. 

For years now people have argued about it, and which D is the worst and who has to go and some of them do go (and often play just fine on other teams) but just once, I'd like to see the discussion of what the forwards do NOT do which is the biggest factor in all of this. 

Every single D man who comes here plays worse than they did on their previous team. Advance stat that. 

Yes, we have had some defensemen leave here and play quite well. Scandella for one.   Bogo, I guess.  Montour was just entering prime, he has had a huge post Sabre improvement.   

[PerreaultForever]

10 minutes ago, Pimlach said:

Yes, we have had some defensemen leave here and play quite well. Scandella for one.   Bogo, I guess.  Montour was just entering prime, he has had a huge post Sabre improvement.   

Montour is a perfect case in point. You see Montour's a wanderer. But he is aggressive and attacks the goal. Adds a lot to their offense. The difference though is they have a system and a forward covers for him when he takes off. Our forwards, very often, don't do squat.

[TheAud]

1 hour ago, JoeSchmoe said:

Interesting. I would have thought there would be a bigger spread. Jokiharju surprises me. Based on Granato's recent decisions it seems like he has him 7th, but based on at least these two data points, he should be ahead of Clifton (which to me is obvious) and E. Johnson (less obvious to my eyeballs) and maybe even Samuelsson (who has sucked this year but demonstrated better in the past). 

[Thorny]

15 hours ago, Pimlach said:

Please.  He is not the problem.  
 

This thing was broke long before him.  

No idea why the OP is getting torched. He’s been statistically bad

He’s not THE problem (why do people always want to pigeonhole the “problem” onto one thing? It never works like that), but he’s a symptom: a symptom of the type of aggressiveness the GM shows in addressing the roster

if you just assemble a bunch of guys that “aren’t the problem” you still haven’t rostered the guys you needed who could be “part of the fix” 

[inkman]

1 hour ago, Captain Caveman said:

Is anyone claiming him?  If not he can play in Rocheseter...

I know that’s what you want and he probably deserves buts it’s not happening.  Kevyn isn’t going to say “whoops, I guess I screwed up my big UFA signing in the off-season, off to Rochester Connor.”

[Captain Caveman]

5 minutes ago, inkman said:

I know that’s what you want and he probably deserves buts it’s not happening.  Kevyn isn’t going to say “whoops, I guess I screwed up my big UFA signing in the off-season, off to Rochester Connor.”

I like to think he's smart enough to know that only winning is going to save his job, and keeping Connor up isn't helping him there.

[inkman]

21 minutes ago, Captain Caveman said:

I like to think he's smart enough to know that only winning is going to save his job, and keeping Connor up isn't helping him there.

All he can do at this point is push him down the depth chart.  Go get someone better.  

[Captain Caveman]

It's clear to me that we have 6 or 7 guys who are better (if everyone's healthy)

 

[GASabresIUFAN]

“Fixing“ the defense is pretty simple.  Go down to 6 D and play Ryan Johnson. Clifton has been our worst D.  Put him in the pressbox.  This isn’t hard.  
 

 

[sabresparaavida]

4 hours ago, TheAud said:

What is "Relative Expected Goals %"?  Is it telling me that Dahlin is the Sabres 5th best defenseman? Because if it is, it's BS. 

Expected goal percentage is the ratio between the expected goals for and against while a player is on the ice. Team relative expected goal percentage is standardizing that based on the team average. 
 

expected goals for/against are calculated by taking past data on how likely shots of different types (slap shot, clean breakaway, wrister, etc.) from different areas of the ice are to be goals. So if there’s a 20% chance of a clean slot shot going in, a slot shot would give an expected goal of 0.2. Adding  up the expected goals of each shot over the course of a game would give your expected goals. 

[JoeSchmoe]

1 hour ago, sabresparaavida said:

Expected goal percentage is the ratio between the expected goals for and against while a player is on the ice. Team relative expected goal percentage is standardizing that based on the team average. 
 

expected goals for/against are calculated by taking past data on how likely shots of different types (slap shot, clean breakaway, wrister, etc.) from different areas of the ice are to be goals. So if there’s a 20% chance of a clean slot shot going in, a slot shot would give an expected goal of 0.2. Adding  up the expected goals of each shot over the course of a game would give your expected goals. 

Nice explanation. Though I'm not as far down the rabbit hole as some others here, I find that the stat doesn't necessarily account for how elite scoring forwards like Thompson and Matthews consistently put pucks in nets above what's expected. Someone also mentioned it doesn't account for zone starts.

I can't think of any stat though that would justify Johnson's low numbers. We're just very likely to be outscored by the opposition when he's likely on the ice. 

[PerreaultForever]

1 hour ago, JoeSchmoe said:

Nice explanation. Though I'm not as far down the rabbit hole as some others here, I find that the stat doesn't necessarily account for how elite scoring forwards like Thompson and Matthews consistently put pucks in nets above what's expected. Someone also mentioned it doesn't account for zone starts.

I can't think of any stat though that would justify Johnson's low numbers. We're just very likely to be outscored by the opposition when he's likely on the ice. 

Looking at so called advanced stats without considering what's actually going on in the game is stupid stuff that should be reserved for bored children and shut ins who spend too much of their lives on computers and nobody else.

Simple example. Face off in your own end. Team puts player on ice because they consider him a better defender. Team loses face off. Opposition controls and moves puck around perimeter and after some time bangs in a rebound. Said player now has bad metrics. On ice for goal against. Team did not possess puck while he was on. Danger chance not prevented. Blah blah blah. It's the team's fault, but when computer nerd looks at stats he or she goes oh man, what lousy advanced stats. This guy sucks. 

and all you had to do was win said face off and all those metrics change. 

(Now the computer nerds can proceed to tell me how I don't understand advanced metrics because I didn't give enough long winded explanations and illustrations and they will think they are smart because I'm lazy and annoyed and just pissed off with this stupid f'n hockey team. Have at it. idc)

[JoeSchmoe]

34 minutes ago, PerreaultForever said:

Looking at so called advanced stats without considering what's actually going on in the game is stupid stuff that should be reserved for bored children and shut ins who spend too much of their lives on computers and nobody else.

Simple example. Face off in your own end. Team puts player on ice because they consider him a better defender. Team loses face off. Opposition controls and moves puck around perimeter and after some time bangs in a rebound. Said player now has bad metrics. On ice for goal against. Team did not possess puck while he was on. Danger chance not prevented. Blah blah blah. It's the team's fault, but when computer nerd looks at stats he or she goes oh man, what lousy advanced stats. This guy sucks. 

and all you had to do was win said face off and all those metrics change. 

(Now the computer nerds can proceed to tell me how I don't understand advanced metrics because I didn't give enough long winded explanations and illustrations and they will think they are smart because I'm lazy and annoyed and just pissed off with this stupid f'n hockey team. Have at it. idc)

I'll try to keep it positive. 

There are scenarios where goals will go in or chances are generated that aren't a players fault. However, with a large enough sample size, those instances equal out. In the scenario you describe though, if its 5 on 5, a truly elite defender will break up the cycling around the perimeter, gain possession, and limit scoring more often than a weak defender (4 on 5, they'd be happy to keep it on the perimeter... no defender will challenge).

You're absolutely right that you have to watch the games, because there's a lot of reasons to take the advanced stats with a grain of salt. As I mentioned before, if a player has elite shooting talent, they will no doubt be better at winning games that someone with no shooting talent but a better expected goal %. 

Maybe someone with a better grasp on the stats than me can justify why EJ sits so low expected goals % wise. The best I can come up with is that he's been paired with Clifton a lot, who is also a trainwreck... But Clifton has actually been marginally better. Possible this is sample size. 

I still think whatever help his "experience" has brought, doesn't seem to outweigh him giving up as so many chances compared to what he creates... and I'll chalk this up as a major KA fail.

Edited November 27, 2023 by JoeSchmoe

[GASabresIUFAN]

I'm not sure why people think EJ is a problem.  He's only -2 and only has 6 giveaways in 21 games.  Power has 17 (with only 5 takeaways).  Dahlin has 15, Mule 11, Joki 10, and Clifton 9.  

EJ, who was often paired with Clifton, must be doing something right if his +/- is -2 compared to Clifton's -9.  He is second on the teams with blocked shots (37) and is 6th on the team with 26 hits.  He is also a stalwart on the PK. So far he is doing exactly what he was hired to do.  

Anyone who wants to start a thread on how disappointing Clifton has been, I'm right there with you, but EJ has passed the eye test regardless of what the often useless fancy stats say.

[PerreaultForever]

2 hours ago, JoeSchmoe said:

I'll try to keep it positive. 

There are scenarios where goals will go in or chances are generated that aren't a players fault. However, with a large enough sample size, those instances equal out. In the scenario you describe though, if its 5 on 5, a truly elite defender will break up the cycling around the perimeter, gain possession, and limit scoring more often than a weak defender (4 on 5, they'd be happy to keep it on the perimeter... no defender will challenge).

You're absolutely right that you have to watch the games, because there's a lot of reasons to take the advanced stats with a grain of salt. As I mentioned before, if a player has elite shooting talent, they will no doubt be better at winning games that someone with no shooting talent but a better expected goal %. 

Maybe someone with a better grasp on the stats than me can justify why EJ sits so low expected goals % wise. The best I can come up with is that he's been paired with Clifton a lot, who is also a trainwreck... But Clifton has actually been marginally better. Possible this is sample size. 

I still think whatever help his "experience" has brought, doesn't seem to outweigh him giving up as so many chances compared to what he creates... and I'll chalk this up as a major KA fail.

A good part of this I wouldn't argue with but your comment about an "elite defender" isn't really true. It's a team game and elite players can only do so much on their own. If they are surrounded by stiffs their stats will still suck. 

What also has to be considered is who is Granato putting out there against who? I don't have those stats but is EJ being put out in more situations where Granato feels he needs a better defender? Is he up against the top forwards? 

I don't think anybody would argue that EJ is no longer a top D man but he wasn't brought in to be a top D man. He was brought in to be a veteran mentor and bottom pairing guy, an upgrade on Stillman/Bryson. No more, no less. 

Bottom line, EJ is not the reason this team is losing or why they looked so pathetic against New Jersey. Not by a longshot. 

[Brawndo]

3 hours ago, PerreaultForever said:

Looking at so called advanced stats without considering what's actually going on in the game is stupid stuff that should be reserved for bored children and shut ins who spend too much of their lives on computers and nobody else.

Simple example. Face off in your own end. Team puts player on ice because they consider him a better defender. Team loses face off. Opposition controls and moves puck around perimeter and after some time bangs in a rebound. Said player now has bad metrics. On ice for goal against. Team did not possess puck while he was on. Danger chance not prevented. Blah blah blah. It's the team's fault, but when computer nerd looks at stats he or she goes oh man, what lousy advanced stats. This guy sucks. 

and all you had to do was win said face off and all those metrics change. 

(Now the computer nerds can proceed to tell me how I don't understand advanced metrics because I didn't give enough long winded explanations and illustrations and they will think they are smart because I'm lazy and annoyed and just pissed off with this stupid f'n hockey team. Have at it. idc)

The designers of the public models frequently posted here such as Natural Stat Trick, Moneypuck Hockey Viz and Evolving Hockey will tell you that their models require a minimum of ten games, preferably twenty before any real patterns or data can be drawn, so the example you outlined would not have a discernible effect unless it happened to the same player multiple times establishing a pattern of poor play. 
 

The “Computer Nerds” employed by the Sabres watch every game the team plays a minimum of 2-3 times, once live and the other times on replay to make sure their data lines up with what they are seeing. 
 

There is quite a bit of IDGAF towards fancy stats on this board which is every posters right. I personally like them as it gives another piece of information about players.
 

That being said most of the successful teams in the league utilize analytics in personnel decisions, deciding lineup configurations and deployment of players. I want My Team doing the same. 

 

[GASabresIUFAN]

My problem with most fancy stats is that they are so often team-dependent.  Good teams have good fancy stats and bad teams don't.  For example, take our friend Eichel.  In Buffalo, his advanced numbers were lousy, just like the team.  Put him on LV and his stats are now wonderful.  

What I find more intriguing when I look at hockey stats is finding positive and negative anomalies.  For example, when Ullmark played here, the Sabres played at a playoff level and when any other goalie was in the net, the Sabres were awful.  It illustrated how vital Ullmark was to the team.  Too bad management didn't see it that way.  

Hockey is pretty simple.  The more shots you get on the net, especially in HD areas, the more likely to win most nights.  Have a goalie who can consistently stop 90% of the shots he faces and you'll win most nights.  Help the goalie by having a defense that limits shots and limits shots from HD areas, the better your chances are.  

The Sabres are 29th in shots per game at 27.8.  The Sabres are 17th in shots allowed at 30.4. Our differential of -2.6 is 25th in the NHL.  (Stats from NHL.com)

The Sabres are 19th in HDCF at 178 and 24th in HDCA at 199.  Our HDCF% of 47.21% is 23rd in the NHL.  (Stats from Natural stat trick)

Goaltending stats - 5 on 5 Save % - Buffalo .904 (NHL average .918); overall save % - Buffalo .890 (NHL average .897) (Stats from Hockey-reference.com) 

I don't need too many advanced stats to show us that the Sabres don't shoot enough, do a mediocre job of getting shots off from high-danger areas, and do a lousy job of suppressing shots and high-danger shots.

Now add that to the Sabres below-average goaltending and terrible PP, it's amazing this team is even close to 500.  

The anomalies are Mitts and JJP who are contributing at both ends of the ice.  Ryan Johnson who has been a positive upgrade on defense.  UPL vs Comrie and Levi.  UPL has a .914 save % while Levi (.876) and Comrie (.877).  This is just like Ullmark vs anyone else during his time here.  On the bad are Krebs and Clifton (no fancy stats needed)

A real GM would be working hard to fix these obvious issues.  Levi would be sent down.  Clifton would be in the pressbox.  Ryan Johnson would be playing a regular shift.  Krebs would be sent down as well.    The GM would waive Comrie and find a veteran backup to play with UPL for the remainder of the season.  The GM would also acquire a center to replace Krebs and Jost and at least one winger to upgrade KO and or Z.  

Edited November 27, 2023 by GASabresIUFAN

[PerreaultForever]

1 hour ago, Brawndo said:

The designers of the public models frequently posted here such as Natural Stat Trick, Moneypuck Hockey Viz and Evolving Hockey will tell you that their models require a minimum of ten games, preferably twenty before any real patterns or data can be drawn, so the example you outlined would not have a discernible effect unless it happened to the same player multiple times establishing a pattern of poor play. 
 

The “Computer Nerds” employed by the Sabres watch every game the team plays a minimum of 2-3 times, once live and the other times on replay to make sure their data lines up with what they are seeing. 
 

There is quite a bit of IDGAF towards fancy stats on this board which is every posters right. I personally like them as it gives another piece of information about players.
 

That being said most of the successful teams in the league utilize analytics in personnel decisions, deciding lineup configurations and deployment of players. I want My Team doing the same. 

 

The Sabres are supposed to be relying heavily on analytics. How's that working out?

I have no issue with the use of advanced analytics. You use every piece of information you have and you look at everything available. This is just common sense. I only have an issue with using them in isolation from the eye test, experience, and character. 

If I'm the coach following that New Jersey debacle, I'm not looking over the advanced stats today, I've got the entire team doing wind sprints until some of them start to puke. 

Close the country club doors until these guys learn a proper work ethic. After that, we can start digging deeper into the nuances of the analytics. 

[Taro T]

1 hour ago, Brawndo said:

The designers of the public models frequently posted here such as Natural Stat Trick, Moneypuck Hockey Viz and Evolving Hockey will tell you that their models require a minimum of ten games, preferably twenty before any real patterns or data can be drawn, so the example you outlined would not have a discernible effect unless it happened to the same player multiple times establishing a pattern of poor play. 
 

The “Computer Nerds” employed by the Sabres watch every game the team plays a minimum of 2-3 times, once live and the other times on replay to make sure their data lines up with what they are seeing. 
 

There is quite a bit of IDGAF towards fancy stats on this board which is every posters right. I personally like them as it gives another piece of information about players.
 

That being said most of the successful teams in the league utilize analytics in personnel decisions, deciding lineup configurations and deployment of players. I want My Team doing the same. 

 

Fully expect, especially with the guys that have been hired in the last couple of years, that the Sabres analytics team is towards the top end.  Also fully believe that they look at stuff the "public" analytics crews don't.  

And, speaking as one of the "eye test" posters that realizes that analytics can be a hugely effective tool, (have every tool you can afford in your toolbox and understand how and when to use that tool) expect that a LOT of the public analytics data is ####.  Because they don't seem to differentiate  between a "high danger" chance by a Girgensons where the goalie didn't have to move to react to it and a "high danger" chance where the puck was moved there by Pasternak because the goalie already had to make a huge move to stop the potential shot from Pasternak. Personally, expect the team's analytics team does account for that.  (Which is a big part of why the guy everyone wants to see in Ra-cha-cha ain't there yet.)

[mjd1001]

21 hours ago, PerreaultForever said:

For years now people have argued about it, and which D is the worst and who has to go and some of them do go (and often play just fine on other teams) but just once, I'd like to see the discussion of what the forwards do NOT do which is the biggest factor in all of this. 

Every single D man who comes here plays worse than they did on their previous team. Advance stat that. 

I agree almost 100% but I'll take it by saying that the D actually isn't that bad Many games.

I posted this earlier in this thread, the forwards really hang the D out to dry.

After you see a few goals where a defenseman is out of position, Go to nhl.com and look at the replay of the goal. Oftentimes they will show it from various angles.... go back 10 to 15 seconds. Watch the play develop. Go frame by frame. You know what I see when I do that a lot?

Forwards are late getting back, Often times without much effort. Forwards are losing battles on the wall in their own end. Forwards chase the puck like a dog chases a squirrel and leave their position open, so now a defenseman has to worry about HIS guy AND whether or not he should step up and take the guy who is wide open because Cozens or Krebs or another young guy made an awful D-zone decision (and yes, Cozens does this a LOT...stop the talk about him being a good 2-way playe. He can be a disaster positionally in his own end... Don't confuse him skating hard with him knowing a single thing about defensive positioning) 

Oftentimes the D man gets caught between those two spots when the goal is scored and on the 5 second replay it looks like the defense didn't know what they were doing, when in actuality they were trying to cover for the awful forward play.  A snapshot of the goal has us think the D don't know where to be or make bad plays, when they are trying to do their job AND the forwards job at the same time.

There are many times the Sabres are defending a 3 on 3 but the defenseman have to play it as if it's a 3 on 2 because they just can't rely on the forwards to make a correct decision. When a goal is scored, it looks easy to place that on the D, But oftentimes it is the consistent lack of smart positioning or effort from the forwards that is more at fault, and we simply blame the defense as being bad, when it's not always true.

 

I want my defense to do 3 primary things:

1. Don't make mistakes in your own zone.

2. Get the puck out of your own zone.

3. Cover up for mistakes that the forwards make in your own zone.

The biggest weakness this defense has is they're not good enough to do #3, and often they look worse than they are when trying.

Edited November 27, 2023 by mjd1001

[SwampD]

2 hours ago, mjd1001 said:

I agree almost 100% but I'll take it by saying that the D actually isn't that bad Many games.

I posted this earlier in this thread but the forwards really hang the D out to dry.

After you see a few goals where defenseman is out of position, Go to nhl.com and look at the replay of the goal. Oftentimes they will show it from various angles and go back 10 to 15 seconds. Watch the play develop. Go frame by frame. You know what I see when I do that a lot?

Forwards are late getting back, Often times without much effort. Forwards are losing battles on the wall in their own end. Forwards chase the puck like a dog chases a squirrel and leave their position open, so now a defenseman has to worry about HIS guy AND whether or not he should step up and take the guy who is wide open because Cozens or Krebs or another young guy made an awful D-zone decision (and yes, Cozens does this a LOT...stop the talk about him being a good 2-way playe. He can be a disaster positionally in his own end... Don't confuse him skating hard with him knowing a single thing about defensive positioning) 

Oftentimes the D man gets caught between those two spots when the goal is scored and on the 5 second replay it looks like the defense didn't know what they were doing, when an actuality they were trying to cover for the awful forward play. 

There are many times the Sabres are defending a 3 on 3 but the defenseman have to play it as if it's a 3 on 2 because they just can't rely on the forwards to make a correct decision. When a goal is scored, it looks easy to place that on the D, But oftentimes it is the consistent lack of smart positioning or effort from the forwards that is more at fault, and we simply blame the defense as being bad when it's not always true.

Good post. Totally agree.

And it’s been this way for a while.

[Brawndo]

8 hours ago, PerreaultForever said:

The Sabres are supposed to be relying heavily on analytics. How's that working out?

Mixed results 

The prospect pool looks a lot better since they started helping with the draft.

Younger Players are being locked into team friendly contracts based on their recommendations. 
 

The UFA  acquisitions that have been made by the team are horrible, so that is a strike against them. 
 

Granato isn’t making the most effective lineup decisions for the team based analytics, AKA He doesn’t always listen to them. . Overplaying Zemgus and Okposo in key situations, continually playing Dahlin and Samuelsson together even though the data shows they are much better apart.

 

8 hours ago, Taro T said:

Fully expect, especially with the guys that have been hired in the last couple of years, that the Sabres analytics team is towards the top end.  Also fully believe that they look at stuff the "public" analytics crews don't.  

And, speaking as one of the "eye test" posters that realizes that analytics can be a hugely effective tool, (have every tool you can afford in your toolbox and understand how and when to use that tool) expect that a LOT of the public analytics data is ####.  Because they don't seem to differentiate  between a "high danger" chance by a Girgensons where the goalie didn't have to move to react to it and a "high danger" chance where the puck was moved there by Pasternak because the goalie already had to make a huge move to stop the potential shot from Pasternak. Personally, expect the team's analytics team does account for that.  (Which is a big part of why the guy everyone wants to see in Ra-cha-cha ain't there yet.

They do have both private data they gather themselves and data that is exclusive to NHL Teams at their disposal. I remember @Neo mentioning He had dinner with someone who worked in Hockey Analytics and confirmed how much more extensive the private data the teams have is. 
 

Of the public models Evolving Hockey and Hockey Viz are the most credible.