Salary Cap: Pens '23-'24 Salary Cap Thread: "But if you don't get the President of the Pittsburgh Penguins on that phone, you know what's gonna happen to you?"

[cygnus47]

I bought xGoals until I watched us be clearly xTremely not good for 3 seasons and never look like it had anything to do with luck or finishing.

It reflects something positive, but it's not all that close to how likely you are to outscore your opponent.

 

[Empoleon8771]

I find the "analytics mean nothing" brigade just as annoying as the "analytics mean everything" brigade. The reality is that "advanced stats" are just some measurement of shot differentials, that's it. CF is just a shot attempt and CF% just shows what percentage of shot attempts your team is taking. FF is just unblocked shot attempts and FF% just shows what percentage of unblocked shot attempts your team is taking. xGF is just shot attempts multiplied by a location factor that converts that shot attempt into an "expected goals".

The biggest issue with analytics is that people don't know how to use them. I think the largest failure of these stats was calling them "advanced stats", because they are not "advanced stats" whatsoever. They're +/- for shots effectively.

 

[Malkinstheman]

McDavid being on his 5th coach in 8 years is wild. Meanwhile Sid's had 5 coaches in 18 years lol

 

[66-30-33]

McDavid being on his 5th coach in 8 years is wild. Meanwhile Sid's had 5 coaches in 18 years lol

McDavid being on his 5th coach in 8 years is wild. Meanwhile Sid's had 5 coaches in 18 years lol

Haha that's cool.

Haha that's cool.

 

[Jacob]

I enjoy the Oilers woes because it makes our success that much better. It proves you can’t just tank. The Oilers tanked for far longer than we did, too. Hope they never win.

Caps too. Only 1 Cup is kinda mediocre for that group.

 

[FunkySeeFunkyDo]

Haha that's cool.

Haha that's cool.

You can say that again!

 

[Gurglesons]

So the Flyers have three better forward lines than the Pens? Lol

Depends if you believe GF% defines a good line.

 

[ChaosAgent]

I find the "analytics mean nothing" brigade just as annoying as the "analytics mean everything" brigade. The reality is that "advanced stats" are just some measurement of shot differentials, that's it. CF is just a shot attempt and CF% just shows what percentage of shot attempts your team is taking. FF is just unblocked shot attempts and FF% just shows what percentage of unblocked shot attempts your team is taking. xGF is just shot attempts multiplied by a location factor that converts that shot attempt into an "expected goals".

The biggest issue with analytics is that people don't know how to use them. I think the largest failure of these stats was calling them "advanced stats", because they are not "advanced stats" whatsoever. They're +/- for shots effectively.

Well said, and I'm glad people are waking up to this.

Over a long enough time horizon I'm always looking at actual GF and GA versus anything else. Short-term you have some luck.

And yes, they are basically just counting and ratio-ing more granular stats (albeit not the one that actually matters, goals).

I'd love to see one game with a shot-by-shot breakdown and a conversion to expected goals. "Here Acciari had a chance. This is worth .12 expected goals." And then let us see the video and see how it compares.

 

[Empoleon8771]

Well said, and I'm glad people are waking up to this.

Over a long enough time horizon I'm always looking at actual GF and GA versus anything else. Short-term you have some luck.

And yes, they are basically just counting and ratio-ing more granular stats (albeit not the one that actually matters, goals).

I'd love to see one game with a shot-by-shot breakdown and a conversion to expected goals. "Here Acciari had a chance. This is worth .12 expected goals." And then let us see the video and see how it compares.

That wasn't me saying analytics were worthless, I'm saying people use them wrong and both sides are wrong with how they interpret them. Just like any set of data, the problem is the people using them, not the data itself.

I've grown iffy on xGF because the location probability function that converts a shot attempt to an xG is a bit of a mystery and probably hasn't been scrutinized enough. But in their purest forms, analytics are useful tools to evaluate players.

 

[ChaosAgent]

That wasn't me saying analytics were worthless, I'm saying people use them wrong and both sides are wrong with how they interpret them. Just like any set of data, the problem is the people using them, not the data itself.

I've grown iffy on xGF because the location probability function that converts a shot attempt to an xG is a bit of a mystery and probably hasn't been scrutinized enough. But in their purest forms, analytics are useful tools to evaluate players.

I think the overall implication is that CF, CF% and Fenwick are very transparent. They aren't advanced stats. They're just +/- for shots.

I think understanding the conversion function on xG would be fantastic. As it is, it's a black box.

 

[Empoleon8771]

I think the overall implication is that CF, CF% and Fenwick are very transparent. They aren't advanced stats. They're just +/- for shots.

I think understanding the conversion function on xG would be fantastic. As it is, it's a black box.

Yeah this is exactly the issue I have with it. I think something like HDCF makes sense because you can designate "high danger" locations and just lump any shot in that area in that group. But how you go from that to converting that to an "expected goal" is something that just isn't shared and should be more heavily scrutinized.

It's not only from the aspect of "how do you calculate that function?", but it's also from the aspect of "the league is constantly evolving". The league today is fundamentally different than it was even 10 years ago, so if you're using a probability function based on data from 10 years ago, it's a flawed system.

You need a large enough sample size to be able to be confident in the function, but going too large of a sample size brings in data that isn't reflective of hockey today. It's a very valid question to ask if there is even a way to come up with that function and have it be legit, it's entirely possible that you're including non-reflective data by the point where you have a large enough sample size. If you need 5 seasons of data to have confidence in the function but goal data is only reflective from the last 2 years, how can you make a good function?

 

[Gurglesons]

That wasn't me saying analytics were worthless, I'm saying people use them wrong and both sides are wrong with how they interpret them. Just like any set of data, the problem is the people using them, not the data itself.

I've grown iffy on xGF because the location probability function that converts a shot attempt to an xG is a bit of a mystery and probably hasn't been scrutinized enough. But in their purest forms, analytics are useful tools to evaluate players.

They are useful to evaluate like minded and similarly used players. They are also useful to project a player with talent being used in a lesser role.

The whole JFresh card thing has ruined most of the discretion because if you are at a certain % you are good or bad. This was largely seen during the Matheson trade.

 

[Empoleon8771]

They are useful to evaluate like minded and similarly used players. They are also useful to project a player with talent being used in a lesser role.

The whole JFresh card thing has ruined most of the discretion because if you are at a certain % you are good or bad. This was largely seen during the Matheson trade.

This is something I really agree with, the player cards are trying to simplify analytics far too much into a "who's good based on their card?" discussion.

I still use the cards because I think it's a simple thing to point at for a support point in a discussion (like by saying "I like Kubalik as a target, here's his player card to give you an idea of what his analytics look like"), but people have really gone way too far into the "he's good because he has a good player card" or "he's bad because he has a bad player card".

 

[ChaosAgent]

Yeah this is exactly the issue I have with it. I think something like HDCF makes sense because you can designate "high danger" locations and just lump any shot in that area in that group. But how you go from that to converting that to an "expected goal" is something that just isn't shared and should be more heavily scrutinized.

It's not only from the aspect of "how do you calculate that function?", but it's also from the aspect of "the league is constantly evolving". The league today is fundamentally different than it was even 10 years ago, so if you're using a probability function based on data from 10 years ago, it's a flawed system.

You need a large enough sample size to be able to be confident in the function, but going too large of a sample size brings in data that isn't reflective of hockey today. It's a very valid question to ask if there is even a way to come up with that function and have it be legit, it's entirely possible that you're including non-reflective data by the point where you have a large enough sample size. If you need 5 seasons of data to have confidence in the function but goal data is only reflective from the last 2 years, how can you make a good function?

I think a lot of the Penguins close-in shots are from tips on point shots. Are those high probability? I don't know. I'm reading some stuff that says shot type and other factors are considered, but it's possible that for modeling purposes the location supersedes everything.

Some of the stats about the Penguins finishing last year is so mind-bogglingly bad that it can't help but overinflate the level of opportunity. That xG isn't accurate.

They are useful to evaluate like minded and similarly used players. They are also useful to project a player with talent being used in a lesser role.

The whole JFresh card thing has ruined most of the discretion because if you are at a certain % you are good or bad. This was largely seen during the Matheson trade.

This is true in any business context. Beware the pretty dashboard. Because the pretty dashboard can be treated as more valid due to the design & decoration, vs. the actual level of insight of the numbers on the screen.

 

[Ugene Magic]

I think a healthy acknowledgement that the player is 80-90% of what he was would be nice. Especially when the player plays an incredibly high event game where his decisions lead to good or bad things typically.

Regardless, this is in the wrong thread.

I omitted 2018-2019 because Kessel obliterated Malkin's year by being terrible at hockey.

This myth is just false. They were equally bad. When Malkin went down Kessel was just fine. His +/- was dropping, but once Geno got back things returned back to normal.

Maybe Malkin was just done with the Kessel partnership, but it was not all Kessel destroying his.

It's kind of hypocritical, because Karlsson is in that same kind of light. A necessary evil who also is a defensive liability. The majority of the mistakes that have cost the team were made by the stars and all equally culpable. At crucial times, they all failed the team. If you're willing to overlook their mistakes.... well.

Karlsson is a + player for the first time in 6 seasons. But that's also inflated by a +5 in the SJS game.

Geno is a +/- 0 being +4 the last 4 games while being -5 the previous 4 after being +1 through the first 5. It's a yo yo back n forth process. Same goes with Sid, Letang, Jake and...

What that say's is this team goes where they go.

 

[Andy99]

Yeah I mean if xGF correlates a lot with winning, the Pens would have made the playoffs last year… I think there’s luck and shooting skill that is not being taken into account

 

[ChaosAgent]

This myth is just false. They were equally bad. When Malkin went down Kessel was just fine. His +/- was dropping, but once Geno got back things returned back to normal.

Maybe Malkin was just done with the Kessel partnership, but it was not all Kessel destroying his.

It's kind of hypocritical, because Karlsson is in that same kind of light. A necessary evil who also is a defensive liability. The majority of the mistakes that have cost the team were made by the stars and all equally culpable. At crucial times, they all failed the team. If you're willing to overlook their mistakes.... well.

Karlsson is a + player for the first time in 6 seasons. But that's also inflated by a +5 in the SJS game.

Geno is a +/- 0 being +4 the last 4 games while being -5 the previous 4 after being +1 through the first 5. It's a yo yo back n forth process. Same goes with Sid, Letang, Jake and...

What that say's is this team goes where they go.

We should be looking at GF/GA at 5-on-5 because it isn't skewed by empty nets. Malkin and Karlsson get hindered by +/- because they're out there in situations (Powerplays, ENs for) when you will always be negative +/-. Whereas Sid, Jake and Letang get the benefit of their empty net entitlement. Which Letang definitely deserves, Sid maybe... and it's a joke that Guentzel is out there.

I'll let Kessel's '19-'20 season vs. Malkin's '19-'20 season speak for itself in terms of determining who the problem was the year before.

 

[Gurglesons]

This myth is just false. They were equally bad. When Malkin went down Kessel was just fine. His +/- was dropping, but once Geno got back things returned back to normal.

Maybe Malkin was just done with the Kessel partnership, but it was not all Kessel destroying his.

It's kind of hypocritical, because Karlsson is in that same kind of light. A necessary evil who also is a defensive liability. The majority of the mistakes that have cost the team were made by the stars and all equally culpable. At crucial times, they all failed the team. If you're willing to overlook their mistakes.... well.

Karlsson is a + player for the first time in 6 seasons. But that's also inflated by a +5 in the SJS game.

Geno is a +/- 0 being +4 the last 4 games while being -5 the previous 4 after being +1 through the first 5. It's a yo yo back n forth process. Same goes with Sid, Letang, Jake and...

What that say's is this team goes where they go.

And would that not be the same thing in a blow out against us?

 

[Empoleon8771]

Yeah I mean if xGF correlates a lot with winning, the Pens would have made the playoffs last year… I think there’s luck and shooting skill that is not being taken into account

An additional problem with xGF is that the mystical "probability function" doesn't take into account shooting skill, either. And I don't think that's the only thing it doesn't take into account.

I still think it's a generally useful thing to mention, but I think there are too many open questions and holes in it to fully treat it as gospel. I'm not 100% confident in this, but I think you could make a stronger argument using HDCF% along with GF% over trying to use xGF%.

I think HDCF% shows how many dangerous chances you're getting/giving and then GF% shows how good you are at scoring them/preventing them from going in. I think that's a stronger argument due to the uncertainties with xGF%.

 

[Ugene Magic]

We should be looking at GF/GA at 5-on-5 because it isn't skewed by empty nets. Malkin and Karlsson get hindered by +/- because they're out there in situations (Powerplays, ENs for) when you will always be negative +/-. Whereas Sid, Jake and Letang get the benefit of their empty net entitlement. Which Letang definitely deserves, Sid maybe... and it's a joke that Guentzel is out there.

I'll let Kessel's '19-'20 season vs. Malkin's '19-'20 season speak for itself in terms of determining who the problem was the year before.

So you are comparing a team with stars still doing what they do compared to a change of scenery and a far worse team as proof it was Kessel. To top it off, shorten seasons.

To even compare like to like is lunacy.

 

[Andy99]

An additional problem with xGF is that the mystical "probability function" doesn't take into account shooting skill, either. And I don't think that's the only thing it doesn't take into account.

I still think it's a generally useful thing to mention, but I think there are too many open questions and holes in it to fully treat it as gospel. I'm not 100% confident in this, but I think you could make a stronger argument using HDCF% along with GF% over trying to use xGF%.

I think HDCF% shows how many dangerous chances you're getting/giving and then GF% shows how good you are at scoring them/preventing them from going in. I think that's a stronger argument due to the uncertainties with xGF%.

Yeah, I’m not sure that’s the answer either…I just think these kind of advanced stats don’t really give a correct picture of who’s actually good because, as I’ve said before, they presume that given the exact same scoring opportunity, Jordan Staal is equally likely to put it in the back of the net as Austin Matthews…

 

[Gurglesons]

I feel like Monahan will be a Penguin tbh.

 

[Ugene Magic]

And would that not be the same thing in a blow out against us?

Possible. But not always the case. Moreso that if it was close they'd put them out there to try and get back into the game. in a 10-2 drubbing probably not forcing that narrative if they are down 7-1 fairly early in the game. Halfway point like it was in SJS. Yeah, Hertl was a minus 4 but he was the only one above -2 at -4 at forward. They dressed 7 D-men and there were 5 at -3 or above or 3 at -4.

So possible.

 

[Jacob]

I feel like Monahan will be a Penguin tbh.

I think it’ll be Zucker at the deadline.

 

[Gurglesons]

I think it’ll be Zucker at the deadline.

Arizona is making the playoffs.