Confirmed with Link: Canucks Sign D Luke Schenn to 2-Year, 1-Way Deal ($850K AAV)

[F A N]

Forgive me for not going too deeply into debates like this. Just wanna make one quick point that Quinn Hughes handled tougher minutes just fine as a rookie, but that was with Tanev of course. I remember reading analysis back then that he took over tougher matchup minutes pretty quickly by Christmas..... at least on par with the Edler pairing, and that was certainly not "sheltered". And you can't say he struggled in the tougher bubble playoff minutes, did he?

Now, obviously, Hughes struggled defensively last season. But I say there is more to it than just him alone.

Right. The thing is Hughes can't avoid playing "tough minutes" when he's leading the team in even strength ice time. It would take some mad couching skills to shelter Hughes when he's playing 19 even strength minutes a game. Even more so if the coach is supposedly sheltering Myers at the same time on another pairing. Green isn't that good.

 

[Melvin]

Correct me if I'm wrong here: You've used these very same nascent analytics in discussions here in order to make sense of events, but you think they're worthless? Am I interpreting what you've written here properly?

Forget for a moment that WAR charts are an easy grab. What you're doing here is throwing out _all_ interim data because it hasn't yet been tested by scientific method. And while that may be due process per scientific method, was this same logic applied when linear weights metrics were used prior to the peer reviewed modern WAR metrics baseball has now?

I have not read the papers on SSRN. Will look into them now.

Fundamentally, more shots leading to more goals leading more wins is the basis for Corsi. If that's still the basis, no matter primitive the captures, we should get to where baseball is now by following a similar path. The building blocks are there.

Like a lot of people I find them interesting and when I know how they are calculated and understand the limitations therein I can sometimes find some use for them, but I don't take them very seriously and I try to avoid making conclusive judgements about players based on metrics that I'm not sure have value.

Baseball is just so much different because of its history and because of the number of people involved. Someone would put out a metric and you would have hundreds of nerds debating it for days, weeks, years, tearing it apart, finding bugs, finding problems, testing all the assumptions, etc. Article after article would get written about the flaws of this metric or that metric. There are no perfect metrics, not in baseball, not in anywhere. That "peer review" period doesn't exist in hockey, as far as I can tell. We skipped all that.

The reason why I put my stuff here instead of on my own website or whatever is because I want people to review it with me. I want to test it out. I want to gather feedback and make corrections. I probably could just put it on a website, make it sound like it's some sort of polished product that it isn't, set up a Patreon and start spamming twitter in the hopes that people will get my projections lots of attention and maybe get me some money, but that's not how it should work.

In Evolving-Wild's introduction of their "WAR" metric, part 3 includes this absolutely hilarious paragraph:

It is a general practice at this point to present studies evaluating the overall performance of the metric that has been described. We will be saving that for its own article. This can be found here [future link goes here], but please don’t wait for that! If you’d like to test out how well our WAR model does, please feel free to test it out. It’s public, it’s (now) explained, and we’re more than willing to answer any questions anyone may have!

Man, testing things suuuuucks. We'll get around to it eventually, I promise! (Narrator: they never got around to it.) But you can test our metrics if you want! Why should we test the things that we are putting out there to be used by the public while asking for money on Patreon? Can't someone else do it!?

It's pathetic. A statement like this should be completely unacceptable. But nobody cares, because I'm the first person who has read that article in 3 years. Take the numbers, don't bother to understand how they were derived, don't bother to test their utility, just assume they must be good because the people using them have a lot of Twitter followers. That's how it works right? M*** McC****y is popular, so what he says must be true. Joe Rogan is popular, so what he says must be true. Donald Trump is popular, so what he says....

 

[MS]

How are you better understanding the context by removing data and relying solely on the eye test?

Your statement says "Context for the advanced stats you are seeing"... What does this look like? Please show me using data.

Is the Horvat sub-replacement example a reference to my past exchanges with Scurr? Interesting if so. You're not saying anything wrong in your first paragraph, but you're speaking about inference, not the data. You're saying WAR needs to be contextualized further, not that WAR itself is a useless stat. Much different than saying the statistics themselves are garbage.

If you hate that WAR is presented without context, provide the context that is generally applicable. What you're doing is being dismissive, providing no context yourself, and then saying everyone else is incorrect for using stats poorly. How is this productive?

Yes, using stats without the perfect contextual basis leaves them open for critique. Leaves the inference subject to question. So add the context. You're only adding eye test qualifiers to dismiss statistical assertions = Does not work.

My biggest memory was y2kcanucks saying that Horvat should be dumped because he was a -WAR player. I've seen it other places as well but don't specifically remember the interaction you're talking about.

When I explain my usage of a stat, I explain the judgements I've come to based on my understanding of the context.

Like I said, raw data has value as a tool if you understand context.

Some guy in his basement taking that raw data and turning it into a WAR stat through measures that aren't properly explained, and generating results that can obviously be debunked as very flawed : worthless.

Baseball is great. This stuff has been developed for decades and the sport itself lends it to it spectacularly (as an aside, someone could make $millions applying this stuff to cricket which is still decades behind). When you're using baseball numbers you can be like 95% comfortable that what you're seeing there is a quality stat.

The WAR stuff in hockey is crap. It's marginally better than just taking raw +/- out of context with a few basic adjustments. If my universal way of evaluating players including eye tests and context is leading me to be 80% accurate in evaluating players, it might not be as good as MLB WAR but it's a hell of a lot better than people who are accurate 55% of the time mindlessly spewing bad graphs out of context.

 

[Bleach Clean]

Like a lot of people I find them interesting and when I know how they are calculated and understand the limitations therein I can sometimes find some use for them, but I don't take them very seriously and I try to avoid making conclusive judgements about players based on metrics that I'm not sure have value.

Baseball is just so much different because of its history and because of the number of people involved. Someone would put out a metric and you would have hundreds of nerds debating it for days, weeks, years, tearing it apart, finding bugs, finding problems, testing all the assumptions, etc. Article after article would get written about the flaws of this metric or that metric. There are no perfect metrics, not in baseball, not in anywhere. That "peer review" period doesn't exist in hockey, as far as I can tell. We skipped all that.

The reason why I put my stuff here instead of on my own website or whatever is because I want people to review it with me. I want to test it out. I want to gather feedback and make corrections. I probably could just put it on a website, make it sound like it's some sort of polished product that it isn't, set up a Patreon and start spamming twitter in the hopes that people will get my projections lots of attention and maybe get me some money, but that's not how it should work.

In Evolving-Wild's introduction of their "WAR" metric, part 3 includes this absolutely hilarious paragraph:

Man, testing things suuuuucks. We'll get around to it eventually, I promise! (Narrator: they never got around to it.) But you can test our metrics if you want! Why should we test the things that we are putting out there to be used by the public while asking for money on Patreon? Can't someone else do it!?

It's pathetic. A statement like this should be completely unacceptable. But nobody cares, because I'm the first person who has read that article in 3 years. Take the numbers, don't bother to understand how they were derived, don't bother to test their utility, just assume they must be good because the people using them have a lot of Twitter followers. That's how it works right? M*** McC****y is popular, so what he says must be true. Joe Rogan is popular, so what he says must be true. Donald Trump is popular, so what he says....

Never worked that way with me. I try to question everything that is said no matter who is saying it.

Anyway, I can respect the endeavor to want to test analytics here. However, I hope you realize that this is not the way it has always come across? You've used stats to shut down discussions here, and so have I. We've used them in the authoritative.

Advanced stats have cracked open stalemates here. Burst them wide open. They've also layered our understanding of players. Therefore, I cannot acknowledge that they are useless. And in a way, you've done the same by stating as long as you understand their limitations and calculations, you find a use as well. Meaning, they're not useless, just conditions have to be met first.

That's what wrestling with the information yields over time. The space includes both groups: The lowest common denominator of fan looking for quick grabs, and the fans who have achieved real understanding. If I see someone struggling with the data, that poster gets my respect and not my derision.

 

[Peen]

My biggest memory was y2kcanucks saying that Horvat should be dumped because he was a -WAR player. I've seen it other places as well but don't specifically remember the interaction you're talking about.

When I explain my usage of a stat, I explain the judgements I've come to based on my understanding of the context.

Like I said, raw data has value as a tool if you understand context.

Some guy in his basement taking that raw data and turning it into a WAR stat through measures that aren't properly explained, and generating results that can obviously be debunked as very flawed : worthless.

Baseball is great. This stuff has been developed for decades and the sport itself lends it to it spectacularly (as an aside, someone could make $millions applying this stuff to cricket which is still decades behind). When you're using baseball numbers you can be like 95% comfortable that what you're seeing there is a quality stat.

The WAR stuff in hockey is crap. It's marginally better than just taking raw +/- out of context with a few basic adjustments. If my universal way of evaluating players including eye tests and context is leading me to be 80% accurate in evaluating players, it might not be as good as MLB WAR but it's a hell of a lot better than people who are accurate 55% of the time mindlessly spewing bad graphs out of context.

The Horvat WAR stuff was really popular on twitter and y2k is active there so it correlates.

 

[Bleach Clean]

My biggest memory was y2kcanucks saying that Horvat should be dumped because he was a -WAR player. I've seen it other places as well but don't specifically remember the interaction you're talking about.

When I explain my usage of a stat, I explain the judgements I've come to based on my understanding of the context.

Like I said, raw data has value as a tool if you understand context.

Some guy in his basement taking that raw data and turning it into a WAR stat through measures that aren't properly explained, and generating results that can obviously be debunked as very flawed : worthless.

Baseball is great. This stuff has been developed for decades and the sport itself lends it to it spectacularly (as an aside, someone could make $millions applying this stuff to cricket which is still decades behind). When you're using baseball numbers you can be like 95% comfortable that what you're seeing there is a quality stat.

The WAR stuff in hockey is crap. It's marginally better than just taking raw +/- out of context with a few basic adjustments. If my universal way of evaluating players including eye tests and context is leading me to be 80% accurate in evaluating players, it might not be as good as MLB WAR but it's a hell of a lot better than people who are accurate 55% of the time mindlessly spewing bad graphs out of context.

I've had my own exchanges with poster Scurr about "sub-replacement level Horvat". That's what I thought you were referencing.

Who is the basement nerd using improperly explained WAR measures? Hohl, Perry or other? This: Evolving-Hockey OR this: Wins Above Replacement

Yes, I understand that they use adjusted plus-minus.

I'm thinking you must see GAR the same as WAR? Forget who is using it, I'm just talking the information the stat conveys.

Your eye test is at once subjective and unassailable given that an alternative eye test or application of statistics wouldn't have changed your opinion. Do you not see that as an issue? No one can change your opinion on a player mid-stream. It's only when the player proves you wrong, then you adjust... but only slightly so.

Example: Hughes. Your opinion of him supplanting Dobson and his development curve came after the season. However, you never changed his ultimate ceiling (I've been looking for this post still): 2nd unit ES/1st unit PP. We don't even have your assessment of what he was that year? Was he just a 2nd unit/1 PP guy?

When I went back to employ stats, some were ill-used (All Situations) and some were dismissed (GAR). So even an attempt to quantify his impact that year was hand waived... Even if that attempt may just lead to him being more than just a 2nd unit ES player.

Do you see how one could conclude that data will not supplant your cap for a given player, even when we are looking at the same season's performance? This is the heart of the issue. Nothing that assails your understanding is valid... until it is, and even then, it can't be data that does it.




In the end, you will change your opinion based upon what is shown on ice, but that's still a narrow assertion.

Let's use Hughes again as an example: His rookie year performance had you recognize him as a better selection than Dobson, afterward. It had you reconsider his development timeline, but not his ultimate ceiling as a 2nd unit/1st PP player. Illogical. If you could be wrong about his transition and his projection, you can be wrong about his performance that year by the numbers too.

And granted, you've admitted you've been wrong before. Respect there, but hasn't been almost exclusively after the fact? Like with Hughes,

 

[jd22]

I appreciate the nuanced discussion here. Thanks to all the posters and particularly Melvin who has been doing this for years.

 

[CRDragon]

Do I trust my eye test more than just using those shit charts in a vacuum? Abso-f***ing-lutely.

I honestly don't know why. But reading this made me think this:

Covid isn't real because I can't see it so I trust that isn't real over real science in charts.

I haven't read the whole thread btw, so not sure what the discussion is about, just wanted to point this out.

 

[Melvin]

I honestly don't know why. But reading this made me think this:

Covid isn't real because I can't see it so I trust that isn't real over real science in charts.

I haven't read the whole thread btw, so not sure what the discussion is about, just wanted to point this out.

I mean, if you catch up on the thread you'll see my later response, but these people are not scientists. Their work is not tested, not subjected to peer-review, not following the scientific method. It's far, far closer to pseudo-science, where someone comes up with a hypothesis and then starts selling products based on the hypothesis without bothering to check and see if it's true or not.

And it's a bad analogy anyway, because MS seems to watch a lot of hockey games. If you somehow could meet a substantial number of people, visit hundreds of hospitals and never come across COVID, it would be reasonable to question it. If someone who watches a lot of hockey games and knows a lot about hockey criticizes the charts, the proper response is to dig into the data to try to answer the questions, not to blindly side with unknown people and their untested hypothesis just because their charts are pretty and they have a lot of followers. Facing criticism of your data and answering the criticism is part of the scientific method.

There is a severe GIGO problem when it comes to hockey analytics. I have talked about this before - there is a phenomenon in statistics where people "make do with what they have," and assume that the best that they have is therefore the best there could be, which causes them to make silly conclusions simply because the limited data they have tells them something that a more rich dataset wouldn't. MS is 100% correct that, fundamentally, all of these statistics are built off of +/- and suffer the same problems as raw +/- = they make the assumption that if the player is on the ice for a particular event he must bear some responsibility for that event. They make that assumption because they have to make that assumption, we don't have any better data that we can use. But that doesn't mean the conclusions drawn from this limited data is any more conclusive than the eye test; that simply has not been proven. The core of science is that every hypothesis must be tested rigorously. That is what science is about.

If all we had to go off was raw +/-, we would make some very silly statements indeed. We would say, in 2020, why is Connor McDavid getting 13M? His +/- is -6!! Yikes! And if that was all the data we had, we would have no choice but to conclude this. And every single person using the eye test that says McDavid is good would be 100% in the right. Because our data is slightly better, we know better than to rely on +/-, but the tools we do use are still very, very crude and unsophisticated, and jumping to the conclusion that it is better than the eye test simply because it's the best that we have is not supported by any evidence. Just because the conclusions you are drawing is based on the best statistical data available, does not necessarily make it better than observational data. That needs to be proven.

I was very excited for the player tracking stats they were going to implement for the 2020 season. This was, IMO, going to be a monumental leap forward in our understanding (after about 3 years of data collection,) but it looks like with COVID this was scrapped. I am sad. But consider for a second the following scenario:

1) WAR charts say Darnell Nurse is bad, MS says Nurse is good.
2) New metrics come out based on granular player tracking.
3) The public shifts to the new metrics and scraps the old WAR charts.
4) New metrics show that Darnell Nurse is good.

This is a simple example, but with better data we might well find that the data we are currently using is actually worse than the observational data. In the same way where people today would be laughing at the (hypothetical) people who base things entirely on raw +/-. Our understanding will improve as our data improves, but simply stating that whatever data we have is better than the eye test is a falsehood. Whether it is or it isn't currently is uncertain.

 

[CRDragon]

I mean, if you catch up on the thread you'll see my later response, but these people are not scientists. Their work is not tested, not subjected to peer-review, not following the scientific method. It's far, far closer to pseudo-science, where someone comes up with a hypothesis and then starts selling products based on the hypothesis without bothering to check and see if it's true or not.

And it's a bad analogy anyway, because MS seems to watch a lot of hockey games. If you somehow could meet a substantial number of people, visit hundreds of hospitals and never come across COVID, it would be reasonable to question it. If someone who watches a lot of hockey games and knows a lot about hockey criticizes the charts, the proper response is to dig into the data to try to answer the questions, not to blindly side with unknown people and their untested hypothesis just because their charts are pretty and they have a lot of followers. Facing criticism of your data and answering the criticism is part of the scientific method.

There is a severe GIGO problem when it comes to hockey analytics. I have talked about this before - there is a phenomenon in statistics where people "make do with what they have," and assume that the best that they have is therefore the best there could be, which causes them to make silly conclusions simply because the limited data they have tells them something that a more rich dataset wouldn't. MS is 100% correct that, fundamentally, all of these statistics are built off of +/- and suffer the same problems as raw +/- = they make the assumption that if the player is on the ice for a particular event he must bear some responsibility for that event. They make that assumption because they have to make that assumption, we don't have any better data that we can use. But that doesn't mean the conclusions drawn from this limited data is any more conclusive than the eye test; that simply has not been proven. The core of science is that every hypothesis must be tested rigorously. That is what science is about.

If all we had to go off was raw +/-, we would make some very silly statements indeed. We would say, in 2020, why is Connor McDavid getting 13M? His +/- is -6!! Yikes! And if that was all the data we had, we would have no choice but to conclude this. And every single person using the eye test that says McDavid is good would be 100% in the right. Because our data is slightly better, we know better than to rely on +/-, but the tools we do use are still very, very crude and unsophisticated, and jumping to the conclusion that it is better than the eye test simply because it's the best that we have is not supported by any evidence. Just because the conclusions you are drawing is based on the best statistical data available, does not necessarily make it better than observational data. That needs to be proven.

I was very excited for the player tracking stats they were going to implement for the 2020 season. This was, IMO, going to be a monumental leap forward in our understanding (after about 3 years of data collection,) but it looks like with COVID this was scrapped. I am sad. But consider for a second the following scenario:

1) WAR charts say Darnell Nurse is bad, MS says Nurse is good.
2) New metrics come out based on granular player tracking.
3) The public shifts to the new metrics and scraps the old WAR charts.
4) New metrics show that Darnell Nurse is good.

This is a simple example, but with better data we might well find that the data we are currently using is actually worse than the observational data. In the same way where people today would be laughing at the (hypothetical) people who base things entirely on raw +/-. Our understanding will improve as our data improves, but simply stating that whatever data we have is better than the eye test is a falsehood. Whether it is or it isn't currently is uncertain.

Appreciate this response! Loving it.

 

[ahmon]

Hockey has just too many variables and factors that can affect the outcomes to use something primitive like shots for and against as base to determine the value of a player.

Just think about if these "advanced stats" ever account for:

1. If its based on shots for/against, how does it account for shot quality? And I've seen those heat maps with shot against/for. Just because you get a shot off in the slot. How close the defenders are to you will highly affect whether that is actually a dangerous shot. Also where that pass came from will actually affect how dangerous that shot is - ie whether the goalie has to travel from the across the net or not etc.

2. Hockey is played as a string of events, its not like baseball where each batter goes to bat and the count starts. How your previous linemates played has quite a big impact on how you start your shift.
Are you jumping over the boards and your team is on your heels, or your jumping over the boards and your team has already hemmed the opposition in for like 1 minute?

Thats just to name a couple elements.

Not only is who your linemates and/or pairing is important, your entire team will affect your performance. Its no surprise generally the players on the best teams will have great analytics and when you move one of these players to a worse team their analytics will drop.

If the goal is to figure out who is actually good, have to be able to actually weed out the team effect. And from what I've seen, practically all the advanced stats are heavily influenced by team, which means they do a very poor job of identifying player performance.

 

[Bleach Clean]

This is a simple example, but with better data we might well find that the data we are currently using is actually worse than the observational data. In the same way where people today would be laughing at the (hypothetical) people who base things entirely on raw +/-. Our understanding will improve as our data improves, but simply stating that whatever data we have is better than the eye test is a falsehood. Whether it is or it isn't currently is uncertain.

Not whatever data, no, but data that has been used appropriately? Yes. The likelihood that it is better than a random watcher's eye test is good.

One more clarification on a point you made: The critique of a chart actually places the onus on both parties to present a case, not just the OP. If the data being critiqued has been 'earned' properly in the public sphere, like GAR, the OP should explain its makeup and the critic should explain why it's not valid. Simply dismissing the presentation of GAR for one's own perception is wrong.

 

[Melvin]

Not whatever data, no, but data that has been used appropriately? Yes. The likelihood that it is better than a random watcher's eye test is good.

One more clarification on a point you made: The critique of a chart actually places the onus on both parties to present a case, not just the OP. If the data being critiqued has been 'earned' properly in the public sphere, like GAR, the OP should explain its makeup and the critic should explain why it's not valid. Simply dismissing the presentation of GAR for one's own perception is wrong.

You're not saying anything here. "The likelihood that it is better than a random watcher's eye test is good." How do you know? What is "appropriately?" What are you basing this on? What is "'earned'?" What is the "public sphere." I mean, this is gibberish. Your arguments in this realm always boil down to dressed-up arguments from popularity. People you know and like use these tools, so these tools must be good. This is wrong. This is as wrong as it gets. This line of reasoning is responsible for an uncountable amount of blunders throughout human history. GAR should be assumed worthless until it has demonstrated value. I don't care if Micah McMurdy uses it or petbugs uses it or any number of really popular twitter people use it, none of that adds up to a fraction of a damn.

As for "dismissing it for one's own perception." Yes, MS does this sometimes. But sometimes he doesn't:

But his defensive game is not represented accurately by that graph. He's out there with McDavid/Draisaitl playing pond hockey and Bear/Barrie as his weak partner, with the entire unit geared toward generating offense and outscoring the opposition. Scoring chances are going to happen the other way and it's a result of the context, not that Nurse is somehow one of the worst defensive defenders in the NHL.

This is very direct, well-reasoned and concrete criticism. The appropriate response would be to study this, study the effects of the play style on the numbers and give an equally reasoned response, not to dismiss the criticism because you think any and all numbers are always better than observational data. The only way to improve these metrics would be to take the criticism to hard and study them. Maybe you can prove MS wrong, maybe you can't, maybe he's correct about this, but you won't know until you study it. Saying "NUMBERS GOOD OPINIONS BAD" gets us nowhere.

 

[vancityluongo]

god i wish there weren't character limits on twitter and hf didn't have a dumb perception as a cesspool of morons just because of the trade proposals board

@Melvin and @MS in long form writing essays about this shit is something i would actually subscribe to as a patreon lol. bring in some of the reasonable people like dom and cam lawrence for guest rebuttals every once in a while.

 

[VanJack]

For a guy who was a serviceable depth d-man on a team that won back-to-back Stanley Cups, a two year contract of $850,000 a season is either outstanding value, or possibly a reflection of his perceived value around the league.

It's kind of hard not to believe that if he'd waited a little longer he'd have done far better on the UFA market--particularly given the dire shortage of right-shot d-men around the league.

But one thing's for certain. Schenn will have a lot more trouble struggling to keep up on the Canucks blueline than he had in Tampa. If Schmidt regressed and was begging to leave, I can't imagine Schenn faring a lot better.

The reality is that Green's defensive system--or what passes for one--exposes blueliners to all kinds of compromised situations out there. It's just a tough team to play defense for.

 

[F A N]

For a guy who was a serviceable depth d-man on a team that won back-to-back Stanley Cups, a two year contract of $850,000 a season is either outstanding value, or possibly a reflection of his perceived value around the league.

It's kind of hard not to believe that if he'd waited a little longer he'd have done far better on the UFA market--particularly given the dire shortage of right-shot d-men around the league.

But one thing's for certain. Schenn will have a lot more trouble struggling to keep up on the Canucks blueline than he had in Tampa. If Schmidt regressed and was begging to leave, I can't imagine Schenn faring a lot better.

The reality is that Green's defensive system--or what passes for one--exposes blueliners to all kinds of compromised situations out there. It's just a tough team to play defense for.

Luke Schenn is from SASK, played junior hockey in Kelowna, got married in Kelowna, and has his off season home in Kelowna (I believe his wife is from Kelowna?). I think he would fare much better playing in the City of Vancouver than Schmidt.

 

[ProstheticConscience]

I can't help but think Luke Schenn would be very, very confused if he casually clicked on a random page in this thread.

 

[dwarf]

I am happy we signed him. If he plays with any intensity at all, he can help win hockey games. And he is tough as nails, he will make our players bigger.

 

[Siludin]

Luke Schenn has been a great option for the Canucks.
He comes with his flaws but he seems to calm the team down when he's in the lineup.
Willie Mitchell lite? I like him.

 

[Zippgunn]

Luke Schenn has been a great option for the Canucks.
He comes with his flaws but he seems to calm the team down when he's in the lineup.
Willie Mitchell lite? I like him.

I think he has been worth every cent of his piddling little contract. I love that: "Willie Mitchell lite"...