Seider vs Power - Today and Future

[frisco]

I think Seider right now, but Power is a year and a half younger and that's a pretty big difference at this stage in their development. So, Power later with more upside to catch and exceed Seider.

My Best-Carey

 

[biturbo19]

This seems like exactly what Philly fans said about Provorov for ages. The analytics never caught up to his reputation.

I think part of the issue is that there are players like this, Seider, Provorov, Trouba, Doughty has fallen into it at times, etc...where they're trusted by coaches to player extremely heavy duty matchup minutes and something about their ~style~ of play just seems to be a bit of a blip that is incongruous with a lot of popularly touted analytics models. They just don't play in a way that flatters. They're the "eye test is better than the analytics" guys. The opposite end of the spectrum from the soft minutes offensive leaches who come off with superlative sparkling analytics profiles, but coaches don't trust them with any kind of matchups and consistently try to "hide" them, and the eye test tends to back up the idea that they're not nearly as good a defenceman as the numbers reflect. Guys like Bouchard, where things look peachy until you realize he plays stapled to the best player in the game and benefits immensely from that...and until you actually watch him play and the pathetic defensive effort involved all too often.

Once you get into the margins, outliers, and extreme cases with unusual context...things can quickly get very distorted with the publicly used analytics.

Agreed to some extent. But getting demolished even in extremely tough minutes doesn’t really do anyone any favors, and makes it impossible to measure any progress. Incrementally wouldn’t throwing Holl and Chiarot out there in the same usage while sheltering Seider a bit be much better for the team?

This seems very stat-brained to me. The idea that a team would somehow be better off getting their bad defencemen caved in massively more than their #1D in those hardest minutes, and would somehow be better off overall because Seider's analytics would be better?

That's completely missing the forest for the trees.

 

[KirkAlbuquerque]

Seider ?

Power ?


I hardly know ‘er !

 

[ViewsFromThe6ix]

I think part of the issue is that there are players like this, Seider, Provorov, Trouba, Doughty has fallen into it at times, etc...where they're trusted by coaches to player extremely heavy duty matchup minutes and something about their ~style~ of play just seems to be a bit of a blip that is incongruous with a lot of popularly touted analytics models. They just don't play in a way that flatters. They're the "eye test is better than the analytics" guys. The opposite end of the spectrum from the soft minutes offensive leaches who come off with superlative sparkling analytics profiles, but coaches don't trust them with any kind of matchups and consistently try to "hide" them, and the eye test tends to back up the idea that they're not nearly as good a defenceman as the numbers reflect. Guys like Bouchard, where things look peachy until you realize he plays stapled to the best player in the game and benefits immensely from that...and until you actually watch him play and the pathetic defensive effort involved all too often.

Once you get into the margins, outliers, and extreme cases with unusual context...things can quickly get very distorted with the publicly used analytics.



This seems very stat-brained to me. The idea that a team would somehow be better off getting their bad defencemen caved in massively more than their #1D in those hardest minutes, and would somehow be better off overall because Seider's analytics would be better?

That's completely missing the forest for the trees.

I agree with some of this (namely the Bouchard stuff), but Seider's xGF% 5on5 is at 43.81%. The worst in the league is about 40%. There is a floor to the absolute minimum a player can perform at as an NHL player. Throwing out 2 replacement level D would certainly give similar (relatively) results, and boost Seider's effectiveness, IMO.

 

[Albatros]

Yeah, embrace the tank and get him a Norris playing with Ghost.

 

[biturbo19]

I agree with some of this (namely the Bouchard stuff), but Seider's xGF% 5on5 is at 43.81%. The worst in the league is about 40%. There is a floor to the absolute minimum a player can perform at as an NHL player. Throwing out 2 replacement level D would certainly give similar (relatively) results, and boost Seider's effectiveness, IMO.

The thing is...there really isn't a "floor" to how many goals a defence pairing can concede in the toughest matchups.

That's a limitation of the metric, not a limitation on reality. xGF% is one of those things like i was saying, that gets really distorted at the boundaries and fringes and weird and unusual situations.


At the end of the day, Seider's actual GF% outperforms his xGF% by a decent amount. As does Holl's in his own much lesser role. So at the end of the day...if you're getting better GF% results deploying them like that, it's better for your team overall.

Just look at what happened to both of Hronek's GF% and xGF% when he went from that Detroit team to a different situation riding shotgun to Quin Hughes in Vancouver. Suddenly they both jumped hugely and the actual % results went through the roof to massive levels.

It's a deeply flawed metric that doesn't seem particularly capable of properly assessing context and deployment. Especially at the extremes like this.

It's often really just more of a metric that like fuddy duddy old +/- is more of an indicator in how well situated a player is in the minutes and role they've been handed. After all, it's basically just...+/- for nerds.

And in that context, i think most would agree that Seider is being buried in minutes that are a bit beyond him, but that's due to lack of better alternatives to play those minutes more effectively. Having him handle those better than anyone else also allows a guy like Holl to play softer minutes and tip the scales heavily to the positive side on GF%. Which means...more goals for than against. By a wide margin. Which is good for the team.

 

[norrisnick]

Interesting article dropped today. Remember folks always keep firm control of A, C, and V...

A deep dive into how usage and teammates affect a player's on-ice numbers

In the NHL, some jobs are easier than others and that needs to be better acknowledged when comparing numbers. We've made the adjustment.

theathletic.com

 

[RedHawkDown]

Interesting article dropped today. Remember folks always keep firm control of A, C, and V...

A deep dive into how usage and teammates affect a player's on-ice numbers

In the NHL, some jobs are easier than others and that needs to be better acknowledged when comparing numbers. We've made the adjustment.

theathletic.com

Very cool article. Explains it very well - Seider's usage is the hardest in the league. He's not killing it in that usage, but comparing him to other defensemen is pointless given that nobody plays the minutes his does, and it's not close. Hopefully when it eases up a bit in the coming years as the team defense improves, we will see what he's actually capable of when he's not just trying to stay afloat.

 

[Aladyyn]

Very cool article. Explains it very well - Seider's usage is the hardest in the league. He's not killing it in that usage, but comparing him to other defensemen is pointless given that nobody plays the minutes his does, and it's not close. Hopefully when it eases up a bit in the coming years as the team defense improves, we will see what he's actually capable of when he's not just trying to stay afloat.

I completely agree with this but then why is Seider winning the "now" portion of the poll so hard if we have no idea what he actually is?

 

[ijuka]

Need to be quite the optimist to take Power over Seider.

 

[norrisnick]

I completely agree with this but then why is Seider winning the "now" portion of the poll so hard if we have no idea what he actually is?

Because even what he is showing now is better than Power currently.

 

[newfy]

I completely agree with this but then why is Seider winning the "now" portion of the poll so hard if we have no idea what he actually is?

We have an idea of what he is, if you actually watch the games. If you rely on flawed advanced stats to judge a player who is one of the biggest outliers deployment wise the last decade, youre not going to know. The athletic article about him touches on it, but I believe it was Jfresh saying hes trying to develop a stat where it judges if you leave your team in a better position than when you got on it to account for how ridiculous Seiders usage has been.

You have a couple big hockey stat nerds essentially saying his usage is so difficult they need to make new stats to account for it, and his GF% is still respectable all things considered.

What has Power done in the NHL to even be considered in the same ball park right now?

 

[Breakers]

Seider's underlying numbers make no sense to me

this guy plays essentially some of the toughest minutes in the league
but his defensive results are atrocious.

 

[norrisnick]

Seider's underlying numbers make no sense to me

this guy plays essentially some of the toughest minutes in the league
but his defensive results are atrocious.

He's being asked to play "bend don't break" defense every single shift because the coaching staff doesn't trust anyone else with meaningful assignments. That's not going to look good via analytics.

And it's not essentially. It's #1 with a bullet.

 

[dgibb10]

The thing is...there really isn't a "floor" to how many goals a defence pairing can concede in the toughest matchups.

That's a limitation of the metric, not a limitation on reality. xGF% is one of those things like i was saying, that gets really distorted at the boundaries and fringes and weird and unusual situations.


At the end of the day, Seider's actual GF% outperforms his xGF% by a decent amount. As does Holl's in his own much lesser role. So at the end of the day...if you're getting better GF% results deploying them like that, it's better for your team overall.

Just look at what happened to both of Hronek's GF% and xGF% when he went from that Detroit team to a different situation riding shotgun to Quin Hughes in Vancouver. Suddenly they both jumped hugely and the actual % results went through the roof to massive levels.

It's a deeply flawed metric that doesn't seem particularly capable of properly assessing context and deployment. Especially at the extremes like this.

It's often really just more of a metric that like fuddy duddy old +/- is more of an indicator in how well situated a player is in the minutes and role they've been handed. After all, it's basically just...+/- for nerds.

And in that context, i think most would agree that Seider is being buried in minutes that are a bit beyond him, but that's due to lack of better alternatives to play those minutes more effectively. Having him handle those better than anyone else also allows a guy like Holl to play softer minutes and tip the scales heavily to the positive side on GF%. Which means...more goals for than against. By a wide margin. Which is good for the team.

The context in both these situations is the fact that detroit and vancouver are shooting significantly above expectation as teams. Simple as that. That's why their goals% are above their xGoals%s. Because as a team vancouver is 20+ goals above expectation and detroit is 30

 

[Pavels Dog]

Seider's underlying numbers make no sense to me

this guy plays essentially some of the toughest minutes in the league
but his defensive results are atrocious.

Defensive analytics (at least publically available ones) are trash. People criticize him based on charts that also tell you Olli Määttä might be in consideration for the Norris.

He's been far from perfect, but it really doesn't take much viewing to understand what he brings to the table.

For example, he's the only player in the NHL this year with both 180+ blocked shots and 180+ hits. The last 3 years combined he's one of just two players with 500+ hits and 500+ blocked shots (the other is Trouba).

He does so much heavy lifting for Detroit it's frankly ridiculous people can't see past the analytics.

 

[biturbo19]

The context in both these situations is the fact that detroit and vancouver are shooting significantly above expectation as teams. Simple as that. That's why their goals% are above their xGoals%s. Because as a team vancouver is 20+ goals above expectation and detroit is 30

That's not the context. That's just reading what the numbers say.

The context is that the Red Wings traded Hronek away...and still saw a major GF% advantage over the xGF% metric. The Canucks received Hronek and also saw GF% increases, and Hronek individual shot through the roof. So math that math out...if it's all an easy statistical zero sum game.


That's where the context says you need to take a much harder look at what's really going on here and why some teams and certain players might see major individual fluctuations or distortions in this metric. Across different teams even. That aren't just tied to some specific team having magic voodoo to create this significant gulf between xGF% and actual GF%.

ie. It's not like Detroit just sent all their "good GF% juju to Vancouver" with Hronek trade. They've done well in that respect anyway. But Vancouver also got much better in that same respect.

Until advanced stats can explain stuff like that...they're only a very tiny part of the picture that will not outweigh actual results.

 

[Aladyyn]

He's being asked to play "bend don't break" defense every single shift because the coaching staff doesn't trust anyone else with meaningful assignments. That's not going to look good via analytics.

And it's not essentially. It's #1 with a bullet.

It's so funny how you could go back to ~2016 Ristolainen threads and find basically the exact same posts.

 

[newfy]

It's so funny how you could go back to ~2016 Ristolainen threads and find basically the exact same posts.

Risto had one season where he actually played the hardest minutes on his team. Seiders minutes are actually noticeably more difficult, so think back to how bad Risto was thrown to the wolves, multiply it, and then have better results. Thats what Seider is doing. We're talking some of the hardest minutes ever tracked, not your eye balls watching Risto get lit up aginst good players type tracking

 

[Aladyyn]

Risto had one season where he actually played the hardest minutes on his team. Seiders minutes are actually noticeably more difficult, so think back to how bad Risto was thrown to the wolves, multiply it, and then have better results. Thats what Seider is doing. We're talking some of the hardest minutes ever tracked, not your eye balls watching Risto get lit up aginst good players type tracking

Yeah he's playing Dan Girardi minutes and getting Dan Girardi results

 

[dgibb10]

That's not the context. That's just reading what the numbers say.

The context is that the Red Wings traded Hronek away...and still saw a major GF% advantage over the xGF% metric. The Canucks received Hronek and also saw GF% increases, and Hronek individual shot through the roof. So math that math out...if it's all an easy statistical zero sum game.


That's where the context says you need to take a much harder look at what's really going on here and why some teams and certain players might see major individual fluctuations or distortions in this metric. Across different teams even. That aren't just tied to some specific team having magic voodoo to create this significant gulf between xGF% and actual GF%.

ie. It's not like Detroit just sent all their "good GF% juju to Vancouver" with Hronek trade. They've done well in that respect anyway. But Vancouver also got much better in that same respect.

Until advanced stats can explain stuff like that...they're only a very tiny part of the picture that will not outweigh actual results.

Both teams are having a large element of good luck. Simple as that.

You could attribute it to a coaching change in vancouver you want, or bringing in Kane and Debrincat. But more realistically it's luck.

 

[Faceboner]

Seider's underlying numbers make no sense to me

this guy plays essentially some of the toughest minutes in the league
but his defensive results are atrocious.

Sabres were high on risto, seider is in a similar situation with a better forward group ristolainen was given the same excuses and reasoning for his poor metrics

 

[TheDawnOfANewTage]

Context for those analytics is a MFer.

This is the quality of competition usage of all the #1 D (by TOI) in the league. I'm going to go out on a limb and say that no one would excel with his deployment. His rookie year he was at Parayko's QoC. Last year he was around Matheson's QoC. The hope is that with Edvinsson on the roster he can start taking some of those shifts so Seider gets a breather now and then and pad the analytics for all the moneyballers.

View attachment 841474

This is a good post.

I’m still picking Power, biased, but this is useful info.

 

[Pavels Dog]

Yeah he's playing Dan Girardi minutes and getting Dan Girardi results

Lack of hockey knowledge isn’t a badge to be worn as proudly as you do it.

 

[nbwingsfan]

It's so funny how you could go back to ~2016 Ristolainen threads and find basically the exact same posts.

Seider is playing Vastly harder minutes than he did while also performing better.

But sure, if you completely ignore that then they’re the same