Speculation: Caps Roster General Discussion (Coaching/FAs/Cap/Lines etc) - 2022-23 Season Part 2: Regular Season

[kicksavedave]

Lord how I've missed these endless arguments about how good a player is based on what ever suits our fancy at the moment. I mean thank god we don't have to talk about the Redskins anymore. /S

 

[twabby]

View attachment 591276
from evolving hockey, eller was at -2.33

Is this Evolving Hockey’s RAPM xG differential?

 

[traparatus]

Sure, it would be nice to see CMM get time on PP2, maybe below the goal line, as he seems to do well in that area.

So, the problem I was referencing is when the PK’ers don’t honor the guy in the bumper, and you have two guys just committed to blocking Ovi’s shot. Oshie often buys him that extra half-second. Without him, even w Willy, the group just becomes totally useless, and the PK’ers barely have to exert themselves. The team was in he bottom 2 (along with Arizona) PP during the first half of last year.

It began to improve once Backstrom came back to control the half-wall. And then once Oshie came back for good (March, I believe), it was top 3-ish from that point on.

Unfortunately, it’s an ongoing problem that our PP is predicated on 5 all-star caliber players each playing specialized roles, and no other looks drawn up by Forsythe ever work. This place was flooded with people brainstorming different looks during that awful stretch last year.

But this team really needs its PP working (it was key to success in the FLA series), as it is the area where Ovi is most impactful, and the age of the vets is mitigated by their chemistry, and w/ lesser importance placed on speed.

I think it's a damn shame that the team went through a prolonged stretch of absolute awful power play performance last season and at no point did they even attempt a different approach. Simply waited for Backstrom/Oshie to come back.

It was dumb when it happened and it looks even dumber now, with the team going into the season with Backstrom sidelined long term, possibly for good, and Oshie injured.

 

[twabby]

I hadn’t really investigated the RAPM numbers on Evolving Hockey because they are incorporated into their WAR stats, but some of the numbers are pretty damning.

Evolving-Hockey.com | RAPM Skater Tables

Skater Regularized Adjusted Plus-Minus metrics for the NHL

evolving-hockey.com

McMichael is 46th out of 380 qualified forwards in terms of RAPM xGF/60, best among all Capitals forwards. This stat basically measures a player’s isolated impact in generating expected goals while controlling for teammates, opponents, score, etc.

The next best Capital forward is Connor Brown, ranked 119th. Only McMichael, Brown, Ovechkin, Hathaway, Strome, and Kuznetsov are even above average in this stat. And only McMichael is in the top 100. When it comes to forwards generating chances they have to be worst or close to it among playoff teams from last year.

For a team that has struggled so much generating chances in the postseason it would be foolish to deny McMichael substantial minutes this year, especially when he has been defensively stout as well (54th of 380 in RAPM xGA/60). They need players who can generate chances reliably in order to make any noise in the postseason. They can’t just rely on good shooting and opportunism like they do in the regular season. Defenses tighten up in the postseason and don’t make as many brainfarts. More goals have to be earned.

They should be playing McMichael with legit finishing talent who may struggle a bit generating their own chances. Guys like Oshie and Wilson fit this bill in the middle 6. Heck, they should have him center Ovechkin and see how that works out. They need to give him a chance to grow beside other skilled players. Typically forwards of his age take a significant step forward when they get ice time. They need to make sure he gets this ice time to build on an already impressive rookie season so that he can potentially be an impact top 6 player this season and postseason. We don’t need this potential blocked by low-ceiling players like Eller and Johansson (or Hagelin if he can return).

 

[Vilica]

I remembered that I went back through the year and worked out which players missed which games, and the power play's results with/without them. Ovechkin missed 5, Kuznetsov 3, and Carlson 4 games, which are too small samples to generate any real results - the Caps were 1/21 when Ovechkin was out, 2/10 when Carlson was out, and 1/11 when Kuznetsov was out. However, Oshie played 54, missed 38, and Backstrom played 47, missed 35, and those are big enough samples.

With Oshie, the Caps were 28/143, or 19.5% effective, without Oshie they were 20/113, 17.7% effective.
With Backstrom, the Caps were 33/153, 21.5% effective, without Backstrom they were 15/103, 14.5% effective.

To be sure, Oshie's a superior bumper player, but Nick's the straw. He's missed 2 significant chunks of time in his career, the 11-12 season and last season, and the Caps power play immediately tanked to bottom 2-3 in the league in that missed time. When he's in the lineup, they're #1.

 

[traparatus]

I remembered that I went back through the year and worked out which players missed which games, and the power play's results with/without them. Ovechkin missed 5, Kuznetsov 3, and Carlson 4 games, which are too small samples to generate any real results - the Caps were 1/21 when Ovechkin was out, 2/10 when Carlson was out, and 1/11 when Kuznetsov was out. However, Oshie played 54, missed 38, and Backstrom played 47, missed 35, and those are big enough samples.

With Oshie, the Caps were 28/143, or 19.5% effective, without Oshie they were 20/113, 17.7% effective.
With Backstrom, the Caps were 33/153, 21.5% effective, without Backstrom they were 15/103, 14.5% effective.

To be sure, Oshie's a superior bumper player, but Nick's the straw. He's missed 2 significant chunks of time in his career, the 11-12 season and last season, and the Caps power play immediately tanked to bottom 2-3 in the league in that missed time. When he's in the lineup, they're #1.

Nice bit of research.

Make sense, right? Ovechkin doesn't want to handle the puck at all and that leaves whoever is on the right boards to do vast majority of facilitating. I feel like that's a very difficult position to put a player in and it simply takes a player of Backstrom's talent to do it effectively.

As for Oshie, Caps just don't have anyone else capable of playing that role. Wilson gives it a go but he isn't particularly good at it.

 

[trick9]

I just want him healthy for the playoffs. He can rest durning the regular season

I don't see us making the Playoffs if we have Wilson and Oshie out long-term on top of not having Backstrom. Even if they somehow manage to squeeze in with all those injuries then how much do you think the other top forwards have left in the tank?

 

[Langway]

The next best Capital forward is Connor Brown, ranked 119th. Only McMichael, Brown, Ovechkin, Hathaway, Strome, and Kuznetsov are even above average in this stat. And only McMichael is in the top 100. When it comes to forwards generating chances they have to be worst or close to it among playoff teams from last year.

Mantha was also below average in this area then? Seems like a significant structural issue then having both him and Oshie below average.

 

[twabby]

Mantha was also below average in this area then? Seems like a significant structural issue then having both him and Oshie below average.

Good catch. Mantha didn’t qualify based on TOI when I filtered originally (I used a minimum of 500 minutes. He clocked in at 481 last year).

Using 480 as the minimum Mantha was 70th of 383. So two players in the top 100.

Still not good for a playoff team.

 

[Langway]

Wilson also below average? Certainly they need a better mix and part of it has be a more ambitious approach than low impact reliability that comes at the expense of a higher ceiling. Johansson vs. McMichael may end up being the litmus test. For now perhaps Oshie being out clears the way initially and maybe with that opportunity Protas & McMichael both manage to upend their usual tendencies.

 

[twabby]

Wilson also below average? Certainly they need a better mix and part of it has be a more ambitious approach than low impact reliability that comes at the expense of a higher ceiling. Johansson vs. McMichael may end up being the litmus test. For now perhaps Oshie being out clears the way initially and maybe with that opportunity Protas & McMichael both manage to upend their usual tendencies.

Wilson is actually well below average in this metric and has been for the past two seasons.

I'll just go ahead and copy the table with their ranks among NHL forwards (minimum 350 minutes TOI, so as to also include Aliaksei Protas):

Very few Capitals forwards are near the top of the league in terms of chance generation. Lots of players in the middle six are well below average (Oshie, Wilson, Sheary, Eller, Backstrom).

They have to get more chance generation from their top 9 in order to have any chance of winning a few rounds in the postseason. I don't need to keep bringing up how Washington is the worst goal scoring team at 5v5 in the postseason since 2018. Connor Brown should help a bit. Mantha should help if he can stay healthy.

But they cannot deny McMichael an opportunity entering the season, especially if it's in favor of middling players like Eller and Johansson who have no realistic shot to be impactful options in the top 9. McMichael earned a shot based on his play last season, and the team will be better off letting him grow throughout the season to build on his impressive rookie campaign rather than benching him for every little mistake he makes.

 

[Langway]

I think this gets at the heart of their offensive approach generally. It remains stuck in the past largely revolving around being carried by a core group no longer dominant with not much happening internally developing secondary weapons. Hence, McMichael getting somewhat lost in the shuffle as it's not an active priority. I'm hopeful Strome+Brown can change the dynamic a bit but very likely they need more.

About the only universal active systematic approach is activating D and interchanging between D/F which, again, brings me back to the additional lack of talent on the back-end to compensate. Carlson & Orlov can bring it but Jensen & Gustafsson aren't enough otherwise. I'm generally not a believer Fehervary brings all that much production offensively, even if metrics it. I don't consider it a strong point and I'm not sure there's much further upside (like ever becoming a 30+ point D). Similarly, you don't really want TVR shouldering offensive chances either. So it's a bit of a losing hand tactically relying on defensive pressure and effort as the difference. They need to raise the standard offensively and become more selective/active in their approach and execution or else not much will change. Unfortunately it seems much more likely they'll stick with effort and defense as the basic standard. It's on MacLellan to see to it, either based on shaping the agenda or making needed personnel moves, that the status quo is disrupted and forced into evolving further.

 

[AlexModvechkin8]

Dom’s updated model from The Athletic has the Caps coming in third in the Metro with 97 points behind Carolina and Pittsburgh. He has them at 12th in the NHL and gives them a 70% chance of making the playoffs.

I’ve actually yet to see a model that has the Caps finishing below 4th in the Metro and out of the playoffs. I think the models are really high on the goaltending upgrade and improved depth.

 

[AlexModvechkin8]

Looks like worst case scenario was avoided but we’ll see. Still, good to see him out there… I was nervous it was going to be another Mantha situation where the next morning we wake up to a report that he’s had surgery and is out indefinitely.

 

[g00n]

Wilson is actually well below average in this metric and has been for the past two seasons.

I'll just go ahead and copy the table with their ranks among NHL forwards (minimum 350 minutes TOI, so as to also include Aliaksei Protas):

View attachment 591409

Very few Capitals forwards are near the top of the league in terms of chance generation. Lots of players in the middle six are well below average (Oshie, Wilson, Sheary, Eller, Backstrom).

They have to get more chance generation from their top 9 in order to have any chance of winning a few rounds in the postseason. I don't need to keep bringing up how Washington is the worst goal scoring team at 5v5 in the postseason since 2018. Connor Brown should help a bit. Mantha should help if he can stay healthy.

But they cannot deny McMichael an opportunity entering the season, especially if it's in favor of middling players like Eller and Johansson who have no realistic shot to be impactful options in the top 9. McMichael earned a shot based on his play last season, and the team will be better off letting him grow throughout the season to build on his impressive rookie campaign rather than benching him for every little mistake he makes.

Some real cherrypicking going on here.

This is again a stat about imaginary goals, and RAPM apparently factors GF while on ice but not GA, because of goaltender variations? How does that work ...if a shitty goaltender gives up a goal you benefit, but if your goaltender gives up a shitty goal there's no penalty. Seems like it's tilted toward offense-only players by default.

Anyway.

 

[LesDiablesRouges]

Some of you just live and die by fancy stats and analytics. It’s rather hysterical. Over half the other part of the sport is actually hockey knowledge and the actual eye test.

 

[RedRocking]

I remembered that I went back through the year and worked out which players missed which games, and the power play's results with/without them. Ovechkin missed 5, Kuznetsov 3, and Carlson 4 games, which are too small samples to generate any real results - the Caps were 1/21 when Ovechkin was out, 2/10 when Carlson was out, and 1/11 when Kuznetsov was out. However, Oshie played 54, missed 38, and Backstrom played 47, missed 35, and those are big enough samples.

With Oshie, the Caps were 28/143, or 19.5% effective, without Oshie they were 20/113, 17.7% effective.
With Backstrom, the Caps were 33/153, 21.5% effective, without Backstrom they were 15/103, 14.5% effective.

To be sure, Oshie's a superior bumper player, but Nick's the straw. He's missed 2 significant chunks of time in his career, the 11-12 season and last season, and the Caps power play immediately tanked to bottom 2-3 in the league in that missed time. When he's in the lineup, they're #1.

Great info, that’s really interesting.

It’s asking a lot of Strome to step into Backy’s spot - hope he’s up to the challenge. For all the frustration at times last year re Nick’s 5v5 play, he still calms down and QB’s the PP like no other. Kuzy has shown he doesn’t have the same vision and patience (though few players do).

Strome has shown a willingness to shoot more from that spot though, so it will be interesting if he can add that dynamic. Would relieve some pressure on the bumper guy, and give Carlson some clearer lanes to shoot.

I know a lot of us would like to see Mantha tried in that spot on PP1 since he can really wire it.

 

[Roshi]

Of course his growing pains severely impacted the team, so if this was the design then perhaps the design was flawed.



I guess I just don't care about stylistic elements like skating, physicality, size, and strength if they don't translate to impacting the scoreboard. Fehervary certainly has elements that make you think he's a special player, but then you look at the numbers and realize how much of a drain he really was for the majority of last season.

Similarly McMichael certainly isn't a physically imposing player, but then you look at the numbers and see that he acquitted himself very well for a 20-21 year old rookie. I don't think that adding a few pounds of muscle and winning a few more board battles is his ticket to improvement.

I know im late for the party, but this is pretty much an overlooked part of advanced stats.

Two young players. One plays heavy minutes in first pairing and looks pretty good by the eye-test, while clocking suspect adv stats. One plays small minutes and doesnt look impactful, but puts in the possession metrics.

Doesnt this also kind of prove that they are using CMM the correct way. Shelter him so he can have success with the metrics. While giving him the chance to develope in NHL. And that maybe Fever would benefit for getting easier load, so he can have time aswell.

Why is the expectation that CMM is better with more ice time? Id figure it might have a net negative effect on his possession meters, even if he would score a goal or couple more.

We dont have the need to rush with CMM, but we did have the need to rush Fever.

 

[twabby]

Dom’s updated model from The Athletic has the Caps coming in third in the Metro with 97 points behind Carolina and Pittsburgh. He has them at 12th in the NHL and gives them a 70% chance of making the playoffs.

I’ve actually yet to see a model that has the Caps finishing below 4th in the Metro and out of the playoffs. I think the models are really high on the goaltending upgrade and improved depth.

That sounds like a fair projection.

Washington has a fairly high floor given their goaltending, defense, and generally above average depth so it would kind of surprise me if they miss this year. They’d need a lot of things to go wrong.

Of course, the problem is they also have a lower ceiling than most contenders because their superstars don’t match up with the superstars on other contenders.

 

[twabby]

Some real cherrypicking going on here.

This is again a stat about imaginary goals, and RAPM apparently factors GF while on ice but not GA, because of goaltender variations? How does that work ...if a shitty goaltender gives up a goal you benefit, but if your goaltender gives up a shitty goal there's no penalty. Seems like it's tilted toward offense-only players by default.

Anyway.

The rationale is that some players can consistently outscore their xG due to superior shooting talent. Ovechkin is a perfect example of this, unless one thinks he’s simply getting incredibly lucky year after year.

The same disparity is generally not seen between xGA and GA over long periods of time when you control for goaltending. Generally these impacts tend toward each other (again, controlling for goaltending.) Thus it doesn’t really make as much sense to look at a player’s GA impact vs. xGA since xGA converges much more quickly to a player’s true defensive talent.

I think there is some flaw in that logic because you’d think defenses should know that Ovechkin can score from distance like few others, but generally looking at xGA impact is a better defensive predictor than looking at GA.

If Connor McMichael ends up being a Carl Hagelin type shooter then yes, the xG models will severely overrate him. However it should be noted that the xG model is an on-ice metric, not an individual one, so he generally increases his team’s ability to generate chances and not just his own. This is why I think he should be playing alongside more skilled shooters, e.g. Tom Wilson, TJ Oshie, and yes even Alex Ovechkin. Even if McMichael continues struggling to score himself he could still put these players in better positions to score and they certainly know how to bury pucks if nothing else.

 

[Hivemind]

Some of you just live and die by fancy stats and analytics. It’s rather hysterical. Over half the other part of the sport is actually hockey knowledge and the actual eye test.

Good thing McMichael has passed the eye test, as well.

 

[g00n]

The rationale is that some players can consistently outscore their xG due to superior shooting talent. Ovechkin is a perfect example of this, unless one thinks he’s simply getting incredibly lucky year after year.

The same disparity is generally not seen between xGA and GA over long periods of time when you control for goaltending. Generally these impacts tend toward each other (again, controlling for goaltending.) Thus it doesn’t really make as much sense to look at a player’s GA impact vs. xGA since xGA converges much more quickly to a player’s true defensive talent.

Who judges this "talent" outside the statistical models, and how?

 

[twabby]

Who judges this "talent" outside the statistical models, and how?

The community as a whole provides input. For instance no one worth their salt looks solely at xGF/60 as an overall measure of a player’s offensive worth because it ignores a lot, such as shooting talent. No one is claiming that Connor McMichael
Is the team’s best offensive forward.

I don’t know exactly what you’re asking. Models are improved by making hypotheses and testing them against future observations, just like in any other discipline.

 

[g00n]

The community as a whole provides input. For instance no one worth their salt looks solely at xGF/60 as an overall measure of a player’s offensive worth because it ignores a lot, such as shooting talent. No one is claiming that Connor McMichael
Is the team’s best offensive forward.

I don’t know exactly what you’re asking. Models are improved by making hypotheses and testing them against future observations, just like in any other discipline.

So visual observation of how the player executes actions, and not just simple measurements of quantifiable metrics, are used to modify the statistics themselves?

 

[AlexModvechkin8]

That sounds like a fair projection.

Washington has a fairly high floor given their goaltending, defense, and generally above average depth so it would kind of surprise me if they miss this year. They’d need a lot of things to go wrong.

Of course, the problem is they also have a lower ceiling than most contenders because their superstars don’t match up with the superstars on other contenders.

I don’t disagree that they could use more dynamic playmakers up front but we’ve also seen situations where a team has stars and those stars go cold or get injured and there’s no depth to pick them up. Best on best this team probably falls short of some of the truly elite teams but they’re deeper and more balanced than some of the teams with great top end talent and replacement level players in the bottom six. I’d love to see them get a Timo Meier type to close the gap on top end talent.