Speculation: Caps General Discussion (Coaching/FAs/Cap/Lines etc) - 2021 "Season" Pt. 3

Status
Not open for further replies.

Ridley Simon

Registered User
Feb 27, 2002
18,125
9,066
Marin County — SF Bay Area, CA
To be honest, I am not sure we would have won another Cup with Trotz, either. Seemed like the stars aligned that post-season and Reirden took over some huge responsibilities and Kuzy went god mode. I'm not sure we would've replicated that run in any other year after. It was the perfect storm, if you will.

Edit: Many forget Trotz was on the hot seat much of that run and were even calling for his head and for him to be fired.
Hell, it wasn’t even a few calling for his head. There was a poll and like 90% of the posters here wanted him canned. It was a mob grabbing the pitchforks and torches and storming the castle.

@EroCaps — where were you in that poll? I’d check but I’m lazy.
 
Last edited:

kicksavedave

I'm just here for the memes and gifs.
Sponsor
Apr 29, 2009
10,606
13,247
Fallbrook, CA
www.tiasarms.org
No, Peter Hassett simply doesn't know how to read a table, or do simple division. From Natural Stat Trick (Player Season Totals - Natural Stat Trick):

PlayerCAHDCA% HDCA
Alex Ovechkin3736517.43%
Nicklas Backstrom3796015.83%
T.J. Oshie3045116.78%
Richard Panik2454317.55%
Jakub Vrana2783914.03%
Conor Sheary2364519.07%
Lars Eller2203616.36%
Nic Dowd3004816.00%
Garnet Hathaway2754516.36%
Carl Hagelin2714115.13%
Evgeny Kuznetsov1943317.01%
Tom Wilson2693814.13%
Daniel Sprong1402417.14%
[TBODY] [/TBODY]
So it looks like Kuznetsov allows more like 1 in 6 shots to be high danger. Same with Backstrom. Same with basically the entire team. I have no idea where he is getting 1 in 4 and 1 in 9, respectively.

But not only that, even if he didn't mess up the numbers I think Peter Hassett made a huge mistake interpreting the numbers. Focusing on the proportion of shots that are high danger rather than the absolute number that are high-danger is uh highly misguided!

For instance, suppose Player A gave up 4 shots every 60 minutes, but 1 of them was high danger. Suppose Player B gave up 90 shots every 60 minutes, but 10 of them were high danger. Ignoring other context, who would you assume is doing better at suppressing chances?

Peter is saying that Player B is better, because only 1 in 9 are high danger shots vs. Player A who is allowing 1 in 4 shots against to be high danger.

Of course, this is nonsense.

On a per 60 basis, Backstrom gives up more high danger chances than Kuznetsov this season. He gives up more scoring chances per 60. He gives up more shots per 60. He gives up more goals per 60 (actual goals, the goals that matter to everyone). He gives up more expected goals per 60.

Does that mean that Kuznetsov is a better defensive player? No, not on its own. But it certainly doesn't show that he is a weaker defensive player either. A deeper analysis is needed to tackle that question, which he doesn't do.

Furthermore, he mentions the "cost" of sheltering Kuznetsov is that Dowd and the 4th line are putting up bad numbers in an absolute sense. He, again, I believe is very mistaken.

Dowd is currently sporting a 48% xGF%. Not great, but better than Backstrom, Ovechkin, and Wilson at this point. Here is Dowd's picture:

View attachment 408053

As mentioned, Dowd gets heavy defensive zone usage, only starting 16% in the offensive zone.

And here is Dowd's competition:

dowdxni90


So to summarize Nic Dowd: he is facing very difficult competition in heavy D zone usage and is almost breaking even with expected goals. I think Dowd is actually doing a fantastic job in his 4th line role.

Peter's note about Dowd and Hathaway taking penalties is duly noted, but they've taken penalties their entire career and this year isn't really out of whack. In fact, Hathaway has taken fewer minor penalties on a per 60 basis this year than last year. Same with Hagelin. Dowd's is up, but it very well could be incidental.

I also think Peter has cause and effect mixed up. He thinks Dowd is getting such difficult matchups because Kuznetsov isn't being trusted with them. I see quite the reverse: I think Kuznetsov (and Backstrom, and Eller) are getting easier matchups because Dowd is being trusted with such difficult matchups. And with Dowd's results absorbing this tough competition, the strategy seems like a sound one. It's allowing the top 9 to be in a better position to score goals. Indeed, the team is the best in the NHL at 5v5 goals per 60.

I mentioned this in an earlier post, but Laviolette's usage of the 4th line reminds me a lot of how the Blackhawks used their 4th line in 2012-13, but especially in 2014-15. Marcus Kruger got 24% offensive zone starts and faced difficult competition, allowing more offensively gifted players like Jonathan Toews (60%), Patrick Sharp (72%), and Patrick Kane (78%) to feast on easier deployments. I don't need to remind you who won the Stanley Cup in 2013 and 2015 either.

TLDR: Peter Hassett can't read numbers from a chart or can't do division or both. And even if he could, he's interpreting the numbers in a really bad way. And I think he's confusing cause and effect with Dowd and Kuznetsov, especially since Backstrom, Oshie, Wilson, and Eller are also getting easy zone starts and no one would claim they are huge defensive liabilities. All of the top 9 is benefitting from Dowd getting those hard matchups and it seems to me to be a strategy that can win a Stanley Cup as evidence by it being a strategy that was employed on at least 2 Stanley Cup winners in the past decade.

Great, I am firmly back on the Kuzy bandwagon.
 

ALLCAPSALLTHETIME

Great Dane! Love that Eller feller.
Oct 10, 2009
9,234
4,898
British Columbia, Canada
No, Peter Hassett simply doesn't know how to read a table, or do simple division. From Natural Stat Trick (Player Season Totals - Natural Stat Trick):

PlayerCAHDCA% HDCA
Alex Ovechkin3736517.43%
Nicklas Backstrom3796015.83%
T.J. Oshie3045116.78%
Richard Panik2454317.55%
Jakub Vrana2783914.03%
Conor Sheary2364519.07%
Lars Eller2203616.36%
Nic Dowd3004816.00%
Garnet Hathaway2754516.36%
Carl Hagelin2714115.13%
Evgeny Kuznetsov1943317.01%
Tom Wilson2693814.13%
Daniel Sprong1402417.14%
[TBODY] [/TBODY]
So it looks like Kuznetsov allows more like 1 in 6 shots to be high danger. Same with Backstrom. Same with basically the entire team. I have no idea where he is getting 1 in 4 and 1 in 9, respectively.

But not only that, even if he didn't mess up the numbers I think Peter Hassett made a huge mistake interpreting the numbers. Focusing on the proportion of shots that are high danger rather than the absolute number that are high-danger is uh highly misguided!

For instance, suppose Player A gave up 4 shots every 60 minutes, but 1 of them was high danger. Suppose Player B gave up 90 shots every 60 minutes, but 10 of them were high danger. Ignoring other context, who would you assume is doing better at suppressing chances?

Peter is saying that Player B is better, because only 1 in 9 are high danger shots vs. Player A who is allowing 1 in 4 shots against to be high danger.

Of course, this is nonsense.

On a per 60 basis, Backstrom gives up more high danger chances than Kuznetsov this season. He gives up more scoring chances per 60. He gives up more shots per 60. He gives up more goals per 60 (actual goals, the goals that matter to everyone). He gives up more expected goals per 60.

Does that mean that Kuznetsov is a better defensive player? No, not on its own. But it certainly doesn't show that he is a weaker defensive player either. A deeper analysis is needed to tackle that question, which he doesn't do.

Furthermore, he mentions the "cost" of sheltering Kuznetsov is that Dowd and the 4th line are putting up bad numbers in an absolute sense. He, again, I believe is very mistaken.

Dowd is currently sporting a 48% xGF%. Not great, but better than Backstrom, Ovechkin, and Wilson at this point. Here is Dowd's picture:

View attachment 408053

As mentioned, Dowd gets heavy defensive zone usage, only starting 16% in the offensive zone.

And here is Dowd's competition:

dowdxni90


So to summarize Nic Dowd: he is facing very difficult competition in heavy D zone usage and is almost breaking even with expected goals. I think Dowd is actually doing a fantastic job in his 4th line role.

Peter's note about Dowd and Hathaway taking penalties is duly noted, but they've taken penalties their entire career and this year isn't really out of whack. In fact, Hathaway has taken fewer minor penalties on a per 60 basis this year than last year. Same with Hagelin. Dowd's is up, but it very well could be incidental.

I also think Peter has cause and effect mixed up. He thinks Dowd is getting such difficult matchups because Kuznetsov isn't being trusted with them. I see quite the reverse: I think Kuznetsov (and Backstrom, and Eller) are getting easier matchups because Dowd is being trusted with such difficult matchups. And with Dowd's results absorbing this tough competition, the strategy seems like a sound one. It's allowing the top 9 to be in a better position to score goals. Indeed, the team is the best in the NHL at 5v5 goals per 60.

I mentioned this in an earlier post, but Laviolette's usage of the 4th line reminds me a lot of how the Blackhawks used their 4th line in 2012-13, but especially in 2014-15. Marcus Kruger got 24% offensive zone starts and faced difficult competition, allowing more offensively gifted players like Jonathan Toews (60%), Patrick Sharp (72%), and Patrick Kane (78%) to feast on easier deployments. I don't need to remind you who won the Stanley Cup in 2013 and 2015 either.

TLDR: Peter Hassett can't read numbers from a chart or can't do division or both. And even if he could, he's interpreting the numbers in a really bad way. And I think he's confusing cause and effect with Dowd and Kuznetsov, especially since Backstrom, Oshie, Wilson, and Eller are also getting easy zone starts and no one would claim they are huge defensive liabilities. All of the top 9 is benefitting from Dowd getting those hard matchups and it seems to me to be a strategy that can win a Stanley Cup as evidence by it being a strategy that was employed on at least 2 Stanley Cup winners in the past decade.

Thanks, Twabby. That was very thorough and well written. :thumbu:
 

Roshi

Registered User
Feb 7, 2013
1,986
1,949
Finland
Great job Twabby.

Im normally whining big about advanced statology and fuming when someone on main boards just links an data sheet and calls it a day, but this is how insightful thinking process with the help of statistical data really has advantages, and it actually backs up the eye test.
 

g00n

Retired Global Mod
Nov 22, 2007
30,430
14,290
No, Peter Hassett simply doesn't know how to read a table, or do simple division. From Natural Stat Trick (Player Season Totals - Natural Stat Trick):

PlayerCAHDCA% HDCA
Alex Ovechkin3736517.43%
Nicklas Backstrom3796015.83%
T.J. Oshie3045116.78%
Richard Panik2454317.55%
Jakub Vrana2783914.03%
Conor Sheary2364519.07%
Lars Eller2203616.36%
Nic Dowd3004816.00%
Garnet Hathaway2754516.36%
Carl Hagelin2714115.13%
Evgeny Kuznetsov1943317.01%
Tom Wilson2693814.13%
Daniel Sprong1402417.14%
[TBODY] [/TBODY]
So it looks like Kuznetsov allows more like 1 in 6 shots to be high danger. Same with Backstrom. Same with basically the entire team. I have no idea where he is getting 1 in 4 and 1 in 9, respectively.

But not only that, even if he didn't mess up the numbers I think Peter Hassett made a huge mistake interpreting the numbers. Focusing on the proportion of shots that are high danger rather than the absolute number that are high-danger is uh highly misguided!

For instance, suppose Player A gave up 4 shots every 60 minutes, but 1 of them was high danger. Suppose Player B gave up 90 shots every 60 minutes, but 10 of them were high danger. Ignoring other context, who would you assume is doing better at suppressing chances?

Peter is saying that Player B is better, because only 1 in 9 are high danger shots vs. Player A who is allowing 1 in 4 shots against to be high danger.

Of course, this is nonsense.

On a per 60 basis, Backstrom gives up more high danger chances than Kuznetsov this season. He gives up more scoring chances per 60. He gives up more shots per 60. He gives up more goals per 60 (actual goals, the goals that matter to everyone). He gives up more expected goals per 60.

Does that mean that Kuznetsov is a better defensive player? No, not on its own. But it certainly doesn't show that he is a weaker defensive player either. A deeper analysis is needed to tackle that question, which he doesn't do.

Furthermore, he mentions the "cost" of sheltering Kuznetsov is that Dowd and the 4th line are putting up bad numbers in an absolute sense. He, again, I believe is very mistaken.

Dowd is currently sporting a 48% xGF%. Not great, but better than Backstrom, Ovechkin, and Wilson at this point. Here is Dowd's picture:

View attachment 408053

As mentioned, Dowd gets heavy defensive zone usage, only starting 16% in the offensive zone.

And here is Dowd's competition:

dowdxni90


So to summarize Nic Dowd: he is facing very difficult competition in heavy D zone usage and is almost breaking even with expected goals. I think Dowd is actually doing a fantastic job in his role. Anyone clamoring for Jay Beagle to come back and replace him is on some sort of substance.

Peter's note about Dowd and Hathaway taking penalties is duly noted, but they've taken penalties their entire career and this year isn't really out of whack. In fact, Hathaway has taken fewer minor penalties on a per 60 basis this year than last year. Same with Hagelin. Dowd's is up, but it very well could be incidental.

I also think Peter has cause and effect mixed up. He thinks Dowd is getting such difficult matchups because Kuznetsov isn't being trusted with them. I see quite the reverse: I think Kuznetsov (and Backstrom, and Eller) are getting easier matchups because Dowd is being trusted with such difficult matchups. And with Dowd's results absorbing this tough competition, the strategy seems like a sound one. It's allowing the top 9 to be in a better position to score goals. Indeed, the team is the best in the NHL at 5v5 goals per 60.

I mentioned this in an earlier post, but Laviolette's usage of the 4th line reminds me a lot of how the Blackhawks used their 4th line in 2012-13, but especially in 2014-15. Marcus Kruger got 24% offensive zone starts and faced difficult competition, allowing more offensively gifted players like Jonathan Toews (60%), Patrick Sharp (72%), and Patrick Kane (78%) to feast on easier deployments. I don't need to remind you who won the Stanley Cup in 2013 and 2015 either.

TLDR: Peter Hassett can't read numbers from a chart or can't do division or both. And even if he could, he's interpreting the numbers in a really bad way. And I think he's confusing cause and effect with Dowd and Kuznetsov, especially since Backstrom, Oshie, Wilson, and Eller are also getting easy zone starts and no one would claim they are huge defensive liabilities. All of the top 9 is benefitting from Dowd getting those hard matchups and it seems to me to be a strategy that can win a Stanley Cup as evidenced by it being a strategy that was employed on at least 2 Stanley Cup winners in the past decade.


Jumping into the middle of this convo so sure to have missed some context, but if Kuzy is getting like 80% O-zone starts and Backstrom is getting 65%, wouldn't you EXPECT a much higher volume of high-danger scoring chances, and shots/chances in general, against Backstrom?

Why is that noted for Dowd but not Backstrom when you're trying to dismiss the HDCA ratio? Or did I miss it?
 

twabby

Registered User
Mar 9, 2010
13,624
14,441
I generally like both RMNB and hockey analytics, but RMNB bungles analytics articles as often as they find something insightful. Japer's tends to do far better.

Pat Holden used to be decent at them, much better than Peter Hassett at least. But I don't think Holden writes for RMNB any more.

Jon Press at JapersRink is usually pretty good.
 

twabby

Registered User
Mar 9, 2010
13,624
14,441
Jumping into the middle of this convo so sure to have missed some context, but if Kuzy is getting like 80% O-zone starts and Backstrom is getting 65%, wouldn't you EXPECT a much higher volume of high-danger scoring chances, and shots/chances in general, against Backstrom?

Why is that noted for Dowd but not Backstrom when you're trying to dismiss the HDCA ratio? Or did I miss it?

Yes, if defensive talent was equal you would expect Kuznetsov's numbers to be better than Backstrom's due to his easier deployment (though it's really not that much easier. Both have faced middling competition, Backstrom's slightly tougher, and Backstrom has 72% o-zone starts, Kuznetsov 89%). How much better Kuznetsov's defensive numbers should be is a more difficult question, and one Hassett doesn't even attempt to answer.

The HDCA ratio is largely irrelevant because:

1. It ignores medium and low danger shots against. Though they are lower danger shots, they still have a chance of going in. Instead of ignoring them, they should be properly weighted. That's what expected goals does.

2. It ignores volume. The HDCA ratio doesn't address how many dangerous chances are allowed, just the proportion of the overall shots that are high danger.

Again, the example of Player A allowing only 4 shots per 60, 1 of which being high danger vs. Player B allowing 90 shots total per 60, 10 of which are high danger should be a simple illustration of why this is a bad way to evaluate defensive play.

The point of this isn't to say Backstrom is bad defensively, or worse than Kuznetsov defensively. I don't believe either of those statements to be true. The point of this is to show that Hassett's conclusion that Kuznetsov has done poorly defensively this year is not supported by any of his reasoning, both because he bungles simple facts and because his rationale based on the numbers he was trying to use is faulty.
 

Devil Dancer

Registered User
Jan 21, 2006
18,439
5,406
Yes, if defensive talent was equal you would expect Kuznetsov's numbers to be better than Backstrom's due to his easier deployment (though it's really not that much easier. Both have faced middling competition, Backstrom's slightly tougher, and Backstrom has 72% o-zone starts, Kuznetsov 89%). How much better Kuznetsov's defensive numbers should be is a more difficult question, and one Hassett doesn't even attempt to answer.

The HDCA ratio is largely irrelevant because:

1. It ignores medium and low danger shots against. Though they are lower danger shots, they still have a chance of going in. Instead of ignoring them, they should be properly weighted. That's what expected goals does.

2. It ignores volume. The HDCA ratio doesn't address how many dangerous chances are allowed, just the proportion of the overall shots that are high danger.

Again, the example of Player A allowing only 4 shots per 60, 1 of which being high danger vs. Player B allowing 90 shots total per 60, 10 of which are high danger should be a simple illustration of why this is a bad way to evaluate defensive play.

The point of this isn't to say Backstrom is bad defensively, or worse than Kuznetsov defensively. I don't believe either of those statements to be true. The point of this is to show that Hassett's conclusion that Kuznetsov has done poorly defensively this year is not supported by any of his reasoning, both because he bungles simple facts and because his rationale based on the numbers he was trying to use is faulty.
Ok, ok, I'm sold. Post a link to twabbysrink or whatever you call your inevitable blog and I promise I'll read it.
 

AlexBrovechkin8

At least there was 2018.
Sponsor
Feb 18, 2012
26,744
24,941
District of Champions
RMNB is great at human interest pieces and making the game more fun and accessible for fans (especially new or very casual fans) but they're pretty brutal at the analytics side of the house. They have their narratives and they will jam round-peg numbers into that square-hole thesis all else be damned in order to make the point they want to make. I don't think there's a ton of actual hockey experience be it playing or coaching so they lack context which is why some of their strong opinions on players or situations are head-shakers.
 

Ovechkins Wodka

Registered User
Dec 1, 2007
17,443
7,148
DC
RMNB is great at human interest pieces and making the game more fun and accessible for fans (especially new or very casual fans) but they're pretty brutal at the analytics side of the house. They have their narratives and they will jam round-peg numbers into that square-hole thesis all else be damned in order to make the point they want to make. I don't think there's a ton of actual hockey experience be it playing or coaching so they lack context which is why some of their strong opinions on players or situations are head-shakers.
Can they stop re posting everything from Ovies wife about his son. Its hitting a creepy stalker level.
 
Last edited:

tenken00

Oh it's going down in Chinatown
Jan 29, 2010
9,863
10,089
We have Sprong under team control next year too if we want him since he’s an RFA until 2022, I think. Probably safe to assume he’s not going to continue shooting 30% like he has been this year but he’s looking like a solid depth signing.

Sprong has 3 goals 4 points in the 5 games he has filled in for Wilson. Talk about nailing the audition.
 

twabby

Registered User
Mar 9, 2010
13,624
14,441
In an attempt to try to figure out how to judge Kuznetsov's xGF% based on zone start ratio, I made a scatterplot based on data from Evolving Hockey:

upload_2021-3-16_12-31-4.png


This is every player-season from 2014-15 until now, with a minimum of 250 minutes played (since Kuznetsov is just above 250 minutes this year at 5v5, that is why I chose the cutoff at 250).

The x-axis is the zone start ratio. Players further to the left get heavy defensive zone starts, and players on the right get a lot of offensive zone starts.

The y-axis is the xGF%. Higher is better, 50% is average.

Circled in red is Kuznetsov from this season. He is certainly getting a ton of o-zone starts, but based on the rest of the plot doesn't it seem like despite getting these soft starts he's doing well with them? At the very least it seems difficult to make the case that he's doing poorly with these starts.
 
Last edited:

g00n

Retired Global Mod
Nov 22, 2007
30,430
14,290
Perhaps there should be a thread for those who refuse to read analytics, rather than attempting to stifle legitimate roster discussion? The stat haters can have their own safe space.

I work with numbers every day. Big numbers. It's not about hating stats it's about knowing the place for them, and respecting that most people are here to talk hockey not take a quiz on terms/methods that a much smaller number of posters are even familiar with.

Plus, there is still much debate about how accurate or useful many of these stats and data points are. That's why we created a thread for it. Otherwise every discussion turns into 2-3 people posting graphs and arguing with everyone else about how meaningful they are.
 

Hivemind

We're Touched
Oct 8, 2010
37,009
13,425
Philadelphia
Circled in red is Kuznetsov from this season. He is certainly getting a ton of o-zone starts, but based on the rest of the plot doesn't it seem like despite getting these soft starts he's doing well with them? At the very least it seems difficult to make the case that he's doing poorly with these starts.
Given how much of an outlier his deployment is, we don’t really have much context on how well he should be doing. It’s in Quenville-levels of deployment shifting.
 

twabby

Registered User
Mar 9, 2010
13,624
14,441
Perhaps there should be a thread for those who refuse to read analytics, rather than attempting to stifle legitimate roster discussion? The stat haters can have their own safe space.

I agree with this. It's easy enough to scroll past a post you don't care to read. I realize that as a moderator @Calicaps probably can't put me on ignore, but all non-moderators are free to do so if they want as well.

It seems like it's more a way to stifle discussion than anything. If someone says something in here that I disagree with, am I not allowed to use analytics to support my case? Do I have to say "I disagree with you, come to the stat thread to see why!" in order to show my case?
 

TheLegendOfPatPeake

Registered User
Jun 12, 2020
3,037
3,076
Washington D.C.
I agree with this. It's easy enough to scroll past a post you don't care to read. I realize that as a moderator @Calicaps probably can't put me on ignore, but all non-moderators are free to do so if they want as well.

It seems like it's more a way to stifle discussion than anything. If someone says something in here that I disagree with, am I not allowed to use analytics to support my case? Do I have to say "I disagree with you, come to the stat thread to see why!" in order to show my case?
Checks out to me.
 

twabby

Registered User
Mar 9, 2010
13,624
14,441
Given how much of an outlier his deployment is, we don’t really have much context on how well he should be doing. It’s in Quenville-levels of deployment shifting.

This is a fair statement. But of course it's tough to make the case that he is doing poorly given the zone start context as well. I think it's more likely than not he's doing well given his context, but I suppose it's still up for debate.

Another thing that is missing from ZSR is that it completely ignores on-the-fly zone starts, which make up the vast majority of every player's shift starts, and also neutral zone starts. While Kuznetsov's ZSR is 88%, only 24% of his shifts start in the offensive zone on a faceoff. 16% start in the neutral zone, 3% start in the defensive zone, and 57% start on the fly.
 
Status
Not open for further replies.

Ad

Upcoming events

Ad

Ad

-->