Speculation: Caps General Discussion (Coaching/FAs/Cap/Lines etc) - 2021 "Season" Pt. 2

[g00n]

Yes, the fact that his teammates enjoy better chance differentials with him than without him could be luck. There is a small amount noise to xGF and other stats so it's possible, though improbable.

But wouldn't you also say that goal-scoring and individual points are also subject to incredible amounts of luck? Especially in an 11 game sample? The same logic applies, except of course that individual point-scoring is much, much noisier than xGF and other "fancy" statistics. Scoring points could be the result of luck, placebo, or any number of factors outside of your control.

I invite you to check out this post, where I identified a few instances where Kuznetsov made very nice plays that did not result in individual points, but very easily could have if his teammates did just a little bit better in their finishing: Speculation: - Caps General Discussion (Coaching/FAs/Cap/Lines etc) - 2021 "Season" Pt. 2. Wouldn't you say that he was a little bit unlucky to not have registered a point on these plays? I'm sure there are many other chances where he was unlucky to have registered a point.

To answer your final question: he is driving goal differential. He has probably not been a product of his teammates, otherwise you would likely see his teammates enjoying similar amounts of success (or better) without him on the ice, and this simply hasn't been the case.

Once again, individual points and mucking aren't tabulated on the final scoreboard, only team goals.

If the sample size is too small why are we even talking about the stats?

Doesn't every player make nice plays that teammates fail to convert?

 

[CapitalsCupReality]

It's not ideal but most years here he blew away the investment. Do we go back and retroactively pay him 10 million for the cup season, or just cherry pick the bad ones (in which he's still playing well)? What about the seasons when he was making 3 million playing like a 7 million dollar player?

The discussion makes some sense if he's not producing and playing like shit, but not when he's playing reasonably well

it makes plenty of sense unless you have low expectations.

is this actually Kuzy, ?

because it’s rare you get to rest on your laurels after a few good years and one great playoff run.

 

[Kuznetsnow]

it makes plenty of sense unless you have low expectations.

is this actually Kuzy, ?

because it’s rare you get to rest on your laurels after a few good years and one great playoff run.

If that's what you call 11 games of driving the play coming back from covid sure.

Should've sent that bum Ovechkin packing after his 65 point season as well

 

[CapitalsCupReality]

If that's what you call 11 games of driving the play coming back from covid sure.

Should've sent that bum Ovechkin packing after his 65 point season as well

hyperbole much?

 

[Kuznetsnow]

hyperbole much?

All of this discussion is. 6 points in 11 games coming off Covid is all ado about nothing. The only noteworthy thing is that he's playing pretty well given that it seems to f*** athletes up more than normal people. Would it be better if he had more points? Sure. Ice is slippery, not everything goes in some times when it would go in other times at the same level of play. Look at Zibanejad for what a concerning turn of events looks like

 

[CapitalsCupReality]

Sorry, this issue goes beyond 11 games....but whatever. Not like I expected an honest discussion with your profile. As I said before....let’s revisit after 20 games. I’m told the production is coming. All is right with great offensive production.

 

[AussieCapsFan]

I got a sense that Kuzy is probably put off by the suspension from international play he got handed. So he's like all "whatever, man".

I get the sense that he's a lazy bastard who wanted to win a Cup and make a shit-ton of money, and he's done that. And now he's going to coast and make f*ck all effort for the remainder of his contract because he wants the $$$ so he can retire.

Which equates perfectly with his on-ice performance since he won a Cup.

 

[Melkor]

Nick Jensen should be a prime candidate for US National Olympic team if we are looking simply on these metrics. But everyone knows he's a 3rd pairing guy who can't shoot, can't pass, can't be physical, but can skate the puck out the zone occasionally only to wet his pants once he crosses the opposition's blue line. Context matters. Without context, these stats within 11 game sample size mean so little, it's not even worth a discussion. But we have a much bigger sample size of Kuznetsov lookong like he doesn't really like hockey anymore, playing accordingly and his actual scoring stats slowly dropping evey year since the cup. That's what we know and see for sure unless some of the posters have glasses with stanley cup highlights on replay and refuse to take em off.

 

[traparatus]

Another note: some players are talented enough setup men or shooters that they can consistently outperform their expected goal metrics over the long run. Similarly, some players are the opposite, and typically underperform their expected goal metrics over the long run.

Kuznetsov, over the course of his career, has been the former: his on-ice goal differential is 8 points higher than his expected goal differential, by far the best among any player on the team since 2007-08, including Ovechkin. In fact there hasn't been a single season where Kuznetsov hasn't outperformed his expected goal metrics.

So the fact that his underlying metrics are so strong should be incredibly exciting for fans of the Capitals.

That's great and let's hope that lasts. It would be a huge boost to our Cup chances.

To me, at this moment in time, Kuznetsov's point production is somewhat secondary to him developing better habits and rounding out his game. He doesn't have to be an elite defensive forward but he can't be the worst defensive top-6 center in the league. He doesn't have to be great on faceoffs but he can't be at 40%. His work ethic and consistency must improve. I've seen enough players put up terrific offensive seasons while barely breaking even in terms of game winning impact and that includes Kuznetsov.

By the way, I appreciate your patience in these discussions. It ain't always easy. The more information the better, IMO. I consider myself pretty decent at watching a hockey game but I still miss more than I see. Where did the line start their shift? No idea. Who were they playing against? Only know that if they were stuck in the defensive zone or got scored on. How many chances did they generate or give up? That stays in my head for about 30 seconds.

 

[ovikovy817]

Don't understand talk about trading Kuzya for futures or whatever.

The ultimate goal in hockey is to win games . In the NHL is to win the Cup.
Kuzya can win you games and can win you the Cup.
Remember the guy who had a conn Smythe performance 2 years ago.?

You can count on the fingers of one hand players that produce every night in this league. You just need to get it going at the right moment. Like DSP who was hot for 6 7 games at the right moment in his career.

I know it's a Hockey Future Board but c'mon. Stop value a pick higher than an actual player.

 

[twabby]

If the sample size is too small why are we even talking about the stats?

Doesn't every player make nice plays that teammates fail to convert?

The sample size is too small to look at individual point totals, but not at expected goals. The predictive power of xG, both on a team and individual level, converges much more quickly than goals for at both a team an individual level: Expected Goals are a better predictor of future scoring than Corsi, Goals

From the article, first at the team level:

As you can see after 10 games, a team's GF% has essentially no predictive power, with an r^2 of approximately 7%. Meanwhile, xGF% has much more predictive power: after only 10 games it has an r^2 of 24%. Not even close to a perfect predictor, but much better than GF% which is essentially random in a 10 game sample. In other words, after 10 games, one would rather know a team's xGF% instead of their GF% (and indeed, their win-loss record) to determine if they are any good, at least at 5v5.

And this makes sense, right? Look at the last 10 games on the NHL standings and we see the following oddities:

Edmonton: 8-2-0
Ottawa: 5-5-0
Montreal: 3-5-2
Chicago: 8-2-0
Dallas: 2-4-4

So now we go to the individual level, where we see the following:

In terms of individual on-ice performance, xGF% is still a much better predictor than GF% in terms of predicting future GF%. Even after just 10 games (Kuznetsov has played 11) xGF shows MUCH better predictive power than GF%. Again it's not perfect, and it gets better after 20-30 games, but it is still far better to look at than actual goals for. In other words, I'd rather look at someone's xGF% after 10 games than their goal differential if I wanted to know how their goal differential might look for the rest of the season.

And indeed, Kuznetsov is currently sporting a healthy 62% xGF%. That'll do.

Now for those who are interested in individual points, there's also good news: expected goals on an individual basis predicts future individual goals much better than current individual goals, even after 10 games:

After only 10 games the r^2 between ixG and future iG is a massive 63%, compared to 22% between iG and future iG. Again, in other words, after a 10 game sample if we want to know how a player is going to score goals for the rest of the season you'd be better off looking at their ixG rather than their actual goals scored. And for Kuznetsov he currently has the third best ixG/60 rate on the team, behind only Ovechkin and Sheary. And he has the second best ixG/60 at 5v5, behind only Sheary. Unfortunately his results have lagged his xG rates: at 5v5 he has only scored 0.4 goals/60 despite having a ixG/60 of 0.78, and at all strengths it's 0.68 and 0.85, respectively.

The above article does not address assists unfortunately, but that's ok because his assist rate right now is actually very good. He's 4th best on the team in 5v5 A1/60 at 0.8, which is a touch below his career average but can also be easily explained by saying his on-ice shooting percentage is 8.5% at 5v5, also below his career average by about the same amount that his A1 rate is. He'll likely pick up a few more primary assists based on his on-ice shooting percentage improving a bit. And he already leads the team in primary assists in all-situations on a per 60 basis, so I don't think there's much to complain about there!

So that leaves secondary assists. Kuznetsov doesn't have a single secondary assist this year, which is largely the reason why his point totals are disappointing on the surface. But fortunately I don't think it really matters much: The noise surrounding secondary assists - TSN.ca

Secondary assists are just one data point, but a critical one to keep in mind when analyzing a player’s true production. There is no doubt at this point that they are a mostly random attribution that aren’t reflective of a player’s true offensive impact, but somehow continue to get lumped into point totals – the same point totals that can drive how much a front office may or may not be willing to pay for a player in the future.

The secondary assists will come, especially if he keeps driving offense the way he has been so far.

And to answer your final question: yes, every player makes nice plays that teammates fail to convert. But I'd argue that Kuznetsov has been especially unlucky in this regard so far, at least when it comes to individual point production.

 

[twabby]

Sorry, this issue goes beyond 11 games....but whatever. Not like I expected an honest discussion with your profile. As I said before....let’s revisit after 20 games. I’m told the production is coming. All is right with great offensive production.

If his underlying numbers were the same as they were for the past two seasons, then I'd also be alarmed. But something has changed this season:

Season	GP	TOI	GF%	SF%	FF%	CF%	xGF%	xGF/60	xGA/60	xGD/60
18-19	76	1046.27	58.56	47.11	47.48	47.94	45.45	2.62	3.15	-0.53
19-20	63	847.02	49.96	43.58	44.43	45.41	45.85	2.42	2.85	-0.43
20-21	11	149.27	88.4	57.06	56.38	56.44	61.79	3.16	1.95	1.21

[TBODY] [/TBODY]

 

[CapitalsCupReality]

Sample size, etc etc....


You don’t get to choose when sample size matters to fit the debate.

And....for a numbers guy, you sure seem to play the luck (magic) card an awful lot....

 

[crab]

Nick Jensen should be a prime candidate for US National Olympic team if we are looking simply on these metrics. But everyone knows he's a 3rd pairing guy who can't shoot, can't pass, can't be physical, but can skate the puck out the zone occasionally only to wet his pants once he crosses the opposition's blue line. Context matters. Without context, these stats within 11 game sample size mean so little, it's not even worth a discussion. But we have a much bigger sample size of Kuznetsov lookong like he doesn't really like hockey anymore, playing accordingly and his actual scoring stats slowly dropping evey year since the cup. That's what we know and see for sure unless some of the posters have glasses with stanley cup highlights on replay and refuse to take em off.

Jensen is a great example and one I used last year against the HFBoards advanced stats champions. Jensen was ranked as one of the top advanced stats defenseman in the last few years. They honestly made the claim he was better than Carlson last season (this season they might have a case lol).

Another is Shattenkirk, advanced stat darling. Caps fans had both Shattenkirk and Carlson playing on the team at the same time to be able to make the assessment. Those same advanced stats champions claimed Shattenkirk is a better player. Shattenkirk was absolute trash every time I seen him play.

 

[TheLegendOfPatPeake]

Don't understand talk about trading Kuzya for futures or whatever.

The ultimate goal in hockey is to win games . In the NHL is to win the Cup.
Kuzya can win you games and can win you the Cup.
Remember the guy who had a conn Smythe performance 2 years ago.?

You can count on the fingers of one hand players that produce every night in this league. You just need to get it going at the right moment. Like DSP who was hot for 6 7 games at the right moment in his career.

I know it's a Hockey Future Board but c'mon. Stop value a pick higher than an actual player.

Who suggested trading him for futures or a draft pick?

 

[AlexBrovechkin8]

Alright I did not come here to read about R-squared values and linear regressions. @twabby maybe when the sample size gets above 50 games you can compare the Z and T scores and report back?? This thread has officially jumped the shark

 

[TheLegendOfPatPeake]

On another note, we could be the Sabres who are paying 27 million for three goals between Skinner, Eichel, and Hall.

 

[twabby]

Sample size, etc etc....


You don’t get to magically choose when sample size matters to fit the debate.

And....for a numbers guy, you sure seem to play the luck (magic) card an awful lot....

Luck, probability, and randomness is probably the biggest reason why we are having this debate.

 

[CapitalsCupReality]

No, lack of production and are we getting a 7.8 mil performance every year is the reason we’re having this discussion.

 

[twabby]

Alright I did not come here to read about R-squared values and linear regressions. @twabby maybe when the sample size gets above 50 games you can compare the Z and T scores and report back?? This thread has officially jumped the shark

Justifying a relatively unpopular opinion usually requires some work.

But your point is well-taken. I think for now we can officially say that the talk of Kuznetsov playing poorly or needing to play better is "malarkey", as Joe Biden would say.

 

[g00n]

The sample size is too small to look at individual point totals, but not at expected goals. The predictive power of xG, both on a team and individual level, converges much more quickly than goals for at both a team an individual level: Expected Goals are a better predictor of future scoring than Corsi, Goals

From the article, first at the team level:

As you can see after 10 games, a team's GF% has essentially no predictive power, with an r^2 of approximately 7%. Meanwhile, xGF% has much more predictive power: after only 10 games it has an r^2 of 24%. Not even close to a perfect predictor, but much better than GF% which is essentially random in a 10 game sample. In other words, after 10 games, one would rather know a team's xGF% instead of their GF% (and indeed, their win-loss record) to determine if they are any good, at least at 5v5.

And this makes sense, right? Look at the last 10 games on the NHL standings and we see the following oddities:

Edmonton: 8-2-0
Ottawa: 5-5-0
Montreal: 3-5-2
Chicago: 8-2-0
Dallas: 2-4-4

So now we go to the individual level, where we see the following:

In terms of individual on-ice performance, xGF% is still a much better predictor than GF% in terms of predicting future GF%. Even after just 10 games (Kuznetsov has played 11) xGF shows MUCH better predictive power than GF%. Again it's not perfect, and it gets better after 20-30 games, but it is still far better to look at than actual goals for. In other words, I'd rather look at someone's xGF% after 10 games than their goal differential if I wanted to know how their goal differential might look for the rest of the season.

And indeed, Kuznetsov is currently sporting a healthy 62% xGF%. That'll do.

Now for those who are interested in individual points, there's also good news: expected goals on an individual basis predicts future individual goals much better than current individual goals, even after 10 games:

After only 10 games the r^2 between ixG and future iG is a massive 63%, compared to 22% between iG and future iG. Again, in other words, after a 10 game sample if we want to know how a player is going to score goals for the rest of the season you'd be better off looking at their ixG rather than their actual goals scored. And for Kuznetsov he currently has the third best ixG/60 rate on the team, behind only Ovechkin and Sheary. And he has the second best ixG/60 at 5v5, behind only Sheary. Unfortunately his results have lagged his xG rates: at 5v5 he has only scored 0.4 goals/60 despite having a ixG/60 of 0.78, and at all strengths it's 0.68 and 0.85, respectively.

The above article does not address assists unfortunately, but that's ok because his assist rate right now is actually very good. He's 4th best on the team in 5v5 A1/60 at 0.8, which is a touch below his career average but can also be easily explained by saying his on-ice shooting percentage is 8.5% at 5v5, also below his career average by about the same amount that his A1 rate is. He'll likely pick up a few more primary assists based on his on-ice shooting percentage improving a bit. And he already leads the team in primary assists in all-situations on a per 60 basis, so I don't think there's much to complain about there!

So that leaves secondary assists. Kuznetsov doesn't have a single secondary assist this year, which is largely the reason why his point totals are disappointing on the surface. But fortunately I don't think it really matters much: The noise surrounding secondary assists - TSN.ca



The secondary assists will come, especially if he keeps driving offense the way he has been so far.

And to answer your final question: yes, every player makes nice plays that teammates fail to convert. But I'd argue that Kuznetsov has been especially unlucky in this regard so far, at least when it comes to individual point production.

Twabbs, I appreciate the rigorous explanation, but the graphs mostly show how xG is slightly better than some other terrible predictors per the R^2 values, and that ixG is still a bit above coin flip. Yes, 63% is a lot better than the team comparison values but it's still only good RELATIVE to the other methods.

So again it comes down to belief in the stat rather than the stat's actual predictive power.

 

[twabby]

Twabbs, I appreciate the rigorous explanation, but the graphs mostly show how xG is slightly better than some other terrible predictors per the R^2 values, and that ixG is still a bit above coin flip. Yes, 63% is a lot better than the team comparison values but it's still only good RELATIVE to the other methods.

So again it comes down to belief in the stat rather than the stat's actual predictive power.

Of course there is almost surely room to improve upon xGF. In fact the definition of the xGF has been tweaked over time to improve upon earlier versions. The version used in the article is from 6 years ago, and I imagine current definitions such as the one from Evolving Hockey (RPubs - A New Expected Goal Model for Predicting Goals in the NHL) improve upon this in terms of predictive power, though I can't say for sure how much.

Of course, the follow-up question becomes: what exactly predicts future goal differential best? How do you know that what your eyes see are a better indicator of future success than xGF? Have you run the numbers and shown that your eye-test is a better predictor than xGF?

And another question becomes how much does randomness, which I'll define as "the set things that are not reasonably in control of the player or team", play a role in current and future outcomes? I suspect the role of randomness is incredibly large, and that there is a practical theoretical maximum to the predictive power of models (including a perfect eye-test), and it might be smaller than all of us think.

 

[zappa4ever]

LOL
Fancy stats are not gospel.
Fancy stats are not 100% predictive or reliable or valid. (and nowhere close to it)
So no, "we" can definitely not officially say "malarkey" to Kuz playing poorly or needing to play better.

"I investigated myself and declare myself officially 100% not guilty"

....I think for now we can officially say that the talk of Kuznetsov playing poorly or needing to play better is "malarkey", as Joe Biden would say.

 

[BiPolar Caps]

On another note, we could be the Sabres who are paying 27 million for three goals between Skinner, Eichel, and Hall.

Eichel's having a down year, but I'd do whatever it would take to acquire him for the Caps. That's a 24 year old 1C who has talent and the best years in front of him. Granted the Skinner contract is a woof, still he's just 28 and two seasons removed from a 40 goal season. Probably the best way to acquire Eichel will be for a team to also be willing to take Skinner as well. That's 19 mil of cap a team would have to take on. For the Capitals to make it work and do something like that, you'd have to include the likes of Kuznetsov (of course acquiring Eichel would remove the need for Kuzy), Orlov, Vrana (soon to be RFA currently around 3.3, Skinner would replace Vrana at LW), Caps may also have to add say Samsonov in order for the Sabres to eat about 3 mil of Skinner's salary so that it would work for the Caps.

Top two lines would look something like:

Ovechkin - Eichel - Wilson
Skinner - Backstrom - Oshie

 

[g00n]

Of course there is almost surely room to improve upon xGF. In fact the definition of the xGF has been tweaked over time to improve upon earlier versions. The version used in the article is from 6 years ago, and I imagine current definitions such as the one from Evolving Hockey (RPubs - A New Expected Goal Model for Predicting Goals in the NHL) improve upon this in terms of predictive power, though I can't say for sure how much.

Of course, the follow-up question becomes: what exactly predicts future goal differential best? How do you know that what your eyes see are a better indicator of future success than xGF? Have you run the numbers and shown that your eye-test is a better predictor than xGF?

And another question becomes how much does randomness, which I'll define as "the set things that are not reasonably in control of the player or team", play a role in current and future outcomes? I suspect the role of randomness is incredibly large, and that there is a practical theoretical maximum to the predictive power of models (including a perfect eye-test), and it might be smaller than all of us think.

If we think randomness and luck are such huge factors then there's not much point in statistical analysis. Or any analysis other than "was that luck or skill?"

Regarding the bold, the questions are a bit circular. If you can't trust the numbers then how can you verify an eye test with them?

The fundamental question about observational analysis vs statistical analysis is not going to be resolved by stats, unless those stats are completely objective and the observations are similarly objective and measured in a way that can be tested against those stats. If randomness is always there to bail out any given analysis it's sort of pointless.

I would say that when assessing how a player is doing, or their potential, a good deal of variation is going to exist because not everyone is equally skilled or perceptive when it comes to such evaluations. My "eye test" is not going to be the same as Lavi's.

But much of sport is art and flow as much as it is metrics and points. Can you quantify art, or flow states? Do you know art when you see it? Can you spot someone in the Zone?

A lot of this perception-based analysis hinges on the same sorts of unconscious information processing that's involved in creating or viewing art/music, or being in the zone. The creativity and proficiency needed for high level performance, and the decision making that happens in split seconds, will elude any chart or table because it comes from the same place. If it didn't then every junior or AHL player who puts up big numbers would just tear up the NHL, and we know that doesn't happen. There's something ELSE missing.

This drives stat people nuts because it can't be pinned down or bottled or fit into a spreadsheet, but it's at least as important as any statistic you can name.

This is what frustrates people about Kuzy. He's clearly got the skill and sense to produce and do the right things on the ice but for some reason he doesn't deliver with the consistency of the top stars.

That said, Kuzy is close. He's a 30 goal talent for sure. At 11 games and 2 goals, he's only about 2 goals off that pace. But he has to pick up his current pace quickly due to the shortened season, otherwise we lose runway for his anticipated streak.

Also, this kind of discussion is why we started that fancy stats thread that's collecting dust in the HF attic...