All Purpose Analytics and Extended Stats Discussion

[twabby]

Appreciate the time and effort that went into this. Gotta concede that it's not bad players throwing the curve off.

I still think the other possible flaw I mentioned - that the WAR stat isn't accurately assessing defensive WAR because it's not adjusting enough for deployment - is a much more likely explanation for the weird defenseman aging curve than hockey defensemen entering the NHL as finished products. As HFTN mentioned, I think the separate aging curves for offensive and defensive WAR in the original article you posted support this. The offensive one is a pretty standard aging curve for both forwards and defensemen:

While the defensive one is this weird plateau from about 20 to 32:

Either defense in hockey is some bizarre innate talent that lasts long after age-related decline dulls other skills, both in hockey and in every other sport I've seen aging curves for, or maybe hockey analytics just aren't at the point yet where they can properly adjust for differences in difficulty of deployment, and coaches naturally give players harder roles as they improve and easier roles as they decline, leading to a smoothing of the aging curve? The hockey analytics community borrows a lot of concepts from baseball analytics, as this article demonstrates, but this is one of the things where there's really no precedent in baseball. Every first baseman on every team in baseball is asked to make the same plays in the field and face the same pitchers, there's no way for teams to shelter a guy if he's bad or give him tougher assignments if he's good, and the only "systems" in place are shifts or tailored pitching approaches that are easy to quantify. As a result, comparing a player to the replacement level at their position is pretty straightforward (although even in baseball, quantifying defense is the still a matter of serious debate and even stat nerds take defensive metrics with a huge grain of salt). In hockey there's a massive difference between what the 1LD and the 3LD on the same team are expected to do and who they face, and there can be a massive difference between the roles of 1LD's on different teams, and their roles might change within a single game based on the game situation or what the opposing team's doing. I'm sure Dawson Sprigings is brilliant, but he can be brilliant and still not get quantifying defensive value right on the first try. There's a lot of really smart guys working on the much easier task of quantifying defensive value in baseball, and they haven't gotten it nailed down yet, so it's probably gonna be a while before we nail it down for hockey.



This paragraph frustrates me, because a lot of people have been saying since the start of this debate "he had covid and a massive increase in workload which made him fall off in the second half." He shouldn't be treated like he played at the same level all season. Look at the Smoothed 5v5 xG/60 in this chart:

A lot of black pre-covid, a lot of red afterwards. I know you hate narratives, but I think there's a pretty clear one here of a player who was coming into his own, then got covid, and the lingering effects of that plus conditioning issues stemming from being limited by injuries to only 24 games the year before caused him to lose a step. If post-covid Fehervary was all we saw no one would be questioning your assessment, but it seems like you're totally writing off what he did before covid and ignoring some pretty obvious explanations for his drop off in the second half.

I appreciate the thoughtful writeup.

Defense probably is harder to quantify, and if I recall correctly there is more variation in defensive performance year-by-year than in offensive performance. However, Fehervary still had a miserable defensive year by most metrics so the baseline for me has to be that he is simply a bad defensive player rather than a good defensive player who, through things that are out of his control such as negative variance, put up bad defensive numbers.

I also don't think it's out of line to suggest that defensive aptitude is an innate skill moreso than offense. Defense seemingly relies on hockey IQ, smarts, and effort level, perhaps more than on things you can really learn with experience and practice. Offensively perhaps you can pick up more things through experience, such as a quicker shooting release, a harder shot, what types of passes are more difficult to defend, how to beat NHL level neutral zone traps and goalies, etc.

Regarding the pre-COVID and post-COVID results: unfortunately I cannot see a way to figure out how to get separate WAR values for these two periods of time on Evolving Hockey. It appears that I can only get the entire year's worth of WAR data on Evolving Hockey and I can't filter by date ranges. However, I can get simple stats like xGF% and the like for the different periods. It's difficult to ignore that even before he got COVID, Fehervary's numbers dropped substantially while away from Carlson (42% away from Carlson, 56% with Carlson), while Carlson's numbers stayed excellent while away from Fehervary (55% away from Fehervary in that same time period). The sample is small (Fehervary only played 80 5v5 minutes away from Carlson pre-COVID), but again it's a data point that suggests that perhaps Carlson was more responsible for the great numbers that the 42-74 pairing put up in the first half of the season. I suspect his WAR values would not have actually been that much different in the first half of the season even if his on-ice results were different because I suspect it was John Carlson who was more responsible for those on-ice results, but again this is only a suspicion based on some simple WOWYs. I am not confident in this claim at all.

Dismissing his second half poor performance because of COVID also misses a grim possibility that I haven't brought up yet: COVID is a bitch and realistically can wreak havoc on athletes for years. Long COVID is a real thing. If COVID truly did impact Fehervary negatively from January to May, I'm not confident he's simply going to return in October and feel all better. Getting COVID could be more similar to sustaining an injury than it is to simply feeling sick and then fully recovering. I would like to see how Fehervary performs early next year and that should give us some indication of whether he's recovered or not.

If on the other hand he recovered from COVID but simply hit the proverbial rookie wall, then I would claim that plenty of other rookies of his ilk have similarly hit a wall, and wouldn't really give cause to treat Fehervary as an exception to his peer group.

 

[twabby]

Pretty haughty stance for someone who asserts NHL defensemen are fully formed on debut.

You’re wrong. Every single person here has told you you’re wrong. This isn’t even a debate, it’s just you plugging your ears while everyone else tries to help you, and just like every other time you’ve been wrong it’s going absolutely nowhere, and with the utmost arrogance

I can't engage with people who don't have an open mind on the topic of player development, and won't even consider it open for debate. In my opinion that is a pretty stubborn and arrogant mindset.

Unless you're willing to even consider the possibility that the popular beliefs about player-development may be wrong, how can I possibly have a productive conversation with you? You've already made your conclusion and it's clear you won't budge from that conclusion.

 

[Langway]

I also don't think it's out of line to suggest that defensive aptitude is an innate skill moreso than offense. Defense seemingly relies on hockey IQ, smarts, and effort level, perhaps more than on things you can really learn with experience and practice.

If there was ever any doubt you weren't trolling...

 

[twabby]

If there was ever any doubt you weren't trolling...

I promise you I'm not trolling. Do you see an exclamation point at the end of that sentence?

I mean hockey IQ and smarts in the sense that some players just get it, and some don't. Instinct, perhaps, for lack of a better description. You don't learn instinct, it's natural. You also don't necessarily learn to become a hard-worker after joining the NHL. Work ethic is likely something that's drilled into a person as a child.

Perhaps defensive instincts are more important than what was previously thought. Perhaps work-ethic is just incredibly difficult to change and players don't really become harder workers once they enter the NHL.

 

[Langway]

Oh, sure. While we're at it let's lower the draft age to 14 or 16 since we're missing out on all of these innately supreme talents that are otherwise wasting their time until they're eligible. Sad really.

 

[twabby]

Oh, sure. While we're at it let's lower the draft age to 14 or 16 since we're missing out on all of these innately supreme talents that are otherwise wasting their time until they're eligible. Sad really.

Straw man - Wikipedia

en.m.wikipedia.org

 

[Langway]

Respectfully, you are just off the rails here.

 

[twabby]

Respectfully, you are just off the rails here.

I feel like I have a pretty good batting average when it comes to some “off the rails” predictions I have.

Makes me feel confident I’m on the right track.

 

[CapitalsCupReality]

Might be time to add up your ledger again bud….you’re heavily in the red lol….

…….but amusing!

 

[Carlzner]

Why do NHL players even bother with working on their game in the offseason if they’re at their peak as a rookie?

Fools!

 

[twabby]

Why do NHL players even bother with working on their game in the offseason if they’re at their peak as a rookie?

Fools!

You’re right. Players will just continue getting better as long as they work hard every offseason. That’s why we see 85 year old Otis MacMillan tearing up the league.

 

[Carlzner]

You’re right. Players will just continue getting better as long as they work hard every offseason. That’s why we see 85 year old Otis MacMillan tearing up the league.

Sure yeah it's not like there's a common age where fitness and athletics start to decline across almost literally every sport

 

[twabby]

It’s nice to know that when I squatted 405 pounds when I was 25 years old (50 years ago), those gains were locked in. I didn’t have to do anything to maintain my squat!

 

[Carlzner]

It’s nice to know that when I squatted 405 pounds when I was 25 years old (50 years ago), those gains were locked in. I didn’t have to do anything to maintain my squat!

Typo??

Pretty common knowledge there is no improvement made in weightlifting after the age of 21.

 

[HTFN]

I can't engage with people who don't have an open mind on the topic of player development, and won't even consider it open for debate. In my opinion that is a pretty stubborn and arrogant mindset.

Unless you're willing to even consider the possibility that the popular beliefs about player-development may be wrong, how can I possibly have a productive conversation with you? You've already made your conclusion and it's clear you won't budge from that conclusion.

You must have brass f***ing balls to accuse someone who just engaged with all your numbers of being closed minded while you’re over here so convinced you’re right you won’t listen to every athlete and observer willing to tell you otherwise.

You’re right. Players will just continue getting better as long as they work hard every offseason. That’s why we see 85 year old Otis MacMillan tearing up the league.

Seriously the f*** is this? Your willing and open mind? You’re not even trying to engage with anyone else in good faith and no, your track record is not all that great. How’d the Calgary Flames do this year?

 

[Kalopsia]

I appreciate the thoughtful writeup.

Defense probably is harder to quantify, and if I recall correctly there is more variation in defensive performance year-by-year than in offensive performance. However, Fehervary still had a miserable defensive year by most metrics so the baseline for me has to be that he is simply a bad defensive player rather than a good defensive player who, through things that are out of his control such as negative variance, put up bad defensive numbers.

I'm not arguing it was due to negative variance, I'm arguing that A) the WAR stat isn't adjusting for deployment correctly, resulting in overvaluing players with easy deployments and undervaluing players like Fehervary with difficult deployments, and B) Fehervary's second half went poorly due to the unique combination of covid, heavy workload, and coming off an injury shortened season creating a conditioning issue. The only bad luck I think is at play is that he caught covid in the middle of what was looking like a breakout season.

I also don't think it's out of line to suggest that defensive aptitude is an innate skill moreso than offense. Defense seemingly relies on hockey IQ, smarts, and effort level, perhaps more than on things you can really learn with experience and practice. Offensively perhaps you can pick up more things through experience, such as a quicker shooting release, a harder shot, what types of passes are more difficult to defend, how to beat NHL level neutral zone traps and goalies, etc.

Really seems like you're grasping at straws here to defend the models. Honestly, which do you think is more reasonable: that defensive aptitude in hockey is an innate skill even though offensive aptitude isn't and no one before you ever noticed it (including the people you're citing for the aging curves, none of whom seem to be drawing this conclusion), or that there are still kinks to work out in defensive WAR? I know one of your big things is not putting too much faith in the opinions of coaches and GMs to evaluate players, but it seems like you're doing the same thing, just with the guys that develop these WAR models. These are all black boxes, right? You can't actually see how they generate the WAR numbers, you just have to trust that their model works? In the end, either way, you're trusting someone's judgement on how to assess a hockey player's value.

Also, what's the difference between hockey IQ and smarts?

Regarding the pre-COVID and post-COVID results: unfortunately I cannot see a way to figure out how to get separate WAR values for these two periods of time on Evolving Hockey. It appears that I can only get the entire year's worth of WAR data on Evolving Hockey and I can't filter by date ranges. However, I can get simple stats like xGF% and the like for the different periods. It's difficult to ignore that even before he got COVID, Fehervary's numbers dropped substantially while away from Carlson (42% away from Carlson, 56% with Carlson), while Carlson's numbers stayed excellent while away from Fehervary (55% away from Fehervary in that same time period). The sample is small (Fehervary only played 80 5v5 minutes away from Carlson pre-COVID), but again it's a data point that suggests that perhaps Carlson was more responsible for the great numbers that the 42-74 pairing put up in the first half of the season. I suspect his WAR values would not have actually been that much different in the first half of the season even if his on-ice results were different because I suspect it was John Carlson who was more responsible for those on-ice results, but again this is only a suspicion based on some simple WOWYs. I am not confident in this claim at all.

Another poster brought this up a few days ago, so I'll just copy from my response to that one. Carlson played with Orlov when he was away from Fehervary, and Fehervary played with Schultz or Jensen when he was away from Carlson. Of course Carlson got better when he played with the best D on the team, and Fehevary got worse when he played with either the worst D in Schultz or a bad stylistic fit in Jensen. All those numbers show is that Orlov > Fehervary, Carlson > Schultz, and Fehervary-Jensen is a bad pairing, none of which is exactly a revelation.

Edit: Looking into this yields some interesting results. Three players got at least 10 minutes on ice with both Fehervary and Carlson prior to Fehervary getting covid, and all three of them did better with Fehervary than with Carlson.

xGF%	w/ Fehervary	w/ Carlson	Difference
Orlov	59.53	56.50	+3.03
TVR	43.66	37.65	+6.01
Schultz	18.43	15.02	+3.41

Small sample sizes of course, but a data point in Fehervary's favor. Side note, fire Justin Schultz into the sun.

Dismissing his second half poor performance because of COVID also misses a grim possibility that I haven't brought up yet: COVID is a bitch and realistically can wreak havoc on athletes for years. Long COVID is a real thing. If COVID truly did impact Fehervary negatively from January to May, I'm not confident he's simply going to return in October and feel all better. Getting COVID could be more similar to sustaining an injury than it is to simply feeling sick and then fully recovering. I would like to see how Fehervary performs early next year and that should give us some indication of whether he's recovered or not.

Fair point. This is definitely my biggest fear for Fehervary.

If on the other hand he recovered from COVID but simply hit the proverbial rookie wall, then I would claim that plenty of other rookies of his ilk have similarly hit a wall, and wouldn't really give cause to treat Fehervary as an exception to his peer group.

My best guess is that covid exhaustion plus limited playing time due to injury the previous year plus the heavy workload added up to the rookie wall to end all rookie walls. As long as he doesn't have long-covid extending into next season, I think he deserves the benefit of the doubt for having the worst-case scenario confluence of events to crush him in the second half.

 

[g00n]

I feel like I have a pretty good batting average when it comes to some “off the rails” predictions I have.

Makes me feel confident I’m on the right track.

I think you need your own thread to log your predictions and track their true accuracy. Like Twabby Babble or something.

Or is this that thread? Maybe a post with the specific predictions, then.

 

[twabby]

I'm not arguing it was due to negative variance, I'm arguing that A) the WAR stat isn't adjusting for deployment correctly, resulting in overvaluing players with easy deployments and undervaluing players like Fehervary with difficult deployments, and B) Fehervary's second half went poorly due to the unique combination of covid, heavy workload, and coming off an injury shortened season creating a conditioning issue. The only bad luck I think is at play is that he caught covid in the middle of what was looking like a breakout season.

Was Fehervary’s deployment really that difficult? In terms of TOI he got 19:30 a night, about 17:13 at 5v5. He certainly wasn’t sheltered, but his TOI wasn’t off the charts either. His QOC and QOT was also similarly pretty ordinary:

And of all rookie defensemen this year who played 250+ 5v5 minutes, Fehervary ranked 17th of 33 players in terms of QOC as measured by his opponents’ WAR/60. He also had a 58% o-zone to d-zone start ratio.

I don’t really think that his deployments were responsible for his poor performance.

Really seems like you're grasping at straws here to defend the models. Honestly, which do you think is more reasonable: that defensive aptitude in hockey is an innate skill even though offensive aptitude isn't and no one before you ever noticed it (including the people you're citing for the aging curves, none of whom seem to be drawing this conclusion), or that there are still kinks to work out in defensive WAR? I know one of your big things is not putting too much faith in the opinions of coaches and GMs to evaluate players, but it seems like you're doing the same thing, just with the guys that develop these WAR models. These are all black boxes, right? You can't actually see how they generate the WAR numbers, you just have to trust that their model works? In the end, either way, you're trusting someone's judgement on how to assess a hockey player's value.

Also, what's the difference between hockey IQ and smarts?

I never said that offensive ability isn’t an innate talent, though I admit I phrased it poorly. All that I said was that it’s possible that experience provides more impact to a player’s offensive game than their defensive game.
There is no difference between hockey IQ and smarts. I meant to delete one of them but left them both in by mistake.

Evolving Hockey has posted their writeup for WAR. They haven’t made their code public, but they have made their methodology public. So I wouldn’t call it a black box.

Evolving-Hockey

evolving-hockey.com

I don’t fully trust the models, but I’ll take them over a lot of other people’s subjective eye-test opinions, including my own. WAR models are optimized to try to measure a player’s impact on goal differential. Most scouts, coaches, executives, and fans are way too emotionally involved to make cold judgments on players. Or focus way too much on the wrong things, like size.

While WAR was not designed to be predictive, it can be used to observe and quantify trends and it can be used to make some educated guesses.

Another poster brought this up a few days ago, so I'll just copy from my response to that one. Carlson played with Orlov when he was away from Fehervary, and Fehervary played with Schultz or Jensen when he was away from Carlson. Of course Carlson got better when he played with the best D on the team, and Fehevary got worse when he played with either the worst D in Schultz or a bad stylistic fit in Jensen. All those numbers show is that Orlov > Fehervary, Carlson > Schultz, and Fehervary-Jensen is a bad pairing, none of which is exactly a revelation.

Edit: Looking into this yields some interesting results. Three players got at least 10 minutes on ice with both Fehervary and Carlson prior to Fehervary getting covid, and all three of them did better with Fehervary than with Carlson.

xGF%	w/ Fehervary	w/ Carlson	Difference
Orlov	59.53	56.50	+3.03
TVR	43.66	37.65	+6.01
Schultz	18.43	15.02	+3.41

Small sample sizes of course, but a data point in Fehervary's favor. Side note, fire Justin Schultz into the sun.

This is exactly why I think WAR stats can be incredibly useful.

In essence all popular WAR models are, to my knowledge, regressions that take all of the players, teammates, competition, game-state, deployment, on-ice results, etc. as input. Instead of hyperfocusing on a specific WOWY it takes them all into account plus a multitude of other factors and tries to make a best fit. Given all of these factors as input, what WAR makes the most sense for all of these players?

Also, despite some pretty good on-ice numbers, Justin Schultz is not looked at kindly by WAR:

Seems about right to me!

 

[Jags]

But perhaps the assumption you’re making is wrong. Are you so out of touch? No, it’s the children who are wrong.

I get wanting to challenge conventional wisdom, but it can't rightfully be supplanted by something that just isn't wise. With regard to the "all NHL defensemen" part of your argument, the data just isn't there yet. You've reached a conclusion like "2 + 2 = 9", shrugged, pointed at it, and said, "Hey, I know everyone thinks the answer is 4, but what if everyone else in the world is wrong and I'm right? The numbers support it."

And with specific regard to Fehervary, you keep coming back to this idea that he had a shitty season. He didn't. He was good for a big chunk of it, and the part that was dodgy was the part where the whole team stunk. We can't even specifically diagnose all the factors that went into that decline, but there's no denying it happened. Did he suffer because Carlson did, vice versa, or do you feel the stat model accounts for that adequately? Is the quality of his minutes affected by Carlson's (like on a night where special teams play 20 minutes)? Were there systems changes that may have exacerbated issues in the attempt to address them? Is there a decent sample of numbers of him playing with another partner? Was he just gassed by injury/COVID or the big swings in the density of the schedule? Which guy is the real guy?

2021-2022 was a tale of two seasons for the Caps as a whole. Acting like Fehervary had an objectively bad season is at least a huge generalization, and at most just a bullshit conclusion you sniffed out and are now trying to justify by any means necessary.

The people who come up with these numbers do so with much more humility than you're deploying here.

 

[twabby]

I get wanting to challenge conventional wisdom, but it can't rightfully be supplanted by something that just isn't wise. With regard to the "all NHL defensemen" part of your argument, the data just isn't there yet. You've reached a conclusion like "2 + 2 = 9", shrugged, pointed at it, and said, "Hey, I know everyone thinks the answer is 4, but what if everyone else in the world is wrong and I'm right? The numbers support it."

And with specific regard to Fehervary, you keep coming back to this idea that he had a shitty season. He didn't. He was good for a big chunk of it, and the part that was dodgy was the part where the whole team stunk. We can't even specifically diagnose all the factors that went into that decline, but there's no denying it happened. Did he suffer because Carlson did, vice versa, or do you feel the stat model accounts for that adequately? Is the quality of his minutes affected by Carlson's (like on a night where special teams play 20 minutes)? Were there systems changes that may have exacerbated issues in the attempt to address them? Is there a decent sample of numbers of him playing with another partner? Was he just gassed by injury/COVID or the big swings in the density of the schedule? Which guy is the real guy?

2021-2022 was a tale of two seasons for the Caps as a whole. Acting like Fehervary had an objectively bad season is at least a huge generalization, and at most just a bullshit conclusion you sniffed out and are now trying to justify by any means necessary.

The people who come up with these numbers do so with much more humility than you're deploying here.

He did have a bad season though. You can segregate it any way you want but in totality he was below replacement level last season. Maybe it was because of COVID, maybe it was because of his injuries the prior season, maybe both, and maybe there isn't an excuse. But it was a bad season. You can make excuses if you want, but every player deals with hardships. Many of them performed much better than Fehervary.

I think the WAR model accounts for the on-ice performance reasonably well. As mentioned above, this particular model takes into account every teammate and every opponent, among other items and regresses to come up with a best guess for a player's isolated performance.

While Washington performed much better in the 2021 portion of the season than the 2022 portion, that doesn't mean individual players performed poorly all year. Carlson had a fantastic year (ignoring the playoffs of course!). Orlov and Jensen had great seasons overall. TVR was solid.

I think a lot of people are naturally invested in Fehervary and that leads to some wonky evaluations of his game. Suppose Fehervary was instead an undrafted free agent that Washington signed last offseason at the age of 22, and then he had the season he had. Would you be singing the same praises for his season? Would he be penciled in for top 4 duty entering next season by management?

 

[Langway]

Was Fehervary’s deployment really that difficult? In terms of TOI he got 19:30 a night, about 17:13 at 5v5. He certainly wasn’t sheltered, but his TOI wasn’t off the charts either.

In terms of EVTOI Fehervary as a rookie defenseman was in the 84th percentile among post-lockout rookies that played at least 60 games. Lower it to 40 games and it borders on the 90th percentile. It was the highest among Caps rookie defenders in the post-lockout era...more than Carlson averaged in his first 100+ games. It was more than Cale Makar, Zach Werenski, Drew Doughty, Duncan Keith, Victor Hedman, Brett Pesce, Adam Fox, PK Subban, Erik Karlsson, Morgan Rielly, Seth Jones, Matt Niskanen, Thomas Chabot, Mattias Ekholm, Devon Toews, Roman Josi, Dmitry Orlov, Jakob Chychrun, Shea Theodore and on in their rookie campaigns. Not altogether off the charts, no, but well above average. Of the 32 rookie D over 17 seasons that logged more EVTOI only 11 did so on playoff teams, where developmental TOI isn't so readily handed out. Of the 11 there are zero flame-outs. Of the 32 there aren't any either. Some have bounced around a bit but that's hockey. That extent of usage for a rookie defender on a playoff team was uncommon.

It's even more significant in light of Laviolette's tendencies when it comes to inexperienced players and his generally high standard of defensive engagement. If one thinks NHL coaches do know what they're doing--and certainly some do not share that belief--then that level of usage and a PK role suggests he's not deemed to be a Madison Bowey level tire fire defensively. If that were the case he would have been a third pair ES-only player, rotating in/out of the lineup and not the top four staple he remained in the second half. He may have been lucky compared to expected goals share or something but I'm wary of chalking that up to sheer luck and one does need to acknowledge what was, again, fairly pronounced usage for a rookie. Pronounced for the Caps, for Laviolette, for pretty well anyone in that situation. Plenty of good teams and D have such 'luck'...often repeatedly. I'm fairly confident of a moderate projection of a Sekera type 3/4 with more offensive flashes and physicality. It's a fair projection based on his tools and his rookie campaign was in line with that. Should he put it together and link up offense/defense better maybe that improves further. Many of the raw attributes are there. Pay and play him along those lines and I don't see the problem. He's cheap for another season and by then they should have a clearer insight as to his role.

Smarts and effort level not being areas that can be improved with further applied experience would seem to deny the existence of learned expertise a.k.a. veteran savvy. A very long list of grizzled Cup winners trash that line of thinking. Not that it's a singularly determinative factor in why teams win but it's often a good part of the reason why those lacking it lose come playoff time. All in all analytics are helpful perspectives to be aware of and to further dissect individual players/attributes but they still do not dictate the agenda for any team and probably will not any time soon. Colorado still dresses Jack Johnson, Andrew Cogliano & Darren Helm. Tampa still dresses Bogosian, Maroon & Bellemare. It takes a blend, particularly at the bottom of the lineup with role players.

 

[g00n]

In terms of EVTOI Fehervary as a rookie defenseman was in the 84th percentile among post-lockout rookies that played at least 60 games. Lower it to 40 games and it borders on the 90th percentile. It was the highest among Caps rookie defenders in the post-lockout era...more than Carlson averaged in his first 100+ games. It was more than Cale Makar, Zach Werenski, Drew Doughty, Duncan Keith, Victor Hedman, Brett Pesce, Adam Fox, PK Subban, Erik Karlsson, Morgan Rielly, Seth Jones, Matt Niskanen, Thomas Chabot, Mattias Ekholm, Devon Toews, Roman Josi, Dmitry Orlov, Jakob Chychrun, Shea Theodore and on in their rookie campaigns. Not altogether off the charts, no, but well above average. Of the 32 rookie D over 17 seasons that logged more EVTOI only 11 did so on playoff teams, where developmental TOI isn't so readily handed out. Of the 11 there are zero flame-outs. Of the 32 there aren't any either. Some have bounced around a bit but that's hockey. That extent of usage for a rookie defender on a playoff team was uncommon.

It's even more significant in light of Laviolette's tendencies when it comes to inexperienced players and his generally high standard of defensive engagement. If one thinks NHL coaches do know what they're doing--and certainly some do not share that belief--then that level of usage and a PK role suggests he's not deemed to be a Madison Bowey level tire fire defensively. If that were the case he would have been a third pair ES-only player, rotating in/out of the lineup and not the top four staple he remained in the second half. He may have been lucky compared to expected goals share or something but I'm wary of chalking that up to sheer luck and one does need to acknowledge what was, again, fairly pronounced usage for a rookie. Pronounced for the Caps, for Laviolette, for pretty well anyone in that situation. Plenty of good teams and D have such 'luck'...often repeatedly. I'm fairly confident of a moderate projection of a Sekera type 3/4 with more offensive flashes and physicality. It's a fair projection based on his tools and his rookie campaign was in line with that. Should he put it together and link up offense/defense better maybe that improves further. Many of the raw attributes are there. Pay and play him along those lines and I don't see the problem. He's cheap for another season and by then they should have a clearer insight as to his role.

Smarts and effort level not being areas that can be improved with further applied experience would seem to deny the existence of learned expertise a.k.a. veteran savvy. A very long list of grizzled Cup winners trash that line of thinking. Not that it's a singularly determinative factor in why teams win but it's often a good part of the reason why those lacking it lose come playoff time. All in all analytics are helpful perspectives to be aware of and to further dissect individual players/attributes but they still do not dictate the agenda for any team and probably will not any time soon. Colorado still dresses Jack Johnson, Andrew Cogliano & Darren Helm. Tampa still dresses Bogosian, Maroon & Bellemare. It takes a blend, particularly at the bottom of the lineup with role players.

The rookie TOI factoring in PLAYOFF teams, and not rebuild projects, is what I asked about in the first thread. Thanks for looking that up. I wonder how that WAR progress plot looks with just the dozen elite guys

 

[twabby]

In terms of EVTOI Fehervary as a rookie defenseman was in the 84th percentile among post-lockout rookies that played at least 60 games. Lower it to 40 games and it borders on the 90th percentile. It was the highest among Caps rookie defenders in the post-lockout era...more than Carlson averaged in his first 100+ games. It was more than Cale Makar, Zach Werenski, Drew Doughty, Duncan Keith, Victor Hedman, Brett Pesce, Adam Fox, PK Subban, Erik Karlsson, Morgan Rielly, Seth Jones, Matt Niskanen, Thomas Chabot, Mattias Ekholm, Devon Toews, Roman Josi, Dmitry Orlov, Jakob Chychrun, Shea Theodore and on in their rookie campaigns. Not altogether off the charts, no, but well above average. Of the 32 rookie D over 17 seasons that logged more EVTOI only 11 did so on playoff teams, where developmental TOI isn't so readily handed out. Of the 11 there are zero flame-outs. Of the 32 there aren't any either. Some have bounced around a bit but that's hockey. That extent of usage for a rookie defender on a playoff team was uncommon.

It's even more significant in light of Laviolette's tendencies when it comes to inexperienced players and his generally high standard of defensive engagement. If one thinks NHL coaches do know what they're doing--and certainly some do not share that belief--then that level of usage and a PK role suggests he's not deemed to be a Madison Bowey level tire fire defensively. If that were the case he would have been a third pair ES-only player, rotating in/out of the lineup and not the top four staple he remained in the second half. He may have been lucky compared to expected goals share or something but I'm wary of chalking that up to sheer luck and one does need to acknowledge what was, again, fairly pronounced usage for a rookie. Pronounced for the Caps, for Laviolette, for pretty well anyone in that situation. Plenty of good teams and D have such 'luck'...often repeatedly. I'm fairly confident of a moderate projection of a Sekera type 3/4 with more offensive flashes and physicality. It's a fair projection based on his tools and his rookie campaign was in line with that. Should he put it together and link up offense/defense better maybe that improves further. Many of the raw attributes are there. Pay and play him along those lines and I don't see the problem. He's cheap for another season and by then they should have a clearer insight as to his role.

Smarts and effort level not being areas that can be improved with further applied experience would seem to deny the existence of learned expertise a.k.a. veteran savvy. A very long list of grizzled Cup winners trash that line of thinking. Not that it's a singularly determinative factor in why teams win but it's often a good part of the reason why those lacking it lose come playoff time. All in all analytics are helpful perspectives to be aware of and to further dissect individual players/attributes but they still do not dictate the agenda for any team and probably will not any time soon. Colorado still dresses Jack Johnson, Andrew Cogliano & Darren Helm. Tampa still dresses Bogosian, Maroon & Bellemare. It takes a blend, particularly at the bottom of the lineup with role players.

Peter Laviolette also chose to give Justin Schultz significant playing time in 20-21 over Nick Jensen, a superior player by pretty much every metric and by pretty much everyone's eye test. Even this season everyone is pretty much in agreement that Fehervary struggled down the stretch yet he was continually played in a top 4 role by Laviolette with tacit approval by MacLellan as shown by his lack of a deadline upgrade at the LD position despite good options being available. This doesn't exactly make me want to put much stock into their opinions of a player when they have already shown plenty of questionable recent player-evaluations.

I think learned experience is almost certainly a thing. As mentioned, it appears that players improve offensively in their early 20s on average and I'm sure that has something to do with experience and veteran savvy. Defensively this could also true, but perhaps the value of experience is simply outweighed by other physical aspects that decline in your early twenties, such as reaction time.

Being as impactful as Sekera is a very good-case scenario for Fehervary's development IMO. It's not what I'd expect him to turn into, on average. Sekera only finished 2 of his 15 seasons with a worse WAR rate than Fehervary finished with this season, and both of those were fewer than 40 games so there is a bit more variance potentially baked into those results. Sekera's rookie campaign in particular was much stronger than Fehervary's from a WAR perspective as well. His age 22 season was similarly more impressive than Fehervary's from a WAR perspective.

I don't think it really matters if one can't fully explain why defensive ability plateaus rather than rises in a player's early 20s. The data is what the data is. For a long time people didn't know why the sun rose and set, but even without that understanding they could reasonably predict when and where the sun would rise. Just because the WAR data doesn't jive with current dogma doesn't mean the data is flawed or the interpretation of the data is flawed.

 

[HTFN]

Peter Laviolette also chose to give Justin Schultz significant playing time in 20-21 over Nick Jensen, a superior player by pretty much every metric and by pretty much everyone's eye test. Even this season everyone is pretty much in agreement that Fehervary struggled down the stretch yet he was continually played in a top 4 role by Laviolette with tacit approval by MacLellan as shown by his lack of a deadline upgrade at the LD position despite good options being available. This doesn't exactly make me want to put much stock into their opinions of a player when they have already shown plenty of questionable recent player-evaluations.

I think learned experience is almost certainly a thing. As mentioned, it appears that players improve offensively in their early 20s on average and I'm sure that has something to do with experience and veteran savvy. Defensively this could also true, but perhaps the value of experience is simply outweighed by other physical aspects that decline in your early twenties, such as reaction time.

I don't think it really matters if one can't fully explain why defensive ability plateaus rather than rises in a player's early 20s. The data is what the data is. For a long time people didn't know why the sun rose and set, but even without that understanding they could reasonably predict when and where the sun would rise. Just because the WAR data doesn't jive with current dogma doesn't mean the data is flawed or the interpretation of the data is flawed.

Wow.

What a load of shit

 

[RedRocking]

The data is what the data is. For a long time people didn't know why the sun rose and set, but even without that understanding they could reasonably predict when and where the sun would rise.

Yea, we know, the data is the moon, the sun, the stars and your religion

But seriously, I think you bring interesting conversations and analysis here - just maybe try being a little more humble, or avoid such grandiose statements? Just a thought. And yes, I will promptly show myself out of this thread.