Prospect Info: Logan Stanley: How has he progressed?

[HannuJ]

Ves and Stanley have played a grand total of 21 NHL games between them. Labelling either of them a bust or success is absurd. Both players look like they could have a future role imo, but plenty remains to be seen.

without looking it up, how many games did Petan get?

 

[Skidooboy]

You have a fundamental misunderstanding of analytics and what predictiveness is.

If I'm rolling a die, I can predict that 33% of the time I will roll above a 4. I can predict that 16.67% of the time I will roll a 2. That doesn't help me predict exactly what the next roll is going to be, just the likelihoods. That is predictiveness. I can predict that over 1000 rolls I'll get ~333 5s or 6s.

In hockey, analytics attempts to look at the variables in a game and say "given these variables, team x has a y% chance of winning". Not "team x is going to win". So what good is that? Well, if you have team x playing a game that has them winning 70% of the time then over 1000 games you can expect them to win 700 games. Even then if you play a game where you're 70% likely to win you're still going to lose almost 1 out of every 3 games.

The questions you are asking are looking for answers that analytics have never claimed to provide.

My point is is that it’s not predictive. You cannot use probability to determine the next action. If you could it wouldn’t gambling.

Particularly when you’re dealing with an incomplete data set, as you would be when you’re dealing with a growing changing developing learning human being. Again stats are useful not definitive everybody keeps talking on this thread like they are definitive. People pull out a three-game average and act like it’s the end all and be all of analysis. It isn’t.

 

[tbone1968]

Well, by the end of his D+4 season, Anderson had played all of 18 games and scored a whopping 5 NHL points, and he didn’t start out in Finland, and spend a year learning the NA game.

Sure, maybe KVes sucks, and never scores another NHL goal. Or maybe he continues to develop within the system and becomes a bigger Harkins with a wicked shot, or a faster and sturdier Armia-lite who can finish on a breakaway, or an Appleton who can play the halfwall on the PP and chip in 10/20 or better in a given year while being punishingly tough to play against. Don’t we want one of those guys? Or is not being Barzal or Heinola or Josh Anderson enough to call him a bust?

Until last year Harkins was being called a bust. Ehlers was a playoff no-show and. Copp was a 3rd-liner and Stanley was booking his economy ticket on the last train to Bustville - all these guys were those things until they weren’t. I don’t like seeing retread vets paid to be mediocre on the Jets but I always, always want to see our prospects succeed because it keeps the team viable and it keeps hope alive.

Great post. Big Stan has been a nice surprise this year and I thought Ves looked fine last night. Nothing flashy but hustled. Big power forward hard along the boards. Give them a chance. Great for the team to have these young guys contributing on affordable contracts.

 

[Evil Little]

without looking it up, how many games did Petan get?

More than I thought, but I wonder what the point of this exercise is.

 

[Skidooboy]

You have a fundamental misunderstanding of analytics and what predictiveness is.

If I'm rolling a die, I can predict that 33% of the time I will roll above a 4. I can predict that 16.67% of the time I will roll a 2. That doesn't help me predict exactly what the next roll is going to be, just the likelihoods. That is predictiveness. I can predict that over 1000 rolls I'll get ~333 5s or 6s.

In hockey, analytics attempts to look at the variables in a game and say "given these variables, team x has a y% chance of winning". Not "team x is going to win". So what good is that? Well, if you have team x playing a game that has them winning 70% of the time then over 1000 games you can expect them to win 700 games. Even then if you play a game where you're 70% likely to win you're still going to lose almost 1 out of every 3 games.

The questions you are asking are looking for answers that analytics have never claimed to provide.

But you can’t actually predict what those roles will be it doesn’t matter if it’s 66.666% of the time you expect to roll a certain value because he might very well roll 300-1s in a row. We say it’s unlikely we like to pass off the idea that that could never happen, but as a lifelong dungeons and dragons player I can assure you it can.

it’s like the card deck. The odds of drawing a royal flush straight off the top of the deck are pretty minimal but they’re actually identical to the odds of drawing any other five cards straight off the deck. We apply meaning to the royal flush as opposed to a two of clubs or three of spades or five of diamonds or seven hearts in and nine are diamonds but the odds actually don’t change. And you can play the odds and play the odds play the odds and still lose. And that’s for a simple thing where there’s only 52 variablesNow imagine hockey.

A dynamic shifting action filled game. With a couple dozen humans running around with knives in thier feet. slumps happen hot streaks happen opponents learn your weaknesses and take advantage of them, you learn your opponents weaknesses and take advantage of them.line matchups change there’s so much happening.

So yeah I think when it comes to drafting and developing a young hockey player, I think if your entire glimpse of what that player could be is bound up in one statistical analysis, I don’t think you seeing the whole picture.

That’s why they don’t put error bars on goalie save percentage because if they did there wouldn’t be anything to talk about.

 

[HannuJ]

More than I thought, but I wonder what the point of this exercise is.

the number of reps needed to determine that a player is a bust. no correlation.
with Petan, there clearly were things happening in practice, etc to lead to the Jets giving up on him (his personal life aside). Stanley, it's the opposite.
vesalainen: tbd

 

[Whileee]

But you can’t actually predict what those roles will be it doesn’t matter if it’s 66.666% of the time you expect to roll a certain value because he might very well roll 300-1s in a row. We say it’s unlikely we like to pass off the idea that that could never happen, but as a lifelong dungeons and dragons player I can assure you it can.

it’s like the card deck. The odds of drawing a royal flush straight off the top of the deck are pretty minimal but they’re actually identical to the odds of drawing any other five cards straight off the deck. We apply meaning to the royal flush as opposed to a two of clubs or three of spades or five of diamonds or seven hearts in and nine are diamonds but the odds actually don’t change. And you can play the odds and play the odds play the odds and still lose. And that’s for a simple thing where there’s only 52 variablesNow imagine hockey.

A dynamic shifting action filled game. With a couple dozen humans running around with knives in thier feet. slumps happen hot streaks happen opponents learn your weaknesses and take advantage of them, you learn your opponents weaknesses and take advantage of them.line matchups change there’s so much happening.

So yeah I think when it comes to drafting and developing a young hockey player, I think if your entire glimpse of what that player could be is bound up in one statistical analysis, I don’t think you seeing the whole picture.

That’s why they don’t put error bars on goalie save percentage because if they did there wouldn’t be anything to talk about.

There's a difference between probabilities like the roll of a die and statistical models like a shot metrics based model. There are more variables, including unknown or unmeasured variables that affect the predictive characteristics. There are also assumptions about the model structures, that are usually simplified, creating more error. There are also outliers, that are sometimes random and sometimes systematic, but they are seldom examined or included in models because it's messy and difficult. Also, most model results aren't accompanied by statistical error estimates, such as confidence intervals, so most of the time there is excessive reliance on point estimates.

So, it's reasonable to be circumspect about the metrics. Yes, they have been demonstrated to be more predictive of future results than other metrics like goals for / against, but they still leave a lot of room for different outcomes. Also, the testing of the models tend to be much more robust for team performances than the relative value of players in different team settings, though that doesn't stop us from using them as a way to assess individual players. We have plenty of examples of players fluctuating from year to year or within a season, so it's not clear how stable these metrics are for a given player (as an example, Scheifele's shot metrics for / against have flipped this year. Is he a different player, playing differently, in different contexts that aren't fully adjusted for?)

I like the data and use them all the time to at least describe what's happening, but I'm more circumspect about using them to make firm conclusions about players or team potential. The really good data scientists tend not to bluster or bully, and recognize and explain the limitations of their analyses.

 

[DannyGallivan]

I disagree. The TNSE approach to Vesa has been nothing but patience. They've spent a season on the Moose reworking his game to be North American in orientation. They acknowledged pretty openly that his game needed a sort of functional rebuild. I think your comparison to Harkins is actually appropriate, but not with respect to the intent you lend to it. Harkins in my mind is exactly what they are trying to achieve with Vesa, only in a bigger body. Harkins is a 2015 and is only now ready to take the next step. His regular playing time likely comes at the expense of not re-signing someone like Perreault. Vesa is getting a shot only now in the NHL and even then it is substantially limited. Even if the Jets were to send Vesa down to the Moose for the remainder of the year I wouldn't take that as a sign that his time has run out, rather than the Jets don't see a regular fit for him in this year's roster so more and better development could be beneficial with the Moose, Covid notwithstanding. What I do expect is he bounces between taxi and regular squad, getting limited games and minutes for a taste at the bigs and to evaluate progress this year.

I think Vesa's ultimate role comes as playing third line wing on a Lowry-like line. Big body, hard to play against, defensively responsible (in time). Perfetti on the other hand has a much higher ceiling, and really, from draft onwards, has always been much more highly regarded than Vesa. While Vesa was a first rounder, he was quite late (24) and as such, a project of sorts, much like Roslovic who had a similar draft spot. But I don't recall anyone hyping Vesa to the extent that we see with Perfetti. I don't even view them as competing for the same role. Perfetti, if all works out, is likely a possession driving winger that would benefit Scheifele, PLD getting someone else assuming he (PLD) signs an extension. Assuming Perfetti doesn't become a center at the NHL level that is, something we were all drooling about prior to the PLD acquisition. With confidence and time Vesa might ultimately become a mini-Wheeler but I think that's somewhat improbable and perhaps delusional on my part, so for the time being I'll stick with my prediction from the start of this paragraph.

Appleton too was a 2015 draft. Holding Vesa to the same standard would give him at least another year to take hold of his opportunity. Fully acknowledging the massive difference in draft position between Appleton and Vesa.

You make some valid points. I would give him another year, but right up to now I haven't seen or heard anything that would give me confidence that he will be a regular NHLer. As you said, Harkins is likely next in line after Perreault. Maybe Ves ends up as a shutdown fourth liner... I haven't seen anything that resembles NHL-calibre offensive instincts in his game.

 

[garret9]

So then you know how many goals a player will score next season already .

cool how many goals will PLD score next game?

This fits right in to my favourite tweet:

Hockey analytics biggest detractors are those that do not know anything about hockey analytics.

 

[HannuJ]

This fits right in to my favourite tweet:

Hockey analytics biggest detractors are those that do not know anything about hockey analytics.

it's funny. you're seeing something similar (but different) with COVID analytics.
experts: if we do nothing, 10000x people will die.
governments: do a little bit.
Joe Public: where were the mass deaths that were predicted? we only had 5000x.

 

[Ducky10]

without looking it up, how many games did Petan get?

More

 

[Skidooboy]

This fits right in to my favourite tweet:

Hockey analytics biggest detractors are those that do not know anything about hockey analytics.

And hockey analytics guys won’t post error bars

 

[Skidooboy]

This fits right in to my favourite tweet:

Hockey analytics biggest detractors are those that do not know anything about hockey analytics.

And statisticians will also tell you that stats workreally well on populations but don’t often work well on individuals

 

[10Ducky10]

Analytics are fine but they are not everything. Merely a tool to be used among many others.

 

[MrBoJangelz71]

Analytics are fine but they are not everything. Merely a tool to be used among many others.

Agreed.

Problem many times is it is the sole tool available and there the only tool used, especially with assessing draft picks.

Stanley's poor appraisal on this forum pre draft was based purely on poor analytics from his first season in the O. His second season didn't take off till the halfway mark, where his development took a steep climb, but the numbers did not improve enough to drastically impact his first poor season.

Jets obviously put more stock in what their scouts were watching than what the numbers were telling them.

But I went back over our draft threads from 2016, and when I first brought up Stanley as a legit target for us, the responses I got on that were beyond comical. He was viewed as a complete hack on this forum, with ZERO skill and it was driven purely through the analytical lens and I was an idiot for suggesting him. Couldn't skate, couldn't pass, zero skill.

His size almost made him a bigger target (no pun intended) because it was automatically assumed he was drafted for his size.

 

[Joe Hallenback]

I had Stanley marked as a late 1st/early 2nd round guy because he is intriguing. He could have went anywhere late 1st especially in that particular draft. I think he is going to be a good NHLer and I think he will add offense to his game as he goes along.

The questions of why we didn't draft say Girard is likely most teams had him as a 2nd round pick. A very slight offensive D man who has difficulty defending. I think the way scouting is done now as opposed to even 4 or 5 years ago is starting to change. Smaller skating d men CAN play in the NHL even with defensive issues. If this draft was 2020 you would likely see Girard go in the mid 1st or higher.

Debrincat is the perfect example of the opposite of Stanley. The sub 5'9 forward who does score a ton. How many guys are sub 5'9 in the NHL currently? 5 or 6 ? Its hard to draft a guy that small and then expect him to continue to do what he does. Likely those guys go 2nd and later and then you look like a genius when they work out or not its a wasted pick

 

[KingBogo]

Agreed.

Problem many times is it is the sole tool available and there the only tool used, especially with assessing draft picks.

Stanley's poor appraisal on this forum pre draft was based purely on poor analytics from his first season in the O. His second season didn't take off till the halfway mark, where his development took a steep climb, but the numbers did not improve enough to drastically impact his first poor season.

Jets obviously put more stock in what their scouts were watching than what the numbers were telling them.

But I went back over our draft threads from 2016, and when I first brought up Stanley as a legit target for us, the responses I got on that were beyond comical. He was viewed as a complete hack on this forum, with ZERO skill and it was driven purely through the analytical lens and I was an idiot for suggesting him. Couldn't skate, couldn't pass, zero skill.

His size almost made him a bigger target (no pun intended) because it was automatically assumed he was drafted for his size.

I remember the ongoing gong show of the Stanley pick well. Even suggesting a wait and see stance was met with scorn.

 

[Whileee]

I remember the ongoing gong show of the Stanley pick well. Even suggesting a wait and see stance was met with scorn.

I was sour on it at first blush, but it quickly became a ridiculous platform for dumping on Chevy, Chipman, Zinger and Maurice. That's not an uncommon theme here. Remember the pages and pages of posts slamming Chevy for acquiring Pionk? Again, I was negative to begin, but it was pretty silly how many posters refused to even consider that maybe the Jets saw something in Pionk and he would turn out better than many predicted. My initial take was wrong on the Pionk acquisition, and I was perhaps overly pessimistic about the Stanley pick initially.

At this stage, the Pionk trade looks very good for the Jets.

 

[KingBogo]

I was sour on it at first blush, but it quickly became a ridiculous platform for dumping on Chevy, Chipman, Zinger and Maurice. That's not an uncommon theme here. Remember the pages and pages of posts slamming Chevy for acquiring Pionk? Again, I was negative to begin, but it was pretty silly how many posters refused to even consider that maybe the Jets saw something in Pionk and he would turn out better than many predicted. My initial take was wrong on the Pionk acquisition, and I was perhaps overly pessimistic about the Stanley pick initially.

At this stage, the Pionk trade looks very good for the Jets.

View attachment 395241

I'm ussually pretty nuetral when I feel I don't have enough knowledge to form a full opinion. While I watch every Jets game intently and pay attention to every scrap of Jets news I feel informed enough to have an opion on everything involving the current Jets. However for a player like Stanley I have no first hand knowledge when he was drafted. I never scout junior hockey nor do I watch much junior hockey other than the WJC so I don't jump to negative opinions on draft picks because rarely have I ever seen them play first hand. So getting mad just seems like at opportunity to kick your most disliked dog. Mostly just uninformed anger.