Prospect Info: The 2024-2025 Prospect Thread: Part 1: Skate or Die!

[ManVanFan]

Those crappy numbers said that Tracey had a 40% chance of being an NHL regular when it was nothing of the sort. Dude was a 5th-6th round type who got inflated due to elite linemates and PP usage.

These numbers are shit. Again : it’s worth understanding baseline percentages as a general thing - understanding that on a very macro level production=good, understanding hit rates per round to understand trade values of picks. But trying to apply them to specific players without any sort of context is dumb as f*** and anyone trying to do it understands neither math nor hockey.

A GROUP OF NHL SCOUTS! With their eye tests from probably multiple NHL scouts. Deemed him worthy of a 1st round selection.

What team do you work for? You make a living off scouting? I don't think so. Otherwise I am guarantee you aren't spending half your day on here.

 

[Yultron]

Mr Second half is catching fire , Riley Patterrson scores again

 

[Tables of Stats]

A GROUP OF NHL SCOUTS! With their eye tests from probably multiple NHL scouts. Deemed him worthy of a 1st round selection.

What team do you work for? You make a living off scouting? I don't think so. Otherwise I am guarantee you aren't spending half your day on here.

NHL scouts miss far more often than they hit and the league still has a ton of hold-outs from the era when being hired as a scout could be a pure act of nepotism or a reward for a loyal soldier. Being an NHL scout means an NHL team likes you and makes you a scout, it doesn't mean you're even in the top 50th percentile for scouting ability.

If you want to talk about scouts using any given analytics source as an authoritative claim that those scouts know something we don't you'd need to look at the hit rate for those specific scouts, their lists, who they like, why they like them, etc. Then you'd have to show that the best scouts use these sources. If you can't do that saying scouts use anything is as useful as saying that Benning used something as justification for why it's good.

 

[ManVanFan]

NHL scouts miss far more often than they hit and the league still has a ton of hold-outs from the era when being hired as a scout could be a pure act of nepotism or a reward for a loyal soldier. Being an NHL scout means an NHL team likes you and makes you a scout, it doesn't mean you're even in the top 50th percentile for scouting ability.

If you want to talk about scouts using any given analytics source as an authoritative claim that those scouts know something we don't you'd need to look at the hit rate for those specific scouts, their lists, who they like, why they like them, etc. Then you'd have to show that the best scouts use these sources. If you can't do that saying scouts use anything is as useful as saying that Benning used something as justification for why it's good.

NHL Scout's Analysis: Inside the player evaluation process

Jason Bukala has 16 years experience scouting at the NHL level for the Nashville Predators and Florida Panthers. Now he's joining Sportsnet to give you an inside look at scouting departments and the player evaluation process.

www.sportsnet.ca

Here's an article from a former head of scouting and about the process.

Analytics have become a very useful tool in the process as well. Team analytics departments routinely communicate with the scouting staff throughout the season and alert the group of developing trends. It’s not one or the other; the more information, the better. In recent years, especially during the pandemic, video scouting has added another layer of intel and assists the process.

That's some real impressive pretending that only bad scouts use analytics. Tin foil hat theory.

 

[Tables of Stats]

NHL Scout's Analysis: Inside the player evaluation process

Jason Bukala has 16 years experience scouting at the NHL level for the Nashville Predators and Florida Panthers. Now he's joining Sportsnet to give you an inside look at scouting departments and the player evaluation process.

www.sportsnet.ca

Here's an article from a former head of scouting and about the process.

Analytics have become a very useful tool in the process as well. Team analytics departments routinely communicate with the scouting staff throughout the season and alert the group of developing trends. It’s not one or the other; the more information, the better. In recent years, especially during the pandemic, video scouting has added another layer of intel and assists the process.

That's some real impressive pretending that only bad scouts use analytics. Tin foil hat theory.

Nobody is arguing against the idea of analytics as a useful tool. We're arguing about how worthless the specific models you keep touting as useful are. You don't seem to understand what makes a model useful, when and where models can be trusted, and how a model can be harmful if blindly followed. You seem stuck at a very sophomoric level of understanding where you find models that agree with your preexisting opinions and then give yourself gold stars for being the smartest person in the room.

Go, gain an understanding of what makes a model useful, and then come back when you can articulate a point without arguing to some vague authority that you think proves your point.

 

[krutovsdonut]

That's why I wasn't comparing them stylistically but rather on value and sure Kudryavtsev is a bit heavier and probably brings more offense but his offense also probably won't translate to the NHL either and he becomes a Jordan Spence type of guy which may be a slightly better version of Stecher.

If he ever gets to the value of a Spence then the Canucks would have a win here.

i mean in one sense you're right. if he plays a single nhl game they have a win. the fact he was worthy of a contract is a win.

but not sure about the comps you are choosing. neither stecher nor spence represent kudryatsev's body type, ceiling or projected style of play. stecher was tiny in junior and didn't blossom until his last season at north dakota when he finally bulked up. i doubt he could have played in the ahl at 20. spence is a special and pretty unique case of elusive and, well, here is an only slightly unfair photo.

 

[ManVanFan]

Nobody is arguing against the idea of analytics as a useful tool. We're arguing about how worthless the specific models you keep touting as useful are. You don't seem to understand what makes a model useful, when and where models can be trusted, and how a model can be harmful if blindly followed. You seem stuck at a very sophomoric level of understanding where you find models that agree with your preexisting opinions and then give yourself gold stars for being the smartest person in the room.

Go, gain an understanding of what makes a model useful, and then come back when you can articulate a point without arguing to some vague authority that you think proves your point.

You pretend as if you have read up on that specific model.
Who's blindly following it? I use it as a projection tool on how a player is progressing. I literally said they aren't set in stone. I have also never said you shouldn't watch a player play. At this point you guys are making up bullshit to pretend you are schooling me about something. "Pre-existing opinions"? Where do you get this shit from?

I've been reading many models from Chatel, Bacon, McCullum, now Nick. Canucks army guy doing the PCS% back in the day. I can't even remember all the other ones. I've read more about models than you know and interacted with the people asking questions about the models. A good decade of understanding the models and how they work and what they can provide me for info. All these flaws that you guys are trying to tell me about and everything else I said were thought's I had half a decade ago when I didn't understand them.

 

[Tables of Stats]

You pretend as if you have read up on that specific model.
Who's blindly following it? I use it as a projection tool on how a player is progressing. I literally said they aren't set in stone. At this point you guys are making up bullshit to pretend you are schooling me about something. "Pre-existing opinions"? Where do you get this shit from?

I've been reading many models from Chatel, Bacon, McCullum, now Nick. Canucks army guy doing the PCS% back in the day. I can't even remember all the other ones. I've read more about models than you know and interacted with the people asking questions about the models. A good decade of understanding the models and how they work and what they can provide me for info. All these flaws that you guys are trying to tell me about and everything else I said were thought's I had half a decade ago when I didn't understand them.

The old, "You're all just too uninformed to understand" argument doesn't hold much water when you consistently fail to demonstrate your understanding. You sound like an AI bro talking about how AI art is good actually and how the normies just don't understand the brilliance of the soulless slop that these models produce. If you want to educate about models, you should start showing this deeper understanding of them that you claim to have.

If Melvin were to come back and make claims, I'd listen because he showed that he could apply his knowledge of models to produce his own model that was probably better than average. I've yet to see you do anything approaching that.

 

[ManVanFan]

The old, "You're all just too uninformed to understand" argument doesn't hold much water when you consistently fail to demonstrate your understanding. You sound like an AI bro talking about how AI art is good actually and how the normies just don't understand the brilliance of the soulless slop that these models produce. If you want to educate about models, you should start showing this deeper understanding of them that you claim to have.

If Melvin were to come back and make claims, I'd listen because he showed that he could apply his knowledge of models to produce his own model that was probably better than average. I've yet to see you do anything approaching that.

What exactly do I fail to understand about the models?

 

[MS]

A GROUP OF NHL SCOUTS! With their eye tests from probably multiple NHL scouts. Deemed him worthy of a 1st round selection.

What team do you work for? You make a living off scouting? I don't think so. Otherwise I am guarantee you aren't spending half your day on here.

1. NHL scouting is a closed industry where 98% of the people doing it are ex-players or relatives of NHL executives. It's such a shitty industry that someone as dumb as Jim Benning can rise to the top of it.

I don't really have much doubt that plenty of us posting on this board understand scouting better than most NHL scouts. In the case of Tracey I watched every minute of that team that I could that year because Jett Woo was on it and it was dead f***ing obvious that this was a player with no projectable skills riding his linemates to impressive production. And I called it repeatedly at the time :

Pre-Game Talk: - 2019 NHL Draft, Pt. V: Got your ticket? (Mod note in pinned post)

Krebs was maybe a borderline candidate at our pick before his injury ... but with a torn Achilles there is no way in hell he should go in the top 15. You realize this isn’t the NFL draft right? Most of these players are two years away from playing in the NHL. If the medical staff is confident...

forums.hfboards.com

2019 NHL Draft, Pt. II

Vasili Podkolzin has the exact same PPG that Arytom Manukyan did at the same age in the MHL

forums.hfboards.com

Bonus discussion there of why Kaedan Korczak was better than Lassi Thompson, something NHL scouts also got wrong that year.

2. There is a huge difference between scouts having complex internal models showing zone entry stuff, etc. and the stuff we're talking about which is just bad math graphs showing that Kudryatsev is comparable to Jordan Subban or blindly tossing Zack MacEwen into a pool of non-comparable players based solely on a production:age ratio.

Also I suspect that some of the stuff that is considered common sense for draft watchers ('production=good', '10 point coke machine defender=bad', 'goalies are juju and taking them in the top half of the first round is a bad idea') is mindblowing when presented statistically to former NHL players.

3. In the case of the sorts of models we're talking about, @Melvin basically did this on this board with his Potato system a couple years ago. And this was a smart guy who worked in statistics in MLB (a sport that actually lends itself to this stuff, unlike hockey) and his whole point was that his system was absolutely terrible but that it was hilarious how it could still basically match the efforts of an average NHL team. But what we see now is a lot of dumb-guys-who-think-they're-smart basically copying that sort of crappy model and presenting it as some sort of really smart stuff they invented, which is totally laughable.

Like, I love math. I love statistics. But that means that I also equally hate bad math and bad statistics and it's so f***ing frustrating to see people present obviously bad math as some sort of smart model.

 

[Tables of Stats]

What exactly do I fail to understand about the models?

I never claimed you don't understand the models, I've claimed you've failed to demonstrate that knowledge. The fact that you'd rather argue yourself in circles, and make claims that NHL scouts use models, therefore... rather than point to times when the models you're agreeing with have made good predictionsshows that either you don't understand the models as well as you think. It's either that, or you do or that you understand them but are so awful at communication you fail to impart that knowledge to others.

So which is it?

 

[ManVanFan]

I never claimed you don't understand the models, I've claimed you've failed to demonstrate that knowledge. The fact that you'd rather argue yourself in circles, and make claims that NHL scouts use models, therefore... rather than point to times when the models you're agreeing with have made good predictionsshows that either you don't understand the models as well as you think. It's either that, or you do or that you understand them but are so awful at communication you fail to impart that knowledge to others.

So which is it?

The failure to listen isn't on me. Here's an example. I said before I came to a better understanding of models, the thoughts/questions/flaws/doubts that have been communicated in here were ones I had also a long time ago. One would have to listen and come in with an open mind and not already made up assumptions.

 

[Tables of Stats]

The failure to listen isn't on me.

This is a complete non-sequitur; nobody here is saying that models = bad.

We're saying that your presentation style hasn't done anything to show that you understand them or to show why any given model you're using to argue for something is worth a damn. You can't just look at a time, in this case a specific pick, when a model was right and the scouts were wrong and say, "Look, model good." You need to look at the overall accuracy of the model as a whole and understand when and where using a model is useful and when and where it's useful to go off model.

You're still just arguing like an AI bro or a Joe Rogan truther and assuming the supremacy of your position is so self-evident that just showing examples without explanation will get your point across.

 

[MS]

The failure to listen isn't on me. Here's an example. I said before I came to a better understanding of models, the thoughts/questions/flaws/doubts that have been communicated in here were ones I had also a long time ago. One would have to listen and come in with an open mind and not already made up assumptions.

I basically consider anyone using NHLe to be mentally handicapped, and using Bader as an example given the number of terrible takes he's had is just hilarioius.

 

[ManVanFan]

This is a complete non-sequitur; nobody here is saying that models = bad.

We're saying that your presentation style hasn't done anything to show that you understand them or to show why any given model you're using to argue for something is worth a damn. You can't just look at a time, in this case a specific pick, when a model was right and the scouts were wrong and say, "Look, model good." You need to look at the overall accuracy of the model as a whole and understand when and where using a model is useful and when and where it's useful to go off model.

You're still just arguing like an AI bro or a Joe Rogan truther and assuming the supremacy of your position is so self-evident that just showing examples without explanation will get your point across.

I believe the words were "those crappy numbers" "these numbers are shit", can't forget, "I am doubting any model's robustness that has Jordan Subban as the 5th most comparable player." Then can't forget the arguing with the owner of the model that because he put names, in the "best statistical comparison" that it is also a meaningful player comparison. (Owner of the model said it's not) (I tried to explain it's not either) Can't forget that they also don't "score adjust" for other leagues. Even though it's been updated multiple times over the years. Even recently as this Dec. Can't forget that "in other words, this model implies the development of this player has been static if not stunted". Lol. I can't make this shit up, it's written on the boards.

post draft development is, by far, the best indicator of eventual nhl success, especially, for obvious reasons, in later round picks. i would say a player showing this kind of development is more likely to succeed than a respectable second rounder he is currently caught up to.

I wouldn't disagree because the model would show that.

We're arguing about how worthless the specific models you keep touting as useful are.
Then there is you who is arguing that your opinion is the model is useless because you don't like the answer. You haven't give me any examples of why the model is useless.

I basically consider anyone using NHLe to be mentally handicapped, and using Bader as an example given the number of terrible takes he's had is just hilarioius.

I basically consider anyone that thinks on the internet that they are a super scout to be just hilarious. I'm talking about you since you won't figure that out. You asked for examples and lost.

 

[MS]

I basically consider anyone that thinks on the internet that they are a super scout to be just hilarious. I'm talking about you since you won't figure that out.

I never said I was a 'super scout'. I'm wrong about things all the time, Mynio being a recent example that I'm taking a few learning points from. But you don't have to be a 'super scout' to be at the NHL standard when the NHL standard is thinking that guys like Brayden Tracey and Lassi Thompson were worthy of 1st round picks.

And I'm sure as f*** smarter than the idiots putting out these models. I can sleep pretty soundly on that. Any one using models as their go-to isn't worth anything.

But also, like, I can show you a post from 2022 when both Dakota Joshua and Kiefer Sherwood were in the AHL as Group VI UFA types where I called them as potential NHL players who I'd target as signings. Good luck showing me a model that did that.

 

[Tables of Stats]

I believe the words were "those crappy numbers" "these numbers are shit", can't forget, "I am doubting any model's robustness that has Jordan Subban as the 5th most comparable player." Then can't forget the arguing with the owner of the model that because he put names, in the "best statistical comparison" that it is also a meaningful player comparison. (Owner of the model said it's not) (I tried to explain it's not either) Can't forget that they also don't "score adjust" for other leagues. Even though it's been updated multiple times over the years. Even recently as this Dec. Can't forget that "in other words, this model implies the development of this player has been static if not stunted". Lol. I can't make this shit up, it's written on the boards.

post draft development is, by far, the best indicator of eventual nhl success, especially, for obvious reasons, in later round picks. i would say a player showing this kind of development is more likely to succeed than a respectable second rounder he is currently caught up to.

I wouldn't disagree because the model would show that.

We're arguing about how worthless the specific models you keep touting as useful are.
Then there is you who is arguing that your opinion is the model is useless because you don't like the answer. You haven't give me any examples of why the model is useless.

Those graphic sheets by guys who won't share how their model works are pretty bad. That they are occasionally correct doesn't invalidate this critique. If you want to sway me, go do the leg work, look at these models you agree with see which picks they've made publically and show how they've tracked compared to what the model said they'd do.

Also, you're still raging at nothing, boxing at shadows, you've yet to put your money down to show anything. You can't seem to articulate what you've learned in your years studying these models, can't seem to articulate why these models are useful, can't seem to muster a word about why one model might be more robust or useful than another model. You're still just standing here stamping your feet insisting that models are good with the seeming justification for this being that sometimes models get it right where scouts get it wrong.

You're going to need to do the leg work. I suggest that you pick a few of your favourite models and show that they consistently beat actual NHL picks as well as lists put out by the media. If you do that, I'll have a reason to think that those specific models are actually useful. Until you can manage that, I'll stick to using methods of judging players and prospects that use the eye test, my own guess about personality and team/development program fit, and statistics to make my calls.

 

[ManVanFan]

I never said I was a 'super scout'. I'm wrong about things all the time, Mynio being a recent example that I'm taking a few learning points from. But you don't have to be a 'super scout' to be at the NHL standard when the NHL standard is thinking that guys like Brayden Tracey and Lassi Thompson were worthy of 1st round picks.

And I'm sure as f*** smarter than the idiots putting out these models. I can sleep pretty soundly on that. Any one using models as their go-to isn't worth anything.

But also, like, I can show you a post from 2022 when both Dakota Joshua and Kiefer Sherwood were in the AHL as Group VI UFA types where I called them as potential NHL players who I'd target as signings. Good luck showing me a model that did that.

Yeah and I was desperate for Ivan Barbashev and Artturi Lekhonen back in 2020. Years before they were traded or good and now are Stanley cup winners and first line players. I only get to dream.

It's not as if the guys with the models are saying they are 100% correct. They give ranges given history. Models would have said that Vasili Podkolzin was not going to be worth 10oa pick.

 

[Tables of Stats]

They give ranges given history. Models would have said that Vasili Podkolzin was not going to be worth 10oa pick.

Many people using traditional scouting methods would have told you the same thing. What's your point?

 

[MS]

Those graphic sheets by guys who won't share how their model works are pretty bad. That they are occasionally correct doesn't invalidate this critique. If you want to sway me, go do the leg work, look at these models you agree with see which picks they've made publically and show how they've tracked compared to what the model said they'd do.

Also, you're still raging at nothing, boxing at shadows, you've yet to put your money down to show anything. You can't seem to articulate what you've learned in your years studying these models, can't seem to articulate why these models are useful, can't seem to muster a word about why one model might be more robust or useful than another model. You're still just standing here stamping your feet insisting that models are good with the seeming justification for this being that sometimes models get it right where scouts get it wrong.

You're going to need to do the leg work. I suggest that you pick a few of your favourite models and show that they consistently beat actual NHL picks as well as lists put out by the media. If you do that, I'll have a reason to think that those specific models are actually useful. Until you can manage that, I'll stick to using methods of judging players and prospects that use the eye test, my own guess about personality and team/development program fit, and statistics to make my calls.

Like, I could 'create a model' that takes PPG of all draft picks and uses a multiplier of 1.0 for the KHL, .08 for other Euro leagues, 0.6 for NCAA, 0.4 for the CHL etc. and it would generate a draft list that wouldn't meaningfully be different from any of these 'models' their owners are touting. And I could point to big misses! And show where my model nailed an evaluation!

But it would still all just be totally useless. All it's telling you is that a) production is good, which everyone knows at this point, and b) some leagues are stronger than others, which everyone also knows. And it's doing so at the expense of literally any sort of context or eye test or understanding of skill sets. It's garbage. NHLe is garbage. Doing bulk player spreadsheets is garbage. Doing things that list Jordan Subban as a comparable to Kudryatsev is garbage. It literally tells you nothing outside of the most basic stuff that almost everyone already knows.

Yeah and I was desperate for Ivan Barbashev and Artturi Lekhonen back in 2020. Years before they were traded or good and now are Stanley cup winners and first line players. I only get to dream.

It's not as if the guys with the models are saying they are 100% correct. They give ranges given history. Models would have said that Vasili Podkolzin was not going to be worth 10oa pick.

Lots of traditional stuff would have told you the same thing. Again, 'guy who doesn't score much might not be great top-10 pick' isn't something revolutionary that you need a shitty model created by some dumbass in his mom's basement to understand.

 

[Tables of Stats]

Like, I could 'create a model' that takes PPG of all draft picks and uses a multiplier of 1.0 for the KHL, .08 for other Euro leagues, 0.6 for NCAA, 0.4 for the CHL etc. and it would generate a draft list that wouldn't meaningfully be different from any of these 'models' their owners are touting. And I could point to big misses! And show where my model nailed an evaluation!

But it would still all just be totally useless. All it's telling you is that a) production is good, which everyone knows at this point, and b) some leagues are stronger than others, which everyone also knows. And it's doing so at the expense of literally any sort of context or eye test or understanding of skill sets. It's garbage. NHLe is garbage. Doing bulk player spreadsheets is garbage. Doing things that list Jordan Subban as a comparable to Kudryatsev is garbage. It literally tells you nothing outside of the most basic stuff that almost everyone already knows.

I've made a model for doing my points predictions each season and it's crap, but I admit that it's crap and openly post exactly what my methods are, how my model is biased, and why it's at best going to tell you "good player expected to still be good next season". The issue with these sports model grifters is that they hide the guts of their models and don't adequately explain what they think the strengths and weaknesses of their models are.

 

[ManVanFan]

Many people using traditional scouting methods would have told you the same thing. What's your point?

Americans set to dominate first round of the NHL Draft - TSN.ca

While NHL scouts can and will weigh and measure the merits of Jack Hughes vs. Kaapo Kakko, there’s no debating this - a Canadian will not be the No. 1 overall pick in the NHL draft for an unprecedented fourth consecutive year.

www.tsn.ca

10 NHL scouts. 8oa average. Again, should we pretend that we are better than NHL scouting?

Those graphic sheets by guys who won't share how their model works are pretty bad. That they are occasionally correct doesn't invalidate this critique. If you want to sway me, go do the leg work, look at these models you agree with see which picks they've made publically and show how they've tracked compared to what the model said they'd do.

Also, you're still raging at nothing, boxing at shadows, you've yet to put your money down to show anything. You can't seem to articulate what you've learned in your years studying these models, can't seem to articulate why these models are useful, can't seem to muster a word about why one model might be more robust or useful than another model. You're still just standing here stamping your feet insisting that models are good with the seeming justification for this being that sometimes models get it right where scouts get it wrong.

You're going to need to do the leg work. I suggest that you pick a few of your favourite models and show that they consistently beat actual NHL picks as well as lists put out by the media. If you do that, I'll have a reason to think that those specific models are actually useful. Until you can manage that, I'll stick to using methods of judging players and prospects that use the eye test, my own guess about personality and team/development program fit, and statistics to make my calls.

I don't need to do anything. I made a post and you argued the validity of it. I tried to help the understanding of the post. If you don't like the answer, you got figure it out yourself. Lol.

 

[MS]

I've made a model for doing my points predictions each season and it's crap, but I admit that it's crap and openly post exactly what my methods are, how my model is biased, and why it's at best going to tell you "good player expected to still be good next season". The issue with these sports model grifters is that they hide the guts of their models and don't adequately explain what they think the strengths and weaknesses of their models are.

Yeah, I basically did this for THN Fantasy Pools when I was 13 years old. Guy in middle of career = continued expected production. Young guy = add 20%. Old guy = subtract 20%. Injury prone guy = subtract 20%.

It's not exactly genius stuff. But some guy with a blue tick calls it a 'model' and/or does it for prospects and posts confidently about it and people get dazzled.

 

[Tables of Stats]

Americans set to dominate first round of the NHL Draft - TSN.ca

While NHL scouts can and will weigh and measure the merits of Jack Hughes vs. Kaapo Kakko, there’s no debating this - a Canadian will not be the No. 1 overall pick in the NHL draft for an unprecedented fourth consecutive year.

www.tsn.ca

10 NHL scouts. 8oa average. Again, should we pretend that we are better than NHL scouting?

Wow! 10 whole scouts. That's not even 1 scout per NHL team, much less a representative sampling of the entire pool of NHL scouts. You're also still failing to show that these models work broadly. You're quick to argue a single point but fail to understand that we're not disputing that single point, but rather the usefulness of these models as a whole.

 

[ahmon]

Models are useful for fans who don't really follow prospects and just want a rough idea of who are producing well in their draft year.

IMO best scouting is done, when you prioritize on hockey sense, and then the rest.

Even the people who put a whole draft list probably wouldn't have been able to watch every single prospect play.

The most informative way would be to watch the prospect yourself and iso watch them shift to shift and see how they process the game.