Prospect Info: Can the Canes find another Pesce/Slavin

[emptyNedder]

The Canes drafting at the end of rounds could mean the organization might have to "reach" on a few picks. A good example is Samuel Bolduc. Most mock drafts/rankings I have seen have him going around 70-75. But I could see the Canes liking him enough to use pick 60.
He is one of the bigger left-shot D-men available (6'4", 210). He tested really well at the CHL top-prospects showcase as he was the fastest backward skater with the puck and also placed in the top 3 in several off-ice tests.
Central Scouting ranked him 42nd for North American skaters. Bolduc's PSF is .92. That is likely a little low due to his team being weak. Given his size and testing, I would tweak the model and put Bolduc at 1.1.
https://hfboards.mandatory.com/threads/ld-samuel-bolduc

 

[Finlandia WOAT]

If you manage to create a model which can ascribe some predictive value to the success rate of draftees with repeatable success against League average/ESPN draft bot, you're looking at an automatic 5 figure job offer from a bunch of teams. And that's a conseevative estimate because the draft is such a crap shoot past the first...again conservative estimate, 20 picks? Maybe 40?

This is the time for your son to sit down with a case of red bull, pop open as many calc and stat courses he can find on youtube, and go nuts. The salary cap is already stagnating, we're due for economic diminution in the near future...if teams can't throw money at players, they'll throw money at nerds who can count, and the guys who'll get the call are nerds who have some magic formula, so as to ensure the team investment represents some form of asset control.

Scouting services also hire entry level jobs. There was one that was advertising on here. No math, they give you a crash course on what they're paying you to count, then you watch hockey 8 hours a day 6 days a week.

The Data Wars are coming. Let us hope General Tulsky proves an effective war time leader.

You really want to get ahead of the curve? This sh!t'll be automated in 20 years. You know what won't be? Character evaluation. Reading people. Study psych, body language, cold reading... then if a career of evaluating the psychological profile of prospects doesn't work out, you can go do cia psyops.

 

[emptyNedder]

You really want to get ahead of the curve? This sh!t'll be automated in 20 years. You know what won't be? Character evaluation. Reading people. Study psych, body language, cold reading...

First--thanks for the post. Second--I actually think the insight that PIM is important gets at some part of Character evaluation. It is mostly about positioning and fundamentals, but using Pesce and Slavin as examples you just don't see either of them out of control on the ice, which I think is due to who they fundamentally are.

And that's a conseevative estimate because the draft is such a crap shoot past the first...again conservative estimate, 20 picks? Maybe 40?

I would say the first two rounds. Because I think scouts can reasonably pay enough attention to the players in that grouping.

 

[DaveG]

First--thanks for the post. Second--I actually think the insight that PIM is important gets at some part of Character evaluation. It is mostly about positioning and fundamentals, but using Pesce and Slavin as examples you just don't see either of them out of control on the ice, which I think is due to who they fundamentally are.

I would say the first two rounds. Because I think scouts can reasonably pay enough attention to the players in that grouping.

PIM is just... Ehhhh, I'll get to it more when I get home but it is not a very useful tool in my books.

 

[emptyNedder]

PIM is just... Ehhhh, I'll get to it more when I get home but it is not a very useful tool in my books.

Dave. I understand your point of view. I do risk analysis for a living. When models began to be incorporated most of us "knew" that several of the components that the models used weren't useful. All the traditionalists in my industry scoffed. More than a decade later the models have greatly improved performance.

I will state again that by using PIM the model will have some obvious false positives (Radko Gudas, John Klingberg, and Dalton Prout in 2010; Erik Gustafsson and Colin Miller in 2012).

So I am not arguing that using PIM identifies all D-men from the middle and late rounds that will succeed. However, it does greatly improve the probability of identifying them. For instance in 2012 4 D-men were identified (as I mentioned above). Three turned out to be Lindell, Gostisbehere, and Slavin). It would have missed Parayko in the third round. But getting the three players mentioned is a definite win--all possible because of PIM.

PIM is not an indicator per se. However, when combined with other data it greatly improves the outcome.

 

[GoldiFox]

Dave. I understand your point of view. I do risk analysis for a living. When models began to be incorporated most of us "knew" that several of the components that the models used weren't useful. All the traditionalists in my industry scoffed. More than a decade later the models have greatly improved performance.

I will state again that by using PIM the model will have some obvious false positives (Radko Gudas, John Klingberg, and Dalton Prout in 2010; Erik Gustafsson and Colin Miller in 2012).

So I am not arguing that using PIM identifies all D-men from the middle and late rounds that will succeed. However, it does greatly improve the probability of identifying them. For instance in 2012 4 D-men were identified (as I mentioned above). Three turned out to be Lindell, Gostisbehere, and Slavin). It would have missed Parayko in the third round. But getting the three players mentioned is a definite win--all possible because of PIM.

PIM is not an indicator per se. However, when combined with other data it greatly improves the outcome.

Do you have actual TOI for those leagues or using GP?

Lindell played in the Finnish Pro leagues in his draft year and Gostibehere played in the NCAA. They both have lowish PIM but they also probably played far, far less overall minutes than Parayko who in the AJHL. It could be in the extremes of Lindell averaging 5 minutes a game in the SM-Liiga versus Parayko averaging 30 minutes a game in the AJHL.

 

[emptyNedder]

Do you have actual TOI for those leagues or using GP?

Lindell played in the Finnish Pro leagues in his draft year and Gostibehere played in the NCAA. They both have lowish PIM but they also probably played far, far less overall minutes than Parayko who in the AJHL. It could be in the extremes of Lindell averaging 5 minutes a game in the SM-Liiga versus Parayko averaging 30 minutes a game in the AJHL.

I don't have TOI--wish I could get it. But I do adjust the formula based on league as junior leagues should produce much more scoring for D-men while a 17/18-year-old isn't likely to put up many points in a professional league or the NCAA where many opponents are 22-25.

 

[emptyNedder]

I really appreciate all the feedback and constructive criticism. The model is evolving. I think FWOAT is correct when he states "some predictive value." I spent the decade before hockey blogging debating economic models in the comments sections of blogs written by the likes of Paul Krugman, Mark Thoma, and James Kwak. So I expect push back. I also know that the best/most complex model is still only partially predictive.

Keep the comments coming. Y'all are great.

 

[Navin R Slavin]

You really want to get ahead of the curve? This sh!t'll be automated in 20 years. You know what won't be? Character evaluation. Reading people. Study psych, body language, cold reading...

o rly?

How AI Will Revolutionize How We Use Body Language | Science of People

 

[Finlandia WOAT]

o rly?

How AI Will Revolutionize How We Use Body Language | Science of People

In 20 years?

(Thanks for the link)

 

[MinJaBen]

o rly?

How AI Will Revolutionize How We Use Body Language | Science of People

To really get ahead of the curve, study machine learning.

 

[emptyNedder]

To really get ahead of the curve, study machine learning.

But if you are studying what machines are learning won't the machines improve and leave you permanently behind the curve

 

[MinJaBen]

But if you are studying what machines are learning won't the machines improve and leave you permanently behind the curve

Not in our lifetimes, I think.

 

[DaveG]

Dave. I understand your point of view. I do risk analysis for a living. When models began to be incorporated most of us "knew" that several of the components that the models used weren't useful. All the traditionalists in my industry scoffed. More than a decade later the models have greatly improved performance.

I will state again that by using PIM the model will have some obvious false positives (Radko Gudas, John Klingberg, and Dalton Prout in 2010; Erik Gustafsson and Colin Miller in 2012).

So I am not arguing that using PIM identifies all D-men from the middle and late rounds that will succeed. However, it does greatly improve the probability of identifying them. For instance in 2012 4 D-men were identified (as I mentioned above). Three turned out to be Lindell, Gostisbehere, and Slavin). It would have missed Parayko in the third round. But getting the three players mentioned is a definite win--all possible because of PIM.

PIM is not an indicator per se. However, when combined with other data it greatly improves the outcome.

So, to get in to it, I'm definitely one of those guys that's in the middle ground on statistical analysis. I'm not going to critique your model simply because 1 - I haven't seen how exactly it gets from point A to point B, but it looks like it has solid grounding overall, and B - while I do lean toward traditional methodology I also lean heavily to analytics being a great basis to pare things down significantly, to help us figure out who is actually worth paying attention to. It's something I tried to do a good bit in the past and just simply did not have the time to really build a quality model of my own that would give solid valuations. In part because of lack of available data at the time (this was mostly around 2010-11) and in part due to simply lack of time to find a way to acquire the data I was missing.

PIMs are very much a gray area. Using the Parayko vs Slavin example, AJHL is typically a lower skilled level than the USHL that also typically has a few NCAA bound exceptional prospects a year (Parayko, Makar). Most importantly with PIMs you need to get information on what type they are, and that information is hard to come by for leagues lower than say the USHL level. There's only so much you can do balance out with a model with those leagues though due to the lack of information available.

 

[emptyNedder]

while I do lean toward traditional methodology I also lean heavily to analytics being a great basis to pare things down significantly, to help us figure out who is actually worth paying attention to.

As I mentioned in the draft post my son often understands things that I don't. When I mentioned that Anton posted about finding another Pesce/Slavin in the middle rounds he went to work. He looked at things for a few minutes and said: "They don't take penalties." I said something like "Duh. They play sound hockey. If we can figure which 17-year-olds will play sound hockey in 5 years, then we got something." He gave me the look he often gives me and said: "They don't take penalties."
Because he is my son and I have experienced these Rain Man moments several other times, I decided to give his idea a test run. He suggested the outline for the formula and I have tried to refine it based on how it would have predicted past drafts. It sounds crazy, but using penalty minutes creates a model that hones in on a certain type of prospect who is more successful than his draft position would suggest.
I admit that the model might not be doing what I think. Unfortunately, we won't know for 4-5 years.

 

[MinJaBen]

As I mentioned in the draft post my son often understands things that I don't. When I mentioned that Anton posted about finding another Pesce/Slavin in the middle rounds he went to work. He looked at things for a few minutes and said: "They don't take penalties." I said something like "Duh. They play sound hockey. If we can figure which 17-year-olds will play sound hockey in 5 years, then we got something." He gave me the look he often gives me and said: "They don't take penalties."
Because he is my son and I have experienced these Rain Man moments several other times, I decided to give his idea a test run. He suggested the outline for the formula and I have tried to refine it based on how it would have predicted past drafts. It sounds crazy, but using penalty minutes creates a model that hones in on a certain type of prospect who is more successful than his draft position would suggest.
I admit that the model might not be doing what I think. Unfortunately, we won't know for 4-5 years.

You’ve probably said too much...

 

[emptyNedder]

After the first 75 picks or so it is difficult to determine draft order.
Two defensemen who could be available in the 4th or 5th round that my model likes are:

Kalle Loponen--a smallish (5'10", 186) right-shot who has played for Finland in several international tournaments and produced offense at every junior level. Central Scouting ranks him 64th among European skaters. His PSF is 1.27.

Mattias Norlinder--Central Scouting really likes this player as he moved from 47th in its mid-term rankings to 25th in its final rankings. However, most mock drafts have him going in the 4th or 5th round. He is a left-shot dman with average size (6'0", 180). He too excels at producing points from the back-end. Norlinder's PSF is 1.13.

 

[emptyNedder]

IMO, it also goes to show how big a mistake it is to take d-men in the Top 10, unless they are "franchise" type guys.

I hope the organization:

1. Sees the merit in this line of thinking;
2. Has someone lurking on these boards.

There are three additional prospects identified by PSF. Based on most rankings and mock drafts they should be available when the Canes have two picks in the 6th round and one in the 7th.

Alexander Lundqvist--PSF 1.44. He has decent size (6'2", 183) and is a left shot. I couldn't find much in the way of scouting reports and no video. NHL Central Scouting has him ranked 59 among European skaters. Niklas Lidstrom is his uncle.

Grayson Ladd--PSF 1.72. A right shot, average size (6'1" 172). https://hfboards.mandatory.com/threads/d-grayson-ladd
Again not a lot of easily accessible information--but what stood out was his plus/minus. While this is usually an unreliable statistic, it was specifically mentioned in an article about the factors the Canes noticed about Slavin. Like Slavin, Ladd had an excellent plus/minus on a team where the majority of players were minus.

Thomas Pelletier--PSF 1.40. Pelletier is only worth consideration in the 7th round as I haven't seen his name in rankings/mock drafts. However, he is one of the younger players available and is a coveted right-shot with good size (6'3", 192).

 

[GoldiFox]

Canes interviewing a Defenseman ranked in the late 1st or early 2nd. What does your PSF say about Ryan Johnson? The scouting reports I find sound... familiar.

What immediately stands out is his skating. He’s an effortless skater who can evade checks and generate clean zone exits with his feet. He’s not a blazing fast skater, but his edges and first step are high end.

Defensively he’s not the biggest guy, but he’s so solid due to his smarts and mobility. His gap control is very good and often closes on guys quicker than they anticipate.

From here Johnson adds in excellent footwork, strong agility, edgework, and pivots. This gives Johnson the ability to transition quickly from offence to defence or vice-versa. Johnson is able to pinch in the offensive zone and still get back defensively to break up plays in his own end.

Johnson’s game is based on his skating ability and smarts. He can carry the puck out of dangerous areas in his own end, avoiding forecheckers and starting the rush. He is also very good at generating clean zone entries with his smarts and quick feet. His skating and stickhandling skills allow Johnson to generate offence off the rush.

Johnson is a very smart player and this helps him to play strong defence. He keeps himself between his man and the net, forcing opponents to the outside and into less dangerous areas of the ice. Johnson has a very quick stick. He is very good at poke-checking or stealing the puck away from an opponent. His positioning is very well-developed for his age as Johnson takes away passing lanes. Johnson takes good angles on his opponents, effectively taking away their time and space.

Effortless 4-way skater known for his gap control, an active stick, and his defensive awareness in shutting down passing lanes.

 

[Nikishin Go Boom]

Canes interviewing a Defenseman ranked in the late 1st or early 2nd. What does your PSF say about Ryan Johnson? The scouting reports I find sound... familiar.







Effortless 4-way skater known for his gap control, an active stick, and his defensive awareness in shutting down passing lanes.

If we get some speedy forwards with our late first and one of the other early 2nds, I would do Johnson. Seems a lot like Martin who didnt pan out. Martin killing batting 1.000 with our 2nd round picks we had going since 2010.

 

[Cardiac Jerks]

How has Martin not planned out exactly? Seems a bit early.

 

[DaveG]

How has Martin not planned out exactly? Seems a bit early.

Seriously.

I mean until this year Eetu was our most widely panned pick since Paradis. It's hard to tell much based on the lack of numerical projection on what's going on with him at Michigan. Part of the reason I want him out of there and in with the checkers. The other part obviously being we need the depth.

 

[Nikishin Go Boom]

How has Martin not planned out exactly? Seems a bit early.

His play time has taken an absolute nose dive this season. Of the NCAA Juniors this year that the Canes could sign, his name wasnt even mentioned as needing to make a decision to return for his senior year or take a contract. Good sign the team isnt considering him for a contract.

 

[WreckingCrew]

Anyone have any info / insight on Ronnie Attard? Maybe he's a late bloomer or maybe it's because he's older (20 y/o) than much of his competition, but he lit it up from the blueline this year in the USHL. 6'3" RHD got 30G / 64PTS and a +46 this year for Tri-City...after last year having only 8G / 15PTS / -9 and LOTS of penalties.

 

[emptyNedder]

Canes interviewing a Defenseman ranked in the late 1st or early 2nd. What does your PSF say about Ryan Johnson? The scouting reports I find sound... familiar.

I have seen shift-by-shift video of Johnson. He looks great. If the Canes go for a D-man with either 36 or 37, Johnson would be a solid pick.

I didn't have a PSF for Johnson. I don't think it adds much value for prospects who are consensus top 50 because those players get scouted a good deal so there aren't likely to be hidden gems. Johnson's PSF is .945.

Again, I believe PSF's real value is in identifying those D-men ranked somewhere between 75-200. For those I have highlighted in this thread I think three: Jones, Lundmark, and Bolduc are probably on most teams' draft board for the end of round 2 or early in round 3. In their case I think PSF is merely separating them from the other 6-8 defensemen who are similarly ranked. The most value comes from the model indicating that Siedem is worth a 3rd round pick, Lundqvist is worth a 5th or 6th, and Ladd should be drafted at all.