Blue Bullet Report(Updated Apr 22: Expected Draft Value of a Forward Selected Top 90)

[Blue Bullet]

Damn, Landeskog is real low on that list. Galchenyuk is also really low, but I wonder if his injury skewed his NHLP.

Galchenyuk is all due to his injury as he had 4 points in 8 games.

Landeskog was rated higher than his numbers due to being NHL ready and for his intangibles.

 

[Blue Bullet]

In my latest post I try to establish benchmarks for NHLP as well as identify trends such as teams taking CHL forwards with less offensive potential too early in the first round.

 

[Blue Bullet]

In my latest post I look into CHL forwards drafted between 1990 to 2010 and what the average games played, points and points/game for each draft position between 1 through 30 can tell us. By re-ordering the draft positions of each year, based on a player's points/game in his draft season, I look into how teams are doing in selecting CHL forwards. In the end, it appears teams are doing a poor job in selecting CHL forwards in the correct order for the first round of the draft.

 

[I am the Liquor]

In my latest post I look into CHL forwards drafted between 1990 to 2010 and what the average games played, points and points/game for each draft position between 1 through 30 can tell us. By re-ordering the draft positions of each year, based on a player's points/game in his draft season, I look into how teams are doing in selecting CHL forwards. In the end, it appears teams are doing a poor job in selecting CHL forwards in the correct order for the first round of the draft.

That might be useful if points were the only measuring stick for what makes a good player.

 

[40oz]

That might be useful if points were the only measuring stick for what makes a good player.

True but it can be a good indicator

 

[I am the Liquor]

True but it can be a good indicator

This statement seems to be a contradiction.

 

[Jet Walters]

That might be useful if points were the only measuring stick for what makes a good player.

Very true.

Say Lawson Crouse and Matt Barzal make their respective teams next year and Crouse scores 16 goals and has 18 assists for 34 points and Barzal scores 10 with 30 a's for a total of 40. Does that mean Florida dropped the ball and should have taken Barzal? Hell no!! There is way more to the process of determining player worth than just point production.

 

[Blue Bullet]

That might be useful if points were the only measuring stick for what makes a good player.

It is however very useful. It does not paint the whole picture but gives a good starting point as a projection of player upside and there is lots of value to be gleemed simply based on point totals. For example there is 119 chl forwards drafted between 1998 and 2010 in the first round. 93 of those 119 players went on to play 100 nhl games with 50 of those 93 players being better than 0.5 career points/per game player. Using NHLP, 30 of these 50 players or 60% come from players with a NHLP over 66 which is the top 37 projections. Therefore 31% of the players with the best NHLP scores between 98 and 2010 account for 60 % of the players who have career points per game over 0.5

To use Barzal vs Crouse. Barzal projects to be a 71 point player. Between 98 and 2010, 92% (34/37) of forwards drafted in the first round with a 66 point NHLP or better went on to playing 100+ games with the lowest being Brule with 299 games. Of those with 100+ games, 88% (30/34) were 0.5 career point per game players or better.

Crouse projects to be a 56 point player. 72% (23/32) of forwards with a NHLP between 52 and 59 play 100+ nhl games. Of those only 30% (7/23) go on to be 0.5 career point per game forwards.

Therefore, based on past evidence Barzal has about an 81% chance of being a career 0.5 point per game player or better while Crouse has a 22% chance. Basically odds are there is a very good chance Barzal turns into a top 6 player while Crouse is more likely to be a 3rd line role player. I know which player I would want based on both the numbers and the eye test.

 

[Jet Walters]

It is however very useful. It does not paint the whole picture but gives a good starting point as a projection of player upside and there is lots of value to be gleemed simply based on point totals. For example there is 119 chl forwards drafted between 1998 and 2010 in the first round. 93 of those 119 players went on to play 100 nhl games with 50 of those 93 players being better than 0.5 career points/per game player. Using NHLP, 30 of these 50 players or 60% come from players with a NHLP over 66 which is the top 37 projections. Therefore 31% of the players with the best NHLP scores between 98 and 2010 account for 60 % of the players who have career points per game over 0.5

To use Barzal vs Crouse. Barzal projects to be a 71 point player. Between 98 and 2010, 92% (34/37) of forwards drafted in the first round with a 66 point NHLP or better went on to playing 100+ games with the lowest being Brule with 299 games. Of those with 100+ games, 88% (30/34) were 0.5 career point per game players or better.

Crouse projects to be a 56 point player. 72% (23/32) of forwards with a NHLP between 52 and 59 play 100+ nhl games. Of those only 30% (7/23) go on to be 0.5 career point per game forwards.

Therefore, based on past evidence Barzal has about an 81% chance of being a career 0.5 point per game player or better while Crouse has a 22% chance. Basically odds are there is a very good chance Barzal turns into a top 6 player while Crouse is more likely to be a 3rd line role player. I know which player I would want based on both the numbers and the eye test.

This is where the numbers diverge from reality. Despite what your numbers suggest I think if polled the majority of GM's in the league would prefer adding Crouse rather than Barzal if given the choice.

For the Oilers it would be a no brainer. A Crouse style of player would be a much better fit than a Barzal type given their current roster make up.

I feel just as a successful NHL club needs a variety of skill sets and player types on the roster, the prospect pool should be equally diverse with players that are suited for a wide range of roles.

 

[I am the Liquor]

It is however very useful. It does not paint the whole picture but gives a good starting point as a projection of player upside and there is lots of value to be gleemed simply based on point totals. For example there is 119 chl forwards drafted between 1998 and 2010 in the first round. 93 of those 119 players went on to play 100 nhl games with 50 of those 93 players being better than 0.5 career points/per game player. Using NHLP, 30 of these 50 players or 60% come from players with a NHLP over 66 which is the top 37 projections. Therefore 31% of the players with the best NHLP scores between 98 and 2010 account for 60 % of the players who have career points per game over 0.5

To use Barzal vs Crouse. Barzal projects to be a 71 point player. Between 98 and 2010, 92% (34/37) of forwards drafted in the first round with a 66 point NHLP or better went on to playing 100+ games with the lowest being Brule with 299 games. Of those with 100+ games, 88% (30/34) were 0.5 career point per game players or better.

Crouse projects to be a 56 point player. 72% (23/32) of forwards with a NHLP between 52 and 59 play 100+ nhl games. Of those only 30% (7/23) go on to be 0.5 career point per game forwards.

Therefore, based on past evidence Barzal has about an 81% chance of being a career 0.5 point per game player or better while Crouse has a 22% chance. Basically odds are there is a very good chance Barzal turns into a top 6 player while Crouse is more likely to be a 3rd line role player. I know which player I would want based on both the numbers and the eye test.

We can disagree on Crouse/Barzal, but what about goalies and dmen?

Your data is using very narrow parameters that only account for one part of the game. Therefore any conclusions drawn are flawed. Im not sure how that can be argued.

 

[Blue Bullet]

We can disagree on Crouse/Barzal, but what about goalies and dmen?

Your data is using very narrow parameters that only account for one part of the game. Therefore any conclusions drawn are flawed. Im not sure how that can be argued.

Yes when used in a bubble of course there is flaws to using just the numbers and I have never suggested that. What I am doing is trying to expose people to what statistical analysis can do to help elevate a team's drafting. When combined with scouting analysis it gives the best results. Using one alone is not the way to go and I believe most people would agree with that.

As for d-men that is a much tougher nut to crack as points are not as important for d-men as they are forwards. I have found zero evidence that junior production with d-men is a significant factor to NHL production. I will be re-ordering the draft with d-men similar to what I did forwards. Preliminary research suggests that taking the d-men with the better points/game in the first round results in a team getting a d-men that plays more career games and puts up more career points. However, of course scouting analysis is needed to help point one in the direction of which d-men are worth a 1st round pick. Once again this is not a one or the other situation as both scouting and stats are needed. I am trying to point people in the best direction on how to use the stats to add to their draft analysis.

As for goalies I have done some research on my drafting strategy based on historical evidence. Teams are good at evaluating who the top goalies are. For the most part never take a 1st round goalie. Have best value in 2nd round. Snatched up by mid 3rd round. If you miss out on those goalies take your top ranked Euro goalie in the last round. Best chance of getting a late round steal. As for stats on the situation that I have yet to figure out. As they say, goalies are voodoo.

http://bluebulletreport.com/2015/06/04/rethinking-how-to-draft-goalies/

 

[Blue Bullet]

This is where the numbers diverge from reality. Despite what your numbers suggest I think if polled the majority of GM's in the league would prefer adding Crouse rather than Barzal if given the choice.

For the Oilers it would be a no brainer. A Crouse style of player would be a much better fit than a Barzal type given their current roster make up.

I feel just as a successful NHL club needs a variety of skill sets and player types on the roster, the prospect pool should be equally diverse with players that are suited for a wide range of roles.

And that is where teams make mistakes as I suggested in my post. Rather than take the better player in Barzal you suggest taking Crouse due to being a better fit. This is how players like Eberle and Giroux are take in the bottom third of the draft. I'll take Barzal any day of the week over Crouse.

 

[I am the Liquor]

Good luck on your quest. I think you will have to find ways that incorporate the other facets of the game into your data. Then you may really have something.

 

[Jephman]

And that is where teams make mistakes as I suggested in my post. Rather than take the better player in Barzal you suggest taking Crouse due to being a better fit. This is how players like Eberle and Giroux are take in the bottom third of the draft. I'll take Barzal any day of the week over Crouse.

I really liked this Eye Test post. I think some people might misconstrue what you are conveying in the post (or perhaps I am) and interpret it as you saying teams should look only at PPG and use that as the success of an NHLer's career. From what I gather, what you are saying is that teams are taking on too much risk in an effort to find those intangible qualities like size and leadership.

Using your metrics, you say Barzal has an 82% chance of being a 0.5 PPG player. If a team was to pass on a guy like Barzal for a guy like Crouse who is 22% to reach that potential, that means that they are betting on the guy four times less likely to become a solid NHLer. For some GMs seeking size and skill, that 22% may just be worth it, but you are saying that this decision happens more often than the risks and rewards dictate it should.

As you and other posters have said, points per game is by no means a perfect evaluator. However, to me, this is a really interesting exploration.

 

[dustrock]

Good luck on your quest. I think you will have to find ways that incorporate the other facets of the game into your data. Then you may really have something.

It depends on whether or not you think these other abilities can be taught down the line. Haven't we been told a million times to always take scoring/skill, because it is the most rare ability?

BB seems to suggest GMs are ignoring that.

 

[I am the Liquor]

It depends on whether or not you think these other abilities can be taught down the line. Haven't we been told a million times to always take scoring/skill, because it is the most rare ability?

BB seems to suggest GMs are ignoring that.

I don't know how you propose to teach Barzal to be like Lucic. They are two different types of players with different skill sets. Both can be valuable. Only one is being recognized by BB's work.

BB seems to be suggesting that GM's should ignore all other facets of the game and just take the guy with the highest point totals.

 

[Jephman]

I don't know how you propose to teach Barzal to be like Lucic. They are two different types of players with different skill sets. Both can be valuable. Only one is being recognized by BB's work.

BB seems to be suggesting that GM's should ignore all other facets of the game and just take the guy with the highest point totals.

I think you are creating a strawman argument here. I do not believe he is saying this at all. Again, I could be wrong, but I interpret him as saying GMs are undervaluing points per game, and not that this should be the sole parameter of scouting.

Take for example Lawson Crouse and Mat Barzal. Crouse was taken a few picks earlier than Barzal. Let's now analyze these players in terms of expected return:

Crouse: as BB's data suggests, Crouse has about a 22% of becoming a 0.5PPG player in the NHL. With those numbers, he becomes a Lucic/Ladd type player, and I think he would be extremely valuable if he reached that potential. If he doesn't reach that potential, he could be a 3rd or 4th liner, or not make the show at all.

Barzal: the data shows he has about an 82% chance of becoming a 0.5PPG player. If he doesn't reach this, it's highly unlikely he makes the NHL. So some could see him as a more "boom or bust" prospect, but then again he has an 82% chance of booming, so it's a pretty good bet.

Given these numbers, Barzal is likely going to be a 2nd line NHLer, maybe a 1st liner. Crouse has a very good chance (about 80% chance) of not reaching his potential, and he is likely to settle in to a 3rd or 4th line role. In terms of expected return on the player, Barzal looks like he is going to be more valuable to his team in the future than Crouse will be.

For the record, I'm not saying picking Crouse is a bad call. If a GM assesses that his team is lacking size and grit, Crouse makes a lot of sense. However, Barzal is much more likely to be the more valuable asset in the future.

 

[Blue Bullet]

My latest post, at the bluebulletreport.com, the Magic Number is 66, is one of the most interesting peices I have done on my point prediction model, NHLP.

Similar to TSN's Scott Cullen's expected value of a draft pick article, I wanted to focus in on the percentage of CHL forwards, drafted between 1998 and 2010, that play over 100 games as well as those players that play 100+ games while having a 0.5+ Pts/G.

Therefore, by looking at where there are large drop offs in the percentage of forwards who meet the criteria of 100+ games and 0.5+ Pts/G occurs, I was able to separate the 119 CHL forwards drafted in the first round into four groups. Those groups are:

NHLP of 66+
NHLP of 60-65
NHLP of 52-59
NHLP of 51 and below

In reading the post you will see how each group has a different level of success in terms of reaching this criteria.

- 81%(30/37) of forwards with a NHLP of 66 or greater meet this criteria.
- 24%(20/82) of forwards with a NHLP of 65 or less meet this criteria.
- Therefore, the 37 forwards with a NHLP of over 66 make up 31% of the 119 first round forwards evaluated. However, they make up 60% of the forwards that reach 100+ games while having a 0.50+ Pts/G.

Finally, I sort the CHL forwards drafted between 2011 and 2015 into the four groups. See which players fall into what group and the likelihood of success based on historical data. For example, see why the odds are saying that forwards such as Meier (9), Barzal(16) and Svechnikov(19), from the 2015 draft, are much more likely of turning into top 6 forwards than Zacha(6), Crouse(11) or Senyshyn(15).

This post, I believe is very useful for draft fans and will help in utilizing the resource of NHLP to assist with your draft analysis. I have been a draft junkie for over a decade now and I know that I would have loved to have this knowledge about the likelihood of players reaching the benchmarks of 100+ games and 0.5+ Pts/G. Combining this knowledge, along with other tools such as scouting reports from draft guides, is a great way for the draft junkie to analyze the NHL draft. In my post next week I will be looking into where the scouts and stats collide and where they do not. Thanks for reading.

 

[Blue Bullet]

I think you are creating a strawman argument here. I do not believe he is saying this at all. Again, I could be wrong, but I interpret him as saying GMs are undervaluing points per game, and not that this should be the sole parameter of scouting.

Take for example Lawson Crouse and Mat Barzal. Crouse was taken a few picks earlier than Barzal. Let's now analyze these players in terms of expected return:

Crouse: as BB's data suggests, Crouse has about a 22% of becoming a 0.5PPG player in the NHL. With those numbers, he becomes a Lucic/Ladd type player, and I think he would be extremely valuable if he reached that potential. If he doesn't reach that potential, he could be a 3rd or 4th liner, or not make the show at all.

Barzal: the data shows he has about an 82% chance of becoming a 0.5PPG player. If he doesn't reach this, it's highly unlikely he makes the NHL. So some could see him as a more "boom or bust" prospect, but then again he has an 82% chance of booming, so it's a pretty good bet.

Given these numbers, Barzal is likely going to be a 2nd line NHLer, maybe a 1st liner. Crouse has a very good chance (about 80% chance) of not reaching his potential, and he is likely to settle in to a 3rd or 4th line role. In terms of expected return on the player, Barzal looks like he is going to be more valuable to his team in the future than Crouse will be.

For the record, I'm not saying picking Crouse is a bad call. If a GM assesses that his team is lacking size and grit, Crouse makes a lot of sense. However, Barzal is much more likely to be the more valuable asset in the future.

You absolutely get what I am trying to do. I think you will really enjoy my new post.

 

[Kerricthebig]

Great read as always BB!

Small table entry corection: Connor McDavid doesn't have 108 GP and 26 Pts

 

[Blue Bullet]

In my latest post at the bluebulletreport.com I look at where scouts and stats align when selecting CHL forwards in the first round. I also look at what occurs when they do not. See how a top 7 selection with a NHLP of 66 or greater, has about double the chance of playing at least 100 NHL games while maintaining a 0.5+ Pts/G in his career, than compared with a top 7 selection with a NHLP of 65 or less. When scouts and stats collide the results are for the best.

 

[Aceboogie]

Very informative. Do you have a list of players in this draft who are projected top 10-15 picks youd stay away from using this model?

 

[Aceboogie]

I don't know how you propose to teach Barzal to be like Lucic. They are two different types of players with different skill sets. Both can be valuable. Only one is being recognized by BB's work.

BB seems to be suggesting that GM's should ignore all other facets of the game and just take the guy with the highest point totals.

I think his work is from a historical statistical stand point. Sure there is more to hockey than just points, and Crouse brings some different things. But if you look at historical examples and players like Crouse, they tend to not work out based on draft position.

And when it comes down to it, points in junior mean a hell of alot. More often than not, players with high point totals have them because they are extremely creative offensively and are able to do other "intangible" stuff that result in points.

If you look at the past, not many low scoring junior players drafted high tend to ever make it to anywhere near draft "potential" whereas high scoring forwards do. This applies even more to low scoring CHL D. If you are under 35 points in junior for a D, you probably arent making it for. Points for a D show they can transition the puck, handle the puck, find teammates etc.

There was a article once that compared Vancouvers drafting history from 2000-2009 to a model where the highest scoring CHL forward available was drafted. The models roster was incredibly better than Vancouvers drafted roster and actually assembled a well-balanced team thru luck

 

[Aceboogie]

The draft is a funny place. I believe it can be taken advantage of by a smart team using statistical models and advanced stats. Total points his used here by BB, but I think you could drill down to even strength points and further by playing time and really find the high offense players.

Model could also incorporate SEL/KHL/FEL leagues where TOI is super distorted

Also, there is several work done showing the actual value of a 3rd rounder and a 7th rounder is insignificant. You can essentially trade back every draft (like we did in 2013) and trade a 3rd for 4th + 6th, then that 4th for 5th + 7th and wind up with 3 picks instead of 1 and be much better off

 

[Blue Bullet]

Very informative. Do you have a list of players in this draft who are projected top 10-15 picks youd stay away from using this model?

In some future posts I will be breaking down the draft into different groups based on odds of success. I will than apply to 2016.

One player the model is not fond of is Max Jones. No player drafted between 98 and 2010 in the first round with his stats went on to have a career average of 0.5+ PPG. He is almost a sure bet to be a bottom six player.