HF Habs: Scouting Discussion Thread

[Mrb1p]

Yeah, that's kinda my problem with it. Taking WTK's trading back example, just because you made some tier doesn't mean that, by trading back, you'll be left with only one of them. You could be left with two, and where's your tier now?

I don't see what I'm gaining by participating in the exercise of separating my list into tiers. How do we establish what a good tier even is? For example, does it mean that the worst player in the tier ahead is better than the best player in the tier below? Well, if you had a perfect ranking, any partition would be a valid ''tier.''

I think your lost as to what were doing here. Were dealing in uncertainty, no list is perfect, but it is a good way to establish value.

 

[NotProkofievian]

I think your lost as to what were doing here. Were dealing in uncertainty, no list is perfect, but it is a good way to establish value.

No, I just don't think you're really doing anything if you can't give a precise definition for what a tier is and how you establish which players go into which tier. And if you can do that, why not just do it for every player?

 

[WeThreeKings]

What does a tier even mean? How do you define a tier? IMO, this is just a common sense idea that's been applied to drafting without actually being precisely defined. I don't see the purpose. You still only get to pick one guy. You don't get to submit a selection like ''player x, but I'd also be fine with these other 3 guys.''

I define a tier by the ultimate upside of the player.

You only get to take one player but again it's when evaluating a trade up or down scenario. Have to know how far down you are willing to go before the you think you are no longer getting who your think has top 6 upside.

 

[NotProkofievian]

I define a tier by the ultimate upside of the player.

You only get to take one player but again it's when evaluating a trade up or down scenario. Have to know how far down you are willing to go before the you think you are no longer getting who your think has top 6 upside.

Okay, you trade down and no one in your ''tier'' was selected, now what do?

 

[Mrb1p]

No, I just don't think you're really doing anything if you can't give a precise definition for what a tier is and how you establish which players go into which tier. And if you can do that, why not just do it for every player?

It's just establishing draft value.
I see a huge difference between Kakko and Hughes vs Turcotte and Byram vs Dach and Cozens and a few others.

As WTK said, it's good to evaluate moving in the draft.

 

[Mrb1p]

Okay, you trade down and no one in your ''tier'' was selected, now what do?

You draft the best player on your top tier ? I don't understand the confusion.

 

[ahmedou]

For me, as an organized guy the tier stuff makes me think of many scenarios, order, plans,etc. But everyone have their own methods. Our common points, it's the eyes test. That never lies...

 

[NotProkofievian]

You draft the best player on your top tier ? I don't understand the confusion.

Just assign values to the individual players then. I don't understand the purpose of a tier. It logically operates on top of a list and doesn't affect its order, so it can't possibly create value for your list.

 

[atrud66]

Just assign values to the individual players then. I don't understand the purpose of a tier. It logically operates on top of a list and doesn't affect its order, so it can't possibly create value for your list.

When a professor gives grades to their students, they can sort the students by their average in the class. Students that are adjacent in the list can be given different grades while maintaining their order. Student 9 in the sorted list may have a 94% average and receive an A while student 10 in the list may have a 88% average and receive a B+. Even though the students are in the proper order, you lose information by simply saying student 10 is the next best student following student 9 because there's clearly a significant difference in their ability. Classifying them into their grades (tiers) gives more information.

 

[WeThreeKings]

Okay, you trade down and no one in your ''tier'' was selected, now what do?

Trade down again, if you have a buyer. If not, then take the highest ranked guy in that tier.

I don't put a whole lot of work into the tiers, the ranking matters most there's definitely value in that when you are on the clock on the draft floor.

There's still rankings within each individual tier, you're just saying there's not a lot separating these guys in terms of upside, they just have different strengths or style of play or positional value

 

[SOLR]

No, I just don't think you're really doing anything if you can't give a precise definition for what a tier is and how you establish which players go into which tier. And if you can do that, why not just do it for every player?

Tiers are indeed useless.

 

[NotProkofievian]

When a professor gives grades to their students, they can sort the students by their average in the class. Students that are adjacent in the list can be given different grades while maintaining their order. Student 9 in the sorted list may have a 94% average and receive an A while student 10 in the list may have a 88% average and receive a B+. Even though the students are in the proper order, you lose information by simply saying student 10 is the next best student following student 9 because there's clearly a significant difference in their ability. Classifying them into their grades (tiers) gives more information.

No, not more information. You already have the test score.

edit: in fact, you lose a lot of information in practice, this way. On your transcripts, you don't see individual test scores contextualized within a cohort, you just see a tier: A,B,C etc in highschool, or 1 through 4 in north america's universities, or 1 through 5 in Finnish universities, or 1 through 10 in Dutch universities. Everyone in every ''tier'' is exactly the same for someone from the outside.

 

[atrud66]

No, not more information. You already have the test score.

edit: in fact, you lose a lot of information in practice, this way. On your transcripts, you don't see individual test scores contextualized within a cohort, you just see a tier: A,B,C etc in highschool, or 1 through 4 in north america's universities, or 1 through 5 in Finnish universities, or 1 through 10 in Dutch universities. Everyone in every ''tier'' is exactly the same for someone from the outside.

I think the misunderstanding is that in a draft list, you don't often give a numerical score to each player like in the grade example I made

 

[ahmedou]

Is there schools for amateur scouts in Canada?

 

[Scriptor]

Is there schools for amateur scouts in Canada?

How about amateur schools for scouts?

 

[NotProkofievian]

I have posted lists, and anyone can string some names together. Also, please tell me more about your scouting process? How you grade etc, because it appears to me all you do is develop a mancrush, fap to him, and attack anyone who disagrees.

Not really. When your ''list'' is the first 10 names on Bobby's list, with maybe a name or two swapped, that's not your list.

As for my scouting process, the first step is actually watching games.

Watching a guy for however many shifts he can fit into 15 minutes you can generally assess the physical tools at the prospect's disposal, but not much else. The remainder of the time that I'm actually watching games is spent in an attempt to assess their understanding of the game. This actually takes many viewings and sometimes re-viewings to make sure of what you're seeing. Whereas most physical traits are considered on a ''yes/no'' basis, hockey I.Q. is what I place all my bets on.

And finally, I look for statistical trends that might indicate that the player is special. Usually this comes up negative, because most players aren't special. For example in Dorofeyev's case, the statistical signal is fairly weak: essentially it just preserves the chance that he is an elite player in the making, but confirms nothing. If I do find such a signal, I try to look for contexts which might be exaggerating it. Was the year dog-shit? Did he play in a weak division? Was he being carried by much older linemates? If I can't find that, and it jives with the eye test, the guy shoots way up the list, other rankings be damned.

Brink, I should say, has the most/strongest signals outside of the top 2 players. If it's a fluke, so be it. It won't always be, and when it hits, it's going to hit big.

 

[DramaticGloveSave]

Not really. When your ''list'' is the first 10 names on Bobby's list, with maybe a name or two swapped, that's not your list.

As for my scouting process, the first step is actually watching games.

Watching a guy for however many shifts he can fit into 15 minutes you can generally assess the physical tools at the prospect's disposal, but not much else. The remainder of the time that I'm actually watching games is spent in an attempt to assess their understanding of the game. This actually takes many viewings and sometimes re-viewings to make sure of what you're seeing. Whereas most physical traits are considered on a ''yes/no'' basis, hockey I.Q. is what I place all my bets on.

And finally, I look for statistical trends that might indicate that the player is special. Usually this comes up negative, because most players aren't special. For example in Dorofeyev's case, the statistical signal is fairly weak: essentially it just preserves the chance that he is an elite player in the making, but confirms nothing. If I do find such a signal, I try to look for contexts which might be exaggerating it. Was the year dog-****? Did he play in a weak division? Was he being carried by much older linemates? If I can't find that, and it jives with the eye test, the guy shoots way up the list, other rankings be damned.

Brink, I should say, has the most/strongest signals outside of the top 2 players. If it's a fluke, so be it. It won't always be, and when it hits, it's going to hit big.

But how do you grade the prospects? Do you even grade the prospects? Or are you just putting a list of names ranked by your feelings that you get when you watch them?

 

[Andrei79]

Can't be wrong when you don't publish a list!

I can't realistically publish a list anymore. I watch a number of these players, I can talk about specific players I've paid more attention to, mostly because they've caught my eye or I'm unsure why theyre getting hype/hate. I can make direct comparisons but not the number needed to make a coherent top 60. I just try to get a hold of a few favorites and try to nuance what will help/limit their chances of making the NHL.

Players who I think cant be top6ers or more than bottom pairing I dont even bother with.

 

[NotProkofievian]

I can't realistically publish a list anymore. I watch a number of these players, I can talk about specific players I've paid more attention to, mostly because they've caught my eye or I'm unsure why theyre getting hype/hate. I can make direct comparisons but not the number needed to make a coherent top 60. I just try to get a hold of a few favorites and try to nuance what will help/limit their chances of making the NHL.

Players who I think cant be top6ers or more than bottom pairing I dont even bother with.

Ideally, you wouldn't have more than about 40 players on your list. Because that's about the number that will be worth the picks you used for them after it's all said and done.

 

[NotProkofievian]

But how do you grade the prospects? Do you even grade the prospects? Or are you just putting a list of names ranked by your feelings that you get when you watch them?

To rank N players, you need to make N(N-1)/2 comparisons. I try to do as many of those direct comparisons as possible. For each time I make a comparison I make a note (player a > player b). If necessary I have to make indirect player comparisons (player a > player b and player b > player c implies that player a > player c). To borrow Bergy's parlance, at de end of dedé that's all you can really do. You can say 1-5 to a bunch of players on a bunch of made up categories but that's not different than just saying ''player a > player b.''

 

[Mrb1p]

Ideally, you wouldn't have more than about 40 players on your list. Because that's about the number that will be worth the picks you used for them after it's all said and done.

Its impossible to have such a list. Its kinda lazy, youd just be ranking the 40ish best players in the draft. Theres risk assesment and development curve prediction that makes picks worth it later or not.

A guy like Dom Fensore is a good example. Would you have him in the top 40? Thats silly. Kid has good upside, but Im not picking him that high anyway.

 

[NotProkofievian]

Its impossible to have such a list. Its kinda lazy, youd just be ranking the 40ish best players in the draft. Theres risk assesment and development curve prediction that makes picks worth it later or not.

A guy like Dom Fensore is a good example. Would you have him in the top 40? Thats silly. Kid has good upside, but Im not picking him that high anyway.

There's two ways you could go about it. Let's say you had ''the perfect list'' given to you by a time traveler, and it just had busts marked with red, and non-busts marked with green. And then you used those markings on all the consensus lists floating around the internet. What you'd see is that there'd be a lot of green towards one end of the list, and some red mixed in here and there, and the red would get more and more prevalent until that's all there is except the odd green blip. You could try and rescue all the green blips from late in the draft, or you could remove the red blips from the front of the draft. The ideal list would do a bit of both, which is what I try to do.

Time saving would prioritize removing reds from the front of the list, though.

 

[DramaticGloveSave]

To rank N players, you need to make N(N-1)/2 comparisons. I try to do as many of those direct comparisons as possible. For each time I make a comparison I make a note (player a > player b). If necessary I have to make indirect player comparisons (player a > player b and player b > player c implies that player a > player c). To borrow Bergy's parlance, at de end of dedé that's all you can really do. You can say 1-5 to a bunch of players on a bunch of made up categories but that's not different than just saying ''player a > player b.''

So you don’t grade specific traits?

 

[NotProkofievian]

So you don’t grade specific traits?

I don't give them specific grades, but I do evaluate all of their traits. Like, take that scouching guy: I don't do what he does with his 1-5 A-D system. Why should I do that?

 

[DramaticGloveSave]

I don't give them specific grades, but I do evaluate all of their traits. Like, take that scouching guy: I don't do what he does with his 1-5 A-D system. Why should I do that?

Having a grading system is beneficial to the process. Every scout on the planet uses a variation of one. Not having one is essentially just ranking based on feeling.