The Advanced Stats Thread Episode VIII:

[Irishguy42]

You guys, it's very clear that it's hard for me to explain what I'm thinking.

This is the last thing I'll say. Think about it this way....

Let's say the Rangers are on the clock at the 2017 draft. They have the 7th OA, as they did, and they are fielding calls from other teams who are interested in trading for the pick. Think about what these teams are offering for the pick. The Rangers decide not to trade the pick, they select Lias Andersson. Now, let's say in this hypothetical world, that the Rangers start fielding calls on Lias Andersson and have him on the block. Are the teams offering the same package just because Lias was picked 7th overall?

No. Because the player and the pick have different values. Separate the player and the pick.

Truth.

Once a player is selected, that pick loses all value as the "_____ overall" pick because teams will have the player selected at that point rated differently on their draft board. Unless it's like...in the Top 3, but even then (Dubois/Puljujarvi)...

High picks are only rated highly because of what teams perceive to be available at that point. Outside of first few picks, draft pick value is almost always about expectations of availability. Not who is actually selected.

I think your example works better with Barrett Hayton/5OA from this past draft instead of Lias/7OA.

 

[Thirty One]

If you think it’s a bad pick at the time of the pick, you don’t get to change your opinion.

Watch me.

I’m still not over the Chytil was a bad pick thing, and I don’t think it was adequately explained.

My feelings are that it wasn’t a thrilling pick to me at the time, but I accept that the Rangers, who employ far more scouts than I employ, may have uncovered something proprietary that the consensus glossed over. Does anyone object to this stance?

 

[Ail]

Watch me.

I’m still not over the Chytil was a bad pick thing, and I don’t think it was adequately explained.

My feelings are that it wasn’t a thrilling pick to me at the time, but I accept that the Rangers, who employ far more scouts than I employ, may have uncovered something proprietary that the consensus glossed over. Does anyone object to this stance?

I object.

 

[Machinehead]

Watch me.

I’m still not over the Chytil was a bad pick thing, and I don’t think it was adequately explained.

My feelings are that it wasn’t a thrilling pick to me at the time, but I accept that the Rangers, who employ far more scouts than I employ, may have uncovered something proprietary that the consensus glossed over. Does anyone object to this stance?

I object to the notion that the New York Rangers have anything proprietary when it comes to drafting based on their history of shitting themselves.

 

[Thirty One]

I object to the notion that the New York Rangers have anything proprietary when it comes to drafting based on their history of ****ting themselves.

Objection noted.

 

[silverfish]

I’m still not over the Chytil was a bad pick thing, and I don’t think it was adequately explained.

There were two things that I really liked about the Chytil pick. The Rangers drafted for upside, and how young he was at the draft. However, I personally felt like there were better players on the board when this pick was made. Thus, I didn't love the pick. But again, we fall into the trap that I'm using absolutely no insights from what has happened since the draft to judge the pick. And I think that's where people are getting most held up by my, admittedly very different, evaluation process. And that's totally cool.

 

[Thirty One]

There were two things that I really liked about the Chytil pick. The Rangers drafted for upside, and how young he was at the draft. However, I personally felt like there were better players on the board when this pick was made. Thus, I didn't love the pick. But again, we fall into the trap that I'm using absolutely no insights from what has happened since the draft to judge the pick. And I think that's where people are getting most held up by my, admittedly very different, evaluation process. And that's totally cool.

Yeah, I definitely get and support evaluating a decision based on the information available at the time. In this case you and the Rangers had different information available to you at the time.

 

[SA16]

Well the question is why would the information the Rangers have be different than the information other teams might have and if they are far off the consensus why should we believe that the Rangers information is more likely to be accurate?

 

[sbjnyc]

Watch me.

I’m still not over the Chytil was a bad pick thing, and I don’t think it was adequately explained.

My feelings are that it wasn’t a thrilling pick to me at the time, but I accept that the Rangers, who employ far more scouts than I employ, may have uncovered something proprietary that the consensus glossed over. Does anyone object to this stance?

It's not just information from scouting but information about other teams' draft boards. On these forums people always seem to assume the rangers can find a partner to draft up or down at some set value and that all other teams are drafting exactly according to the consensus of a handful of self-proclaimed draft experts on social media.

 

[Uncle Dru]

It's not just information from scouting but information about other teams' draft boards. On these forums people always seem to assume the rangers can find a partner to draft up or down at some set value and that all other teams are drafting exactly according to the consensus of a handful of self-proclaimed draft experts on social media.

100% agree. I'm willing to bet outside of the first few picks team boards are all over the place. I think we underestimate how difficult it is to evaluate and compare 18 year olds from so many different leagues playing around the globe. There's more than enough disagreement when ranking NHL talent...

 

[GeorgeKaplan]

Well the question is why would the information the Rangers have be different than the information other teams might have and if they are far off the consensus why should we believe that the Rangers information is more likely to be accurate?

Is the consensus really the consensus though? The network lists tend to favor North American kids because they don’t send people to watch random games in Russia for one kid like teams will do. They also tend to have less people’s input into actually making their list. McKenzie’s list always tends to be the closest to the actual draft because he gets input from teams scouts

 

[silverfish]

Is the consensus really the consensus though? The network lists tend to favor North American kids because they don’t send people to watch random games in Russia for one kid like teams will do. They also tend to have less people’s input into actually making their list. McKenzie’s list always tends to be the closest to the actual draft because he gets input from teams scouts

It's essentially a crapshoot after the top-10, and that might be too generous. Maybe after the top-3. That's why models, like NHLe, are important.

 

[GeorgeKaplan]

It's essentially a crapshoot after the top-10, and that might be too generous. Maybe after the top-3. That's why models, like NHLe, are important.

I’m definitely more in line with this line of thinking, and to expand I think it’s better to think of draft rankings in terms of tiers rather than a hard line between individual numbers. I’ve said it a few times in different threads, but right now I think the draft is still more art than science

 

[Thirty One]

Well the question is why would the information the Rangers have be different than the information other teams might have and if they are far off the consensus why should we believe that the Rangers information is more likely to be accurate?

They take in the same or sometimes different inputs and form different opinions than the other teams do. I’m not making the argument that they do it better, only that the should be credited when the results suggest theirs might have some merit over the consensus. When the opposite is true, they deserve blame.

 

[Tob]

I make lists of the most attractive players and evaluate the overall draft based on it. We did pretty well this year imo. Chytil was a reach at 21 but he's grown into his facial structure over the last year.

 

[silverfish]

Somebody smarter than me needs to do this for hockey



Miss - Save - Goal, and bucket it by shot type.

 

[silverfish]

Why Quality of Competition doesn’t matter to analytics experts anymore

Title is more heavy-handed than the article itself, but a lot of this is why I don't see a major reversal of Pionk's underlying numbers coming this year. Sure, he's going to stop playing against 'top competition', but he's also going to lose playing with KZB, which is far more important.

 

[Mac n Gs]

Micah sums it up pretty well:

“It matters, on average, about three or four times less than quality of teammates does,” said Micah Blake McCurdy, “even after you take into account how you have five opponents at 5v5 but only four teammates.”

“You shouldn't ever be allowed to talk about a guy's competition without making a quantitative comparison to his teammates,” said McCurdy. “I don't think you can say anything fairly unless you observe that rule of thumb.”

 

[silverfish]

Micah sums it up pretty well:

People really ignore this when talking about Pionk. Thanks for posting that quote, also stood out to me when I read the article. You try and make that point about Pionk and all you hear is "yEaH bUt He PlAyEd AgAiNsT cRoSbY".

Again, does it matter? Sure. But what matters more, playing 300 minutes with Kreider vs. 300 minutes with Spooner or 6 minutes against Crosby vs. 6 minutes against Granlund? Answer should be pretty obvious.

 

[Mac n Gs]

Why Quality of Competition doesn’t matter to analytics experts anymore

Title is more heavy-handed than the article itself, but a lot of this is why I don't see a major reversal of Pionk's underlying numbers coming this year. Sure, he's going to stop playing against 'top competition', but he's also going to lose playing with KZB, which is far more important.

I still think he’ll be tough to predict until we see how the team shakes out. I really can’t see him getting used the same way, and I find it hard to believe that we’ll be as bad defensively as we were last season. Just by those two factors, I see him shifting more towards even. It won’t he a huge reversal, but I think we’ll see he’s not detrimental to the team.

He only spent 23 min at 5v5 with KZB and they were good, but not great, together. They surprisingly didn’t have boosted OZ starts and saw more DZ starts.

He actually spent more TOI with just Mika and Kreids and did well from a quality for/against standpoint. With all of them together, their HDCF was poor while their HDCA was good.

He somehow managed to nuke how dangerous KZB was when they were astronomically better than any other line we iced last season. If only they didn’t have a negative shooting% relative to the team.

 

[silverfish]

I still think he’ll be tough to predict until we see how the team shakes out. I really can’t see him getting used the same way, and I find it hard to believe that we’ll be as bad defensively as we were last season. Just by those two factors, I see him shifting more towards even. It won’t he a huge reversal, but I think we’ll see he’s not detrimental to the team.

He only spent 23 min at 5v5 with KZB and they were good, but not great, together. They surprisingly didn’t have boosted OZ starts and saw more DZ starts.

He actually spent more TOI with just Mika and Kreids and did well from a quality for/against standpoint. With all of them together, their HDCF was poor while their HDCA was good.

He somehow managed to nuke how dangerous KZB was when they were astronomically better than any other line we iced last season. If only they didn’t have a negative shooting% relative to the team.

I think we were KZF at that point in the season and not KZB. But that was Pionk's "issue". He wasn't necessarily atrocious in terms of relCA60, but he was almost league-worst in terms of relCF60. With him on the ice, the Rangers could not generate offense. And if he's doing that to KZ, the team's two best offensive generators, that's a major issue.

 

[Mac n Gs]

I think we were KZF at that point in the season and not KZB. But that was Pionk's "issue". He wasn't necessarily atrocious in terms of relCA60, but he was almost league-worst in terms of relCF60. With him on the ice, the Rangers could not generate offense. And if he's doing that to KZ, the team's two best offensive generators, that's a major issue.

Ah, right. I always forget that they were a line later in the season. Re: Pionk, he was still positive in CF/60 rel with KZF on the ice, although they were much, much better without him.

Do you know who his linemates were where he truly had bad shot generation rates? I’ll try to look later tonight.

I’m shifting the goal posts a bit now, but I do think it’ll be more important in the long run if we can see Pionk consistently be a shot suppressor. That’s where I try to view his stats in a positive light. While he was heavily used in the DZ against “tougher comp,” he still didn’t hurt the team defensively. I just don’t know who he’d play with to show that effect. Smith or Claesson?

 

[silverfish]

Ah, right. I always forget that they were a line later in the season. Re: Pionk, he was still positive in CF/60 rel with KZF on the ice, although they were much, much better without him.

Do you know who his linemates were where he truly had bad shot generation rates? I’ll try to look later tonight.

I’m shifting the goal posts a bit now, but I do think it’ll be more important in the long run if we can see Pionk consistently be a shot suppressor. That’s where I try to view his stats in a positive light. While he was heavily used in the DZ against “tougher comp,” he still didn’t hurt the team defensively. I just don’t know who he’d play with to show that effect. Smith or Claesson?

I always default to the hockeyviz WOWY viz, but I think Nat Stat Trick keeps combo/pairing stats...

NYR 1718 Pionk With-or-Without You

Smith-Pionk, at first glance, doesn't really make me angry. I'm really excited for the 180 HFNYR does on Smith this year, who wasn't even that bad last year. HFNYR taeks about Smith make me dizzy. Can't keep up.

 

[Mac n Gs]

I always default to the hockeyviz WOWY viz, but I think Nat Stat Trick keeps combo/pairing stats...

NYR 1718 Pionk With-or-Without You

Smith-Pionk, at first glance, doesn't really make me angry. I'm really excited for the 180 HFNYR does on Smith this year, who wasn't even that bad last year. HFNYR taeks about Smith make me dizzy. Can't keep up.

I didn’t watch Harftord last year outside of like three games, so I don’t remember who he played with and if it was with Smith. I’m sure there’s some lineups on twitter that will have that answer, so I’ll check later.

 

[DanielBrassard]

People really ignore this when talking about Pionk. Thanks for posting that quote, also stood out to me when I read the article. You try and make that point about Pionk and all you hear is "yEaH bUt He PlAyEd AgAiNsT cRoSbY".

Again, does it matter? Sure. But what matters more, playing 300 minutes with Kreider vs. 300 minutes with Spooner or 6 minutes against Crosby vs. 6 minutes against Granlund? Answer should be pretty obvious.

I agree with you here SF, in that I don't think Pionk's struggles were because of his usage as much as the rest of HF wants to make it out to be. But I really like Pionk's skill set, and I wouldn't be surprised if he was actually good next season. I think a lot of his problems last season was his decision making with the puck in the defensive zone, seemed like he was always looking to make the safe play, or held onto the puck too long. I hope with more time in the league, he makes quicker and better decisions, which could change the entire outlook on him. But who knows.