The other thing here is the idea that PP points don't matter.

They matter, they just aren't useful for evaluations and predicting the future. If you get PP time, you'll get PP points - it's a given. (barring bad luck, you're going to score on 15-20% of your PPs)

If you want to be able to predict how a player will do in the future, it's essential to ignore PP points because they could easily mislead you.

Say you want to re-sign a player, and he got 50 points, 20 on the PP. Say your team is healthy and you don't plan on using him on the PP anymore. You better think of him as a 30 point player in this case.
- James_Tanner

I don't disagree that it's not situationally useful. But a lot of people just use it broadly that they don't matter. And maybe one guy gets a lot more than another because he's just better and he's more suited to it than someone else. My main problem is we don't have enough data.

Certainly not enough good data. And by "we" I mean the general public. We know the NHL and the teams have more in depth numbers.

As far as shot attempts go, is a player, let's say a defenseman, who just wings it from the point as soon as he gets it, but gets a very high percentage blocked better than a guy who might take a second, takes less overall shots, but gets more on net? Your boy Franson fits into the latter category, he doesn't take a lot of shots, and he takes his time, but he gets them through a lot of the time.

Everyone excludes data. What are you even talking about? I believe the statistical term for it is "noise" as in it's information that is irrelevant.

The reason you exclude PP points and second assists in player evaluations is because they are like false positives. The whole point is to find out what's repeatable and whether or not you should get excited about a production surge.

He does have four 5v5 points in five games - that is good. I am only tempering expectations here, because if people think they got a point per game player, they are wrong.

Even with this small sample size, you can see how what I'm doing is effective because two 5v5 primary points in 5 games is a lot better expectation for this player than that he'll be a point per gamer.

If you think I'm just working backwards to fit my narrative, that's ridiculous. I mean, what is my narrative? "bad player is still bad"?

All I have done is give put the players good looking streak into a perspective that is very very likely to be more indicative of what we can expect going forward.
- James_Tanner

That's why I had the qualifier of relevant.

If you're a team evaluating a player hopefully you have someone, many people, watching every game and see every goal and see who actually contributed. I agree that sometimes guys get secondary assists just because they happened to touch the puck, but it's also true that guys who don't get assists nevertheless contributed to the goal, but does that make them less valuable?

And I wasn't speaking specifically about Burmistrov. You say Franson is better than Ristolainen -which is patently ridiculous- because Ristolainen gets a lot of PP points and ice time in general. Has it ever struck you that he gets all that ice time, PP time and those PP points because he's just better than Franson? Is Claude Giroux bad? Kessel? Malkin? Wennberg and Atkinson? They're the top 5 in PP points. You're always preaching subtlety yet you often throw it out the window. We don't need analytics to tell us that good players are good, and even the best have flaws in their game, we know who's good already, we need it to tell us who the tweeners are and if they're any good or not.

I don't disagree that it's not situationally useful. But a lot of people just use it broadly that they don't matter. And maybe one guy gets a lot more than another because he's just better and he's more suited to it than someone else. My main problem is we don't have enough data.

Certainly not enough good data. And by "we" I mean the general public. We know the NHL and the teams have more in depth numbers.

As far as shot attempts go, is a player, let's say a defenseman, who just wings it from the point as soon as he gets it, but gets a very high percentage blocked better than a guy who might take a second, takes less overall shots, but gets more on net? Your boy Franson fits into the latter category, he doesn't take a lot of shots, and he takes his time, but he gets them through a lot of the time.
- Wetbandit1

I have a hard time buying the fact that teams have this great set of secret data the public doesn't have. Frankly, there are so many shot attempts in each game, and the sample size so massive, that I don't think you'd be able to do much more with better or different data. The game is about scoring, you score with shots. If there's a margin of error due to human calculations, it's so small in a giant sample size that it doesn't matter.

I am sure in the minute, in the individual, and in certain ways (perhaps you could get more useful data out of small sample sizes) this data could have some value. However, people act like any use of public statistics is fraudulent because teams have this special data - and I call bullpoop on that.

That's why I had the qualifier of relevant.

If you're a team evaluating a player hopefully you have someone, many people, watching every game and see every goal and see who actually contributed. I agree that sometimes guys get secondary assists just because they happened to touch the puck, but it's also true that guys who don't get assists nevertheless contributed to the goal, but does that make them less valuable?


Not really, but goals are random anyways, so you should be making your evaluations on shot-attempts, not goals/assists or points. In this case, we're talking about a very specific example. Honestly, I only used it because it made a stronger cases - for me, the fact that his Corsi is so, so bad, and that every teammate gets worse with him than without him is enough for me to know the 5 in 5 is a fluke.

And I wasn't speaking specifically about Burmistrov. You say Franson is better than Ristolainen -which is patently ridiculous- because Ristolainen gets a lot of PP points and ice time in general. Has it ever struck you that he gets all that ice time, PP time and those PP points because he's just better than Franson? Is Claude Giroux bad? Kessel? Malkin? Wennberg and Atkinson? They're the top 5 in PP points. You're always preaching subtlety yet you often throw it out the window. We don't need analytics to tell us that good players are good, and even the best have flaws in their game, we know who's good already, we need it to tell us who the tweeners are and if they're any good or not.
- Wetbandit1

Ristolainen is amazing on the PP. But at even-strength he is terrible. He allows more shots, goals, scoring chances than he produces and 100% I guarantee that Franson is better than him at 5v5. If Buffalo was interested in winning today, they might deploy them effectively. Because they aren't, and because Risto is - they hope - their future #1 D - they play him the most minutes.

Fact is though, at 5v5, or 80% the game, he isn't even an average defenseman.

I have a hard time buying the fact that teams have this great set of secret data the public doesn't have. Frankly, there are so many shot attempts in each game, and the sample size so massive, that I don't think you'd be able to do much more with better or different data. The game is about scoring, you score with shots. If there's a margin of error due to human calculations, it's so small in a giant sample size that it doesn't matter.

I am sure in the minute, in the individual, and in certain ways (perhaps you could get more useful data out of small sample sizes) this data could have some value. However, people act like any use of public statistics is fraudulent because teams have this special data - and I call bullpoop on that.
- James_Tanner

With parity as strong as it is, minutiae is how teams win. It's not that they have "secret" data. It's more that they have more in depth data. If that makes any sense.
I linked an article recently here about it. It's more about the minutiae of stats. Like maybe where a blocked shot was taken from, instead of where the block occurred. A lot of blocked shots happen in the middle of the ice right? Because players are shooting at the net. Makes sense. But if you trace it back a lot of shots are of such low danger as to be inconsequential. To me it's more of the quality of the chance you take away. Same goes with shooting. I know it's not a good example, but it's the only one I have, and I'm sure other teams do it too. The Sabres preach shots from the slot. They're trying to pass up less dangerous chances to try and get a high danger one. And it's working, they just can't put the puck in the net(their SH% is .9% less than the league average, that's bad luck). Last month they were 3rd in 5v5 scoring chances for. Chances and shot volume are not necessarily inextricably linked.

And there are other stats that I haven't come across either. Like who allows the most odd-man rushes. Who scores most on odd-man rushes. Things of that nature which I believe you need to have a full picture. I get what you're saying, shots are the most common offensive stat. So you have the biggest sample size therefore you should be able to get the most repeatable results. I just think hockey is was more subtle than just shot attempts.

I'm not anti-stats by any means. They are very useful at times. I just don't think we, at this time, have the full picture. I don't think anyone does. To claim you know everything is nothing short of absurd, since it's impossible to know everything. And if for some crazy reason you had all the individual data points you still have to be able to link them together correctly. And using incomplete data to make absolute statements seems really silly to me.

Ristolainen is amazing on the PP. But at even-strength he is terrible. He allows more shots, goals, scoring chances than he produces and 100% I guarantee that Franson is better than him at 5v5. If Buffalo was interested in winning today, they might deploy them effectively. Because they aren't, and because Risto is - they hope - their future #1 D - they play him the most minutes.

Fact is though, at 5v5, or 80% the game, he isn't even an average defenseman.
- James_Tanner

He definitely has areas he can improve. No doubt. However a lot of people talk about him like he's 27 and not 22(and he's just barely that, his birthday is in late October). He plays an absolute ton of minutes, is a monster on the PP, and he is making strides in his own zone. Last year he got beat so very many times on the rush. Almost always because he was out of position. This year so far he's cut that almost completely out of his game his positioning is much better. I'm not saying he never gets beat, everyone gets beat. He's just not the sieve on the rush he was last year.

But again he's only 22 and he's already this good offensively. I'm not comparing the 2 but remember Karlsson's problems early in his career? It takes those guys a long time to get it right. He hasn't even played 3 full seasons yet, just 241 games.

I missed the bolded about Burmistrov. Some players you can argue just don't play well together, their styles don't match, one slept with the other's girlfriend, whatever, and both their numbers take a hit when playing together, and then when apart they rise. Another Sabres example is Ristolainen and Gorges. They are both better apart than they are together. Burmistrov seems to just be bad.

Chychrun and a top-3 protected pick for Duchene and a second?
- jcragcrumple

No way! The value is way off, plus Duchene goes to free agency right when the the coyotes are projected to start contending (and when chychrun will be coming in to his prime).

I actually love duchene and can't believe Colorado would get rid of him, but I wouldn't do it for chycrun and his contract just doesn't make sense for the coyotes.

With that said...if they can swing a trade and sign for him next year, I'd be all for it!
A top offensive prospect and high pick, but only if he's locked up.

Really enjoy Tanner's secondary assist notion. Pretty comical how he's just going to now discredit Burmi's secondary assists because it validates his point further.

I don't have a pony in this particular race so idc really. Just funny because I'm sure OEL has gathered some secondary assists in his career, but we just don't mention that.

I get his point, the advanced stats and overall history point to Burmi not being that good, but to discredit a nice start on a new team is just a pretty crappy thing to do.

Ristolainen is amazing on the PP. But at even-strength he is terrible. He allows more shots, goals, scoring chances than he produces and 100% I guarantee that Franson is better than him at 5v5. If Buffalo was interested in winning today, they might deploy them effectively. Because they aren't, and because Risto is - they hope - their future #1 D - they play him the most minutes.

Fact is though, at 5v5, or 80% the game, he isn't even an average defenseman.
- James_Tanner

Probably due to him being over used, play Risto a lot less his #'s would be better would be my guess. He plays way too many mins.

Probably due to him being over used, play Risto a lot less his #'s would be better would be my guess. He plays way too many mins.
- Garnie

That was the plan coming into the season, but the injuries and complete lack of talent on the blue line have made that impossible. He played 35+ minutes the other night. It was like 10th most since they started keeping track in '97 for a regular season game. I remember Bylsma saying he wanted to reduce his minutes by 3 or 4. And he's playing even more than he was last year. 1:45 a game to be exact, to 27:01, just 17 seconds shy of 1st, Byfuglien.

but Arizona's corsi is 46% for the season (5 on 5), so he is just playing average so far on the team. That isn't bad, especially considering he played some tough minutes.

That was the plan coming into the season, but the injuries and complete lack of talent on the blue line have made that impossible. He played 35+ minutes the other night. It was like 10th most since they started keeping track in '97 for a regular season game. I remember Bylsma saying he wanted to reduce his minutes by 3 or 4. And he's playing even more than he was last year. 1:45 a game to be exact, to 27:01, just 17 seconds shy of 1st, Byfuglien.
- Wetbandit1

20-22 mins max...played 29 against Nashville on Monday and -23 in shot attempts.

I know his QOC was tops but he's just too young and not ready for these type of mins.

With parity as strong as it is, minutiae is how teams win. It's not that they have "secret" data. It's more that they have more in depth data. If that makes any sense.
I linked an article recently here about it. It's more about the minutiae of stats. Like maybe where a blocked shot was taken from, instead of where the block occurred. A lot of blocked shots happen in the middle of the ice right? Because players are shooting at the net. Makes sense. But if you trace it back a lot of shots are of such low danger as to be inconsequential. To me it's more of the quality of the chance you take away. Same goes with shooting. I know it's not a good example, but it's the only one I have, and I'm sure other teams do it too. The Sabres preach shots from the slot. They're trying to pass up less dangerous chances to try and get a high danger one. And it's working, they just can't put the puck in the net(their SH% is .9% less than the league average, that's bad luck). Last month they were 3rd in 5v5 scoring chances for. Chances and shot volume are not necessarily inextricably linked.

In a game-to-game data set, I agree. Over the long term, they almost certainly are, because of parity and how good the players are.


And there are other stats that I haven't come across either. Like who allows the most odd-man rushes. Who scores most on odd-man rushes. Things of that nature which I believe you need to have a full picture. I get what you're saying, shots are the most common offensive stat. So you have the biggest sample size therefore you should be able to get the most repeatable results. I just think hockey is was more subtle than just shot attempts.

I definitely agree. And I agree that in a league of parity, even a slight edge can make a huge difference. However, the danger is that the team that scores the most breakaways or 2-1's over 20 games probably changes all the time. I doubt there's any consistency.


I'm not anti-stats by any means. They are very useful at times. I just don't think we, at this time, have the full picture. I don't think anyone does. To claim you know everything is nothing short of absurd, since it's impossible to know everything. And if for some crazy reason you had all the individual data points you still have to be able to link them together correctly. And using incomplete data to make absolute statements seems really silly to me.
- Wetbandit1

There are some absolutes you can make though. If a player is getting crushed in possession, goals, etc. then it's unlikely he secretly does this one thing that makes a difference that doesn't show up there. Not impossible. And I agree that the more broad, certain and extreme your position, the least likely it is to be true.

I don't think stats are any more than a guide. However, they certainly should hold more weight than a guys collective reputation.

Also, this has been a great conversation about stats - one of teh better ones i've had on this site.

Really enjoy Tanner's secondary assist notion. Pretty comical how he's just going to now discredit Burmi's secondary assists because it validates his point further.

I don't have a pony in this particular race so idc really. Just funny because I'm sure OEL has gathered some secondary assists in his career, but we just don't mention that.

I get his point, the advanced stats and overall history point to Burmi not being that good, but to discredit a nice start on a new team is just a pretty crappy thing to do.
- SteveJ123

Not trying to discredit him.

But come on, what makes more sense:

Burnmistrov as a ppg player, or a 2/5 guy?

I would definitely look into OEL's secondary assists if there was a reason to, but no one is overrating him. If anything, most NHL fans don't even consider him the top five dman that he is.

Probably due to him being over used, play Risto a lot less his #'s would be better would be my guess. He plays way too many mins.
- Garnie

Like with Morgan Rielly, I think teams try to force a role on a certain player. If they could accept what they have rather than what they want, and find a well to shelter his minutes so he can be the most effective, they'd have a better team.

Look at Columbus putting Sam Gagner on the 4th line and using him on the PP. Which is something I've been advocating for years.

Like with Morgan Rielly, I think teams try to force a role on a certain player. If they could accept what they have rather than what they want, and find a well to shelter his minutes so he can be the most effective, they'd have a better team.

Look at Columbus putting Sam Gagner on the 4th line and using him on the PP. Which is something I've been advocating for years.
- James_Tanner

Yep...I'm not sure it's a team accepting what they have but expecting it too early in their careers maybe. I expect both Rielly and Risto to be able to handle bigger mins but maybe later on in 3/4/5 years.

I was the one sad fan, when Burmistrov was claimed by the Coyotes.

He was one of my favorite Jets and I don't think he got a fair shake in Winnipeg. Hope he continues to average a point per game while in Arizona.

Burmistrov scored a goal tonight VS Vancouver. Now 6 points in 6 games. I don't care what it means on the stat sheet, but that's pretty good. And he really looked good. Confident and skilled. Who knows if it will last. So far he looks likes a steal.

Burmistrov scored a goal tonight VS Vancouver. Now 6 points in 6 games. I don't care what it means on the stat sheet, but that's pretty good. And he really looked good. Confident and skilled. Who knows if it will last. So far he looks likes a steal.
- Gomey

It won't and he's not. That being said, I hope I'm wrong.

Red Flag #2: His Corsi-For Rating ifor every player he's been on the ice with in a Coyotes uniform is 46% or less, and it always gets better for the other player when they get away from Burmistrov.
- Tanner

Hey James... explain to me how that stat is relevant please.

I just sorted players by this season's Corsi, TMCF%, (5v5 at least 400 minutes played)

THE TOP 28 PLAYERS are all on Boston or Los Angeles (except 1 Hab).

63 players are better than Alex Ovechkin

Sidney Crosby comes in at a nice round number... 200

Tarasenko 262
Matthews 274
Eichel 335
Taylor Hall 462

.. I still haven't found McDavid. I probably skipped over him.

I could go on and on and on with examples of how dumb the players rankings are when ranked by this stat.

Did I use the wrong stat? (I don't think so, but hey maybe?).

BTW... if I DID use the wrong stat... that doesn't change the fact that player rankings based on this Corsi stat, just like the last one I looked at, are ridiculous. This isn't an outlier problem. It is a problem with this stat being utterly meaningless.

Hey James... explain to me how that stat is relevant please.

I just sorted players by this season's Corsi, TMCF%, (5v5 at least 400 minutes played)

THE TOP 28 PLAYERS are all on Boston or Los Angeles (except 1 Hab).

63 players are better than Alex Ovechkin

Sidney Crosby comes in at a nice round number... 200

Tarasenko 262
Matthews 274
Eichel 335
Taylor Hall 462

.. I still haven't found McDavid. I probably skipped over him.

I could go on and on and on with examples of how dumb the players rankings are when ranked by this stat.

Did I use the wrong stat? (I don't think so, but hey maybe?).

BTW... if I DID use the wrong stat... that doesn't change the fact that player rankings based on this Corsi stat, just like the last one I looked at, are ridiculous. This isn't an outlier problem. It is a problem with this stat being utterly meaningless.
- Aetherial

Is this a post from 2008? Or are you asking this question for real?

Is this a post from 2008? Or are you asking this question for real?
- James_Tanner

Dude, that is player rankings by the stats YOU SPECIFICALLY referenced.

It is entirely possible I am using the wrong stat. If I am, please tell me which stat YOU are using, and I will look at the player rankings that way, and we can resume our discussion.

Regardless, the rankings on this Corsi stat are stupid. That is not even debatable.


Last time I tore apart a stat, you advised me that literally no one would use that single stat (It was CF% ROFL!).

This time, you specifically used a stat.

Let's see how well this stat stands up to scrutiny ok?

Dude, that is player rankings by the stats YOU SPECIFICALLY referenced.

It is entirely possible I am using the wrong stat. If I am, please tell me which stat YOU are using, and I will look at the player rankings that way, and we can resume our discussion.

Regardless, the rankings on this Corsi stat are stupid. That is not even debatable.


Last time I tore apart a stat, you advised me that literally no one would use that single stat (It was CF% ROFL!).

This time, you specifically used a stat.

Let's see how well this stat stands up to scrutiny ok?
- Aetherial

NO one is using one stat to rank a player. All I was saying was that if you look into his ppg rate so far, it's obviously not going to last. I also used five stats, if you want to be technical: CF, points, assists, secondary assists, and his corsi vs when he's on and off the ice with different teammates.

I have never heard anyone list the top corsi players and then suggest they are the best players in the NHL, in that order.

If you want to know if CF stands up to scrutiny, consider this: in the last four seasons, only 3 teams have winning records over those 320 odd games with their CF under 50. Those three teams were at 49. something and all had elite goalies.

NO one is using one stat to rank a player. All I was saying was that if you look into his ppg rate so far, it's obviously not going to last. I also used five stats, if you want to be technical: CF, points, assists, secondary assists, and his corsi vs when he's on and off the ice with different teammates.

I have never heard anyone list the top corsi players and then suggest they are the best players in the NHL, in that order.

If you want to know if CF stands up to scrutiny, consider this: in the last four seasons, only 3 teams have winning records over those 320 odd games with their CF under 50. Those three teams were at 49. something and all had elite goalies.
- james_tanner1

James,

I know you fully understand the argument I am making. Address my comments about the efficacy of this stat at a player level, or don't, but please don't veer off of the target of my statements. I was very specific and I didn't stutter.


Re: the bolded. No, you haven't because the list would be ridiculed. So, if a ranking list (as you admit) would never be used to make such a suggestion... then what does that say about the ranking? What can you actually conclude about the use of the stat?

The rest of your response has not addressed my point. In fact, your last paragraph, you are now talking about teams. The debate is clearly about individual rankings based on a stat you used to support your assessment of an individual.

Furthermore, it is irrelevant whether or not you chose 5 things. I did not challenge your overall assessment of him or any of the other 4 things. In fact, I specifically quoted one of your "red flags".

My point is simple. If you judge a player, and use a statistic to support your judgement, you have to be certain that statistic is meaningful at a player level. In this case, we discover that 27 of the top 28 are from 2 teams, and there are literally hundreds of rankings that belie what we know about these players talent level, effectiveness and value to their organizations.

Clearly there is something very wrong with using this stat to assess individuals.

James,

I know you fully understand the argument I am making. Address my comments about the efficacy of this stat at a player level, or don't, but please don't veer off of the target of my statements. I was very specific and I didn't stutter.


Re: the bolded. No, you haven't because the list would be ridiculed. So, if a ranking list (as you admit) would never be used to make such a suggestion... then what does that say about the ranking? What can you actually conclude about the use of the stat?


- Aetherial

Uh, try doing that with any stat - goals, points whatever, and it won't work. Does this mean they aren't good stats. I can't believe you don't see how bad of an argument you are making here.