TSN: Jake Gardiner or: How I Learned to Stop Worrying and Love the Advanced Stats

[91Kadri91*]

I believe this is a myth and I challenge it every time it comes up. Can anyone provide any statistical evidence that the impact of QoC is negligible?

Quality of competition matters in short-sample sizes, but it regulates itself over the course of a season; by the time the season is over, the differences in competition level from player-to-player is almost always negligible.

But here’s the key thing: While it matters if a player is facing Sidney Crosby instead of John Scott at any given moment, the range of competition that a player faces over the course of a season is EXTREMELY SMALL. The gap between the players facing the hardest competition and those facing the weakest competition is the same as facing an average player at most like 4 shot attempts per 60. In other words, the guy with the toughest competition in the league will face an average opponent who is +2 corsi/60, while the guy facing the weakest will face an average opponent who is -2 corsi/60. And nearly all players won’t be in these extremes – most will be within -1 corsi/60 and +1 corsi/60. And as you might expect the gap between opponents who are +1 shot attempts per 60 and those -1 is practically nothing.

Yet you’ll hear people talk about how one player plays “really weak” competition or another player’s bad #s are because he takes “the toughs” – this doesn’t really mean anything.

As you can see, basically nothing changes (not surprising for Gunnarson, who mostly plays WITH Phaneuf) for any of the D men despite facing the tougher competition. The only exception is Paul Ranger, which appears to be a sample size quirk that is unlikely to continue (he has very little time vs tough competition in which he has done hilariously poorly which is almost certainly a sample size quirk, but this method acts as if it’s real and uses that # for a large portion of Ranger’s theoretical ice time.) Jake Gardiner goes from the Leafs’ best D Man to….the Leafs’ best D man. It barely makes a dent.

Again, the zone starts for all but Morgan Rielly are basicaly identical (and Rielly’s gap isn’t very large either) and these are with the same teammates. Quite simply put anyone who points out Dion faces top competition as an argument in his favor needs to understand that Phaneuf has been below average against all levels of competition and the top competition he’s faced really doesn’t do much to his numbers.

http://hockey-graphs.com/2014/01/06...-players-stats-toronto-maple-leafs-d-edition/

However, although the differences between players in quality of competition are real, they are not very large. The plot above goes from +10 to -10, but actual year-end totals go from roughly +1.5 to -1.5, with most players falling in a much narrower range than that. Everyone faces opponents with both good and bad shot differential, and the differences in time spent against various strength opponents by these metrics are minimal. The Flyers were relatively focused on matching lines in 2010-2011 – Andreas Nodl ranked 15th in competition among NHL forwards with at least 500 even strength minutes played, while Blair Betts ranked 290th (out of 337). And yet a histogram showing how much of their ice time was spent against opponents of various strengths shows scarcely any difference between them:

http://nhlnumbers.com/2012/7/23/the-importance-of-quality-of-competition

 

[91Kadri91*]

As for QoC, it's been calculated to be not an important factor. Zone starts, minutes played and QoT are the three main factors.

You should read these articles:

http://puckplusplus.com/2015/01/15/how-much-do-zone-starts-matter-i-maybe-not-as-much-as-we-thought/

http://puckplusplus.com/2015/01/20/how-much-do-zone-starts-matter-part-ii-a-lot-on-their-own-not-that-much-in-aggregate/

If anyone doesn't want to, or doesn't have the time, this graph should provide you the gist of the articles:

 

[Gary Nylund]

91Kadri91 - if you don't mind my asking, what do you do for living? I've always been good with numbers but some of your posts make my head hurt.

 

[p.l.f.]

Gary Nylund lol
Speaking of defenseman then never panned out

 

[Beleafer4]

Just a glimpse at what Gardiner is on the ice:
http://public.tableau.com/shared/2GSHKJKJW?:display_count=yes

Of course it doesn't tell everything, but it summarizes it real well.
A lot of people are fooled with the eye test on players like Gardiner, Stralman, Letang and even Subban because of the way they play.

Ironically I am a huge fanboy of all those players when it comes to the eye test. But that may be because I am more left-winged

My personal hypothesis is that conservative people don't like defenseman like gardiner. I have done a lot of research for a psyc course I did before and it is proven that conservative people are more easily startled and are more prone to produce a hostile respons e to that anxiety. Those who are left winged aren't as easily startled or as hostile in anxiety producing situations.

Which makes sense why they wouldn't like a guy like gardiner despite the stats. He plays a very anxiety inducing game. He holds on to the puck a lot even in dangerous situations, makes risky passes and just seems to hate the safe play. For me, what he does works and he makes way more good plays then bad so I really like him. For others, he plays on the edge - it makes them worried and they don't like that. They much prefer the "chip it off the boards + knock down all the forwards on their butts where they can't do anything" style of play. I remember watching a game live last year when gardiner was fending off 2 attackers and was holding on the puck for a bit too long. In the end, nothing bad happened and he ended up making a great play and getting the puck out, but there were still a lot of people muttering their discontent - probably because that sequence scared them..

And even on this board. It seems that the gardiner dispatchers preach a more conservative approach whereas those who support him (like nithoniel and 91kadri91) are imo more left-winged.

Just something I noticed which could explain why a guy like gardiner is so polarizing amongst fans.

 

[PJJ]

Damn you are right. Right now Ive had an offseason without watching Gardiner pull off some ******** in the dzone so im pretty excited to see him this season, but ill probably be back to yelling at the tv by week 2

 

[sommervr]

Gardiner is a Rorschach inkblot test for idiots. The new management aren't idiots.

 

[Macallan18]

Ironically I am a huge fanboy of all those players when it comes to the eye test. But that may be because I am more left-winged

My personal hypothesis is that conservative people don't like defenseman like gardiner. I have done a lot of research for a psyc course I did before and it is proven that conservative people are more easily startled and are more prone to produce a hostile respons e to that anxiety. Those who are left winged aren't as easily startled or as hostile in anxiety producing situations.

Which makes sense why they wouldn't like a guy like gardiner despite the stats. He plays a very anxiety inducing game. He holds on to the puck a lot even in dangerous situations, makes risky passes and just seems to hate the safe play. For me, what he does works and he makes way more good plays then bad so I really like him. For others, he plays on the edge - it makes them worried and they don't like that. They much prefer the "chip it off the boards + knock down all the forwards on their butts where they can't do anything" style of play. I remember watching a game live last year when gardiner was fending off 2 attackers and was holding on the puck for a bit too long. In the end, nothing bad happened and he ended up making a great play and getting the puck out, but there were still a lot of people muttering their discontent - probably because that sequence scared them..

And even on this board. It seems that the gardiner dispatchers preach a more conservative approach whereas those who support him (like nithoniel and 91kadri91) are imo more left-winged.

Just something I noticed which could explain why a guy like gardiner is so polarizing amongst fans.

lol, no. The divide isn't left and right. It's the crowd that relies on its gut and counts the number of mistakes Gardiner makes, and the crowd that sees his positive values amply illustrated by the results.
Next you'll be arguing that liberals are smarter then conservatives!!

 

[Nithoniniel]

I thought we were talking a 5 on 5 ranking? Were we not?

Bolded-That is what I have been saying and it has been you saying there is no subjectivity because all variables have been asked and accounted for. Yes, you will say stats don't show everything, but then follow up with a stat to show or disprove something. You don't see how your statement is confusing when now it is you that is saying stats don't show everything and you can't evaluate play on them alone? It is now you saying there is no clear guidelines, when earlier, they couldn't be clearer and I just don't understand them.

I do find It funny though that they seem to have a formula to compare players in all areas for everything except one that compiles all stats per player and then ranks them against each other. You would think it would be very useful. Maybe the results would show players like Tanev, Gardiner etc more effective than Weber, Keith, etc and then advanced stats would be in question to how well they actually evaluate players.

I didn't take it as a 5-on-5 ranking. Sorry, must have misunderstood you.

And well, you can't just decide which parts you want to be subjective. That's what I'm trying to say here. It's all about knowing where there is room for subjectivity and where there is not. Talking about how valuable it is to have a disciplined guy who can seriously limit shots against versus someone who has better overall numbers? I can't tell you that, it's situational and subjective. Talking about how you don't believe a player is good because he gets easy minutes? That's where I will correct you.

To be honest, you seem to infer a lot of meaning in my posts. I usually never have problems with people understanding my point.

 

[RLF]

Quality of competition matters in short-sample sizes, but it regulates itself over the course of a season; by the time the season is over, the differences in competition level from player-to-player is almost always negligible.

http://hockey-graphs.com/2014/01/06...-players-stats-toronto-maple-leafs-d-edition/

http://nhlnumbers.com/2012/7/23/the-importance-of-quality-of-competition

First off, I do appreciate all the stats you show on the this board, I may not put as much faith in the numbers as you do, but do appreciate the info.

Here is a couple of examples of where I find analysis contradicts itself or context of how the game is played is not taken into consideration and why I can't put as much faith in all this as others.

Contradiction
from one of the articles you suggested a poster should read on zone starts.

http://puckplusplus.com/2015/01/15/how-much-do-zone-starts-matter-i-maybe-not-as-much-as-we-thought/

From the article

"The main issue with the current approach to measuring zone starts is that the measurement is often skewed (and sometimes heavily) by the performance and talent of the player in question. Bad players tend to end up with more defensive zone faceoffs because their opponents tend to get more shot attempts against them, which leads to more opportunities for their goalie to freeze the puck and more defensive zone faceoffs. The same idea is true in reverse for good players, and it all adds up to a false correlation between the traditional zone start measure and possession numbers."

According to this, QOC matters greatly in shot attempts, possession and zone time, but we are told to disregard this over a season because it all evens out and has little affect on overall numbers. Why do they consider QOC here?...because their theory falls apart without it. Again, subjectivity is used in order to determine what is important and what isn't in order to reach a desired result.

Context

This is an article linked on the article above page.

http://www.todaysslapshot.com/from-the-ice/one-timers/nhl-defensemen-who-should-be-trusted-less/

In this article the writer gives what he considers definitive facts on which NHL d-men should be used less or are ineffective when holding a lead. The two stats used to determine this...TOI while holding a lead and Corsi.

When holding a lead late in a period or especially late in the game...shot attempts for, are not the main focus when defending the lead and controlling the puck is not as much a concern. It's get the puck out, dump it in, one man forecheck, no dangerous plays, be safe with the puck. Of course possession stats and shot attempts will diminish for the team defending the lead.

Also here is where QOC has to be considered. A guy like Weber, is going to be on the ice every time the other coach puts out his best, which will happen often when trying to tie a game, so of course he is out there more to "defend". He is not put out to score or create offence, he is put out to defend his end, win the puck battles and dump it out...protect the lead and kill the clock.

The second defending pairing will see more of the second tier offensive players from the trailing team. An easier time of it because QOC is less and the defending team will likely have better possession time and may even create some offence because their QOC is lower and therefore, does not create as many shots against...as the other article mentions. So sure, some guys stats will look better. In all honesty though, who would you rather have on the ice protecting the lead...Suter and Weber or Schenn and Krug? Stats say Suter and Weber should not be counted on as much and Schenn and Krug should get more time. Also, better teams will have the lead more often late in games and therefore the stats get skewed.

This is just a couple of examples of why I will respond with some of the things I do when confronted with stats to prove a point...statistical analysis is determined by an individuals desire to prove a point. Even the formulas are created this way and many will even say, "when we project this or that we can see how"...No, projections are not facts, they very often do not turn out as assumed when put into practice. This is why there are so many failures prior to success, because projections were wrong. So to completely discount the eye test for evaluation as some suggest, makes little sense to me. They should be used together, not one being completely more definitive than the other.

The only true way to know how a player would react in a different role is to put him in that role and actually play the games and see what happens. He may do as expected, he may fall flat on his face. Stats can often contradict each other and it is we who decide which is the right one, which is very subjective, no different than the eye test that way.

 

[RLF]

I didn't take it as a 5-on-5 ranking. Sorry, must have misunderstood you.

And well, you can't just decide which parts you want to be subjective. That's what I'm trying to say here. It's all about knowing where there is room for subjectivity and where there is not. Talking about how valuable it is to have a disciplined guy who can seriously limit shots against versus someone who has better overall numbers? I can't tell you that, it's situational and subjective. Talking about how you don't believe a player is good because he gets easy minutes? That's where I will correct you.

To be honest, you seem to infer a lot of meaning in my posts. I usually never have problems with people understanding my point.

Nor do I.

I never said Gardiner isn't any good, more the opposite really, I have said he could be a #3 d-man, but I rate him a 4/5 at this moment. I have said I believe QOC matters and that it should be considered. This is where you "correct me" because you say it is not subjective and stats show this. Take a look back though, I am the one saying that things are subjective and situational and you have been telling me, not really, the stats show otherwise. Now you are saying it is too hard to say who is more valuable because it is situational and subjective. So yes, that is confusing, but let's just say you're right and I am unable to figure out which ones are subjective or situational and which aren't.

 

[Nithoniniel]

Nor do I.

I never said Gardiner isn't any good, more the opposite really, I have said he could be a #3 d-man, but I rate him a 4/5 at this moment. I have said I believe QOC matters and that it should be considered. This is where you "correct me" because you say it is not subjective and stats show this. Take a look back though, I am the one saying that things are subjective and situational and you have been telling me, not really, the stats show otherwise. Now you are saying it is too hard to say who is more valuable because it is situational and subjective. So yes, that is confusing, but let's just say you're right and I am unable to figure out which ones are subjective or situational and which aren't.

You are still struggling with the very thing I'm about to say for the third time now. There is a time and place for subjectivity, and a time and place where there is not.

Yes you have said that things are subjective and situational in some events, where they haven't been. That doesn't mean that there are not at all parts of evaluation that is subjective and situational.

You can't tell me it's subjective whether or not the sun is yellow. But whether or not it is pretty is. You have been saying that while the sun sure is yellow, you disagree on the extent due to contextual reasons such as the amount of rainbows currently visible. Now I tell you that no, the sun is always yellow. Rainbows don't matter. At which point you jump down my throat since I say I find the sun pretty and therefor negating some previous point that there are no subjectivity at all, that I never made.

And that's where I'm overly done with this particular conversation. Sorry for the excessive grumpiness. I'm excessively grumpy.

 

[Nithoniniel]

Also, the point of a scientific approach is not producing a theory, criticize it, choose the side you like and then fight it out.

You study the material, use it to create a theory, try to poke holes in it, test those attempts, and then draw conclusions.

People seem to think that attempts to poke holes in a theory means that the conclusion is wrong.

Edit: Very simplified here. Just trying to make a point.

 

[RLF]

Also, the point of a scientific approach is not producing a theory, criticize it, choose the side you like and then fight it out.

You study the material, use it to create a theory, try to poke holes in it, test those attempts, and then draw conclusions.

People seem to think that attempts to poke holes in a theory means that the conclusion is wrong.

Edit: Very simplified here. Just trying to make a point.

Fair enough, which is my point on stats and how they are subjective at any given moment. How many times have we been told a scientific theory is fact, then some time later, another scientific theory comes along to counter the previous and prove it false or at minimum, question its authenticity? It happens all the time.

Even in hockey statistical analysis it is happening constantly because advanced stats are still very new. It was not long ago Corsi was the be all end all, now, not so much. Why? Because someone else tested and decided enough variables weren't considered. This is my point on subjectivity. At this point, you have decided, through analysis which is subjective and which is not and arguing based on that conclusion, then telling me where I am off base. It could easily be determined by another scientific theory that all you believe now to have no subjectivity, is not accurate, just like what happened with Corsi. For the record, if you look back, I have not argued things that are hard to question such as zone FO%, possession stats, pts etc. I have questioned things like QOC's overall impact, the correlation of shots against vs effectiveness to defend goals against, shooting zone% to show Jake makes good offensive decisions etc.

That said, my argument is not what stats may show, it is your aversion to someone/me questioning certain stats as fact and you telling me they cannot possibly have subjectivity.

 

[deletethis]

Ironically I am a huge fanboy of all those players when it comes to the eye test. But that may be because I am more left-winged

My personal hypothesis is that conservative people don't like defenseman like gardiner. I have done a lot of research for a psyc course I did before and it is proven that conservative people are more easily startled and are more prone to produce a hostile respons e to that anxiety. Those who are left winged aren't as easily startled or as hostile in anxiety producing situations.

Which makes sense why they wouldn't like a guy like gardiner despite the stats. He plays a very anxiety inducing game. He holds on to the puck a lot even in dangerous situations, makes risky passes and just seems to hate the safe play. For me, what he does works and he makes way more good plays then bad so I really like him. For others, he plays on the edge - it makes them worried and they don't like that. They much prefer the "chip it off the boards + knock down all the forwards on their butts where they can't do anything" style of play. I remember watching a game live last year when gardiner was fending off 2 attackers and was holding on the puck for a bit too long. In the end, nothing bad happened and he ended up making a great play and getting the puck out, but there were still a lot of people muttering their discontent - probably because that sequence scared them..

And even on this board. It seems that the gardiner dispatchers preach a more conservative approach whereas those who support him (like nithoniel and 91kadri91) are imo more left-winged.

Just something I noticed which could explain why a guy like gardiner is so polarizing amongst fans.

Please, oh, please tell me that this post was meant to be satirical. Analytics are heavily used in businesses by people who undoubtedly in the majority have more right leaning world views. Although no doubt dogmatism plays a huge role in many right leaning people, the left is populated with many whose arguments appeal to "feelings" as much if not more. There are few analytics to support many elements of humane governing associated with left leaning policies.

 

[Gary Nylund]

Please, oh, please tell me that this post was meant to be satirical. Analytics are heavily used in businesses by people who undoubtedly in the majority have more right leaning world views. Although no doubt dogmatism plays a huge role in many right leaning people, the left is populated with many whose arguments appeal to "feelings" as much if not more. There are few analytics to support many elements of humane governing associated with left leaning policies.

Where did you hear this? Sounds fishy to me.

 

[deletethis]

Where did you hear this? Sounds fishy to me.

Fishy, how so.

 

[Beleafer4]

lol, no. The divide isn't left and right. It's the crowd that relies on its gut and counts the number of mistakes Gardiner makes, and the crowd that sees his positive values amply illustrated by the results.
Next you'll be arguing that liberals are smarter then conservatives!!

Ironically that's what I do too and I liked him before the advanced stats

My hypothesis has to do with how two people see the same thing and come up with entirely different conclusions. I think its because the emotions of anxiety he stirs in them, then they rationalize why they feel an unease with him after.

And when you say "counts the number of mistakes" that rings true to me. I personally never counted many - but our board almost unanimously used to say he makes a "ton of turnovers and bone-headed mistakes" when in reality- if you put pen on paper- he makes very few. So what we perceived and what the facts show are very different. Must be a reason why we perceive he does - and I was trying to guess why by making a connection to another study. Now this is just me trying to throw darts on the board and is in no way a scientific study - but what the heck I am on hf, not presenting my thesis.

Please, oh, please tell me that this post was meant to be satirical. Analytics are heavily used in businesses by people who undoubtedly in the majority have more right leaning world views. Although no doubt dogmatism plays a huge role in many right leaning people, the left is populated with many whose arguments appeal to "feelings" as much if not more. There are few analytics to support many elements of humane governing associated with left leaning policies.

If you read many proven psyc theories about us, you would think they are satirical. Not that Sigmund freud bull - I mean the stuff they found out back when it was legal to experiment on humans (obviously now it is rightfully unethical - but because we can no longer experiment and can only do correlations, psychology is a really weak science). Seriously- we are ridiculous (or rather the better term is "irrational"). We feel first and then try to justify by making pseudologic after.

My hypothesis had nothing to do with analytics (you got that from me referencing Nith and 91 kadri) but more on how people perceived gardiner using the eye-test only. And it was also a huge reach - because in order to actualy do it properly I would have to separate those that like and dislike gardiner into two separate groups and see if they are left/right winged. Its a big leap but like I said screw it Im on hf.

Heres the original study btw:
http://onlinelibrary.wiley.com/doi/10.1111/ajps.12016/full
"Fear as a Disposition and an Emotional State: A Genetic and Environmental Approach to Out-Group Political Preferences"

 

[Cams]

I came to the thread to point out the bolded comment on Gardiner by Babcock. Yup, our all-world coach seems to agree that Gardiner is a great defenceman.

Gardiner looked absolutely great tonight, too. As did Rielly. He might not be able to turn around Phaneuf's game, but the younger pair has already clearly benefited.

Uhm...no he did NOT say that. He said he liked him. Nowhere in those comments did he say he is great. Why? Because he is not great....at least not yet. I don't see him ever being "great". Great is Robinson, Neidermeyer, Lidstrom, Orr...... currently maybe Doughty.

 

[Gary Nylund]

Gary Nylund lol
Speaking of defenseman then never panned out

To say Nylund didn't "pan out" is misleading IMO. He never had the career he could have because of injuries. Had he been healthy, he would have likely been the franchise d-man we've been looking for since Salming.

 

[91Kadri91*]

First off, I do appreciate all the stats you show on the this board, I may not put as much faith in the numbers as you do, but do appreciate the info.

Here is a couple of examples of where I find analysis contradicts itself or context of how the game is played is not taken into consideration and why I can't put as much faith in all this as others.

Contradiction
from one of the articles you suggested a poster should read on zone starts.

http://puckplusplus.com/2015/01/15/how-much-do-zone-starts-matter-i-maybe-not-as-much-as-we-thought/

From the article

"The main issue with the current approach to measuring zone starts is that the measurement is often skewed (and sometimes heavily) by the performance and talent of the player in question. Bad players tend to end up with more defensive zone faceoffs because their opponents tend to get more shot attempts against them, which leads to more opportunities for their goalie to freeze the puck and more defensive zone faceoffs. The same idea is true in reverse for good players, and it all adds up to a false correlation between the traditional zone start measure and possession numbers."

According to this, QOC matters greatly in shot attempts, possession and zone time, but we are told to disregard this over a season because it all evens out and has little affect on overall numbers.

That's not what they're saying (I don't even know how you got there). They're saying that good players will have better zone-start metrics, because they'll push the play up ice, and they're mid-shift face-off numbers will look better. At the same time, bad players will have more defensive zone starts, because the ice will be slanted in the oppositions favour and, as a result, there will be more shots against (which causes whistles that create defensive zone face-offs). Even if that wasn't the case, the argument is that zone-starts have a very limited effect on performance, so this finding would refute your QoC claims, not assist them.

Why do they consider QOC here?...because their theory falls apart without it. Again, subjectivity is used in order to determine what is important and what isn't in order to reach a desired result.

You mean your argument would fall apart? Their argument is that traditional zone-starts aren't a legitimate measure of zone-starts, because most changes happen 'on the fly', and that when you consider that, the actual value of zone-starts is questionable at best (over a large sample).

Context

This is an article linked on the article above page.

http://www.todaysslapshot.com/from-the-ice/one-timers/nhl-defensemen-who-should-be-trusted-less/

In this article the writer gives what he considers definitive facts on which NHL d-men should be used less or are ineffective when holding a lead. The two stats used to determine this...TOI while holding a lead and Corsi.

You do know there's a difference between QoC and Usage, right? I use usage metrics all the time.

When holding a lead late in a period or especially late in the game...shot attempts for, are not the main focus when defending the lead and controlling the puck is not as much a concern. It's get the puck out, dump it in, one man forecheck, no dangerous plays, be safe with the puck. Of course possession stats and shot attempts will diminish for the team defending the lead.

Well, if that's the case, Gardiner does a damn good job in these situations, despite the listed attributes being his apparent shortcomings:

Also here is where QOC has to be considered. A guy like Weber, is going to be on the ice every time the other coach puts out his best

And he'll be put out there when trying to score too, right? If he's such a great defenseman, he'll be out on the ice in every situation.

The Predators spent 1193 minutes defending the lead, and 1158 minutes seeking the tie, this season.

which will happen often when trying to tie a game, so of course he is out there more to "defend". He is not put out to score or create offence, he is put out to defend his end, win the puck battles and dump it out...protect the lead and kill the clock.

Weber has spent 1133 minutes defending the lead over the past three seasons, and 1274 minutes trailing.

Try again.

The second defending pairing will see more of the second tier offensive players from the trailing team. An easier time of it because QOC is less and the defending team will likely have better possession time and may even create some offence because their QOC is lower and therefore, does not create as many shots against

Nobody is arguing that QoC doesn't matter in short-samples: it does. The argument- it's not really an argument, there's just a few people who can't accept facts- is that over the course of a season (we're using seasons worth of data) QoC's impact is negligible, because the difference in QoC between the vast majority of players is virtually non-existent.

For example, the difference between Gardiner's QoC and Phaneuf's QoC last season was less than 2 goals (barely more than 1; 1.1 or so) per 1200 minutes played (only 79 players with 1200+ MP at even-strength last season; only 7 forwards).

...as the other article mentions. So sure, some guys stats will look better.

With a short-sample, yes.

In all honesty though, who would you rather have on the ice protecting the lead...Suter and Weber or Schenn and Krug?

I'd rather have Erik Karlsson. You know, the guy who so many insist is incapable of playing defense.

Stats say Suter and Weber should not be counted on as much and Schenn and Krug should get more time.

And they're right. What a shocker.

That's not to say Krug and Schenn should receive Suter and Weber's minutes, just that Suter and Weber clearly shouldn't be receiving the minutes they are receiving in those situations, and Krug and Schenn should be receiving more minutes.

Also, better teams will have the lead more often late in games and therefore the stats get skewed.

Sure, but significantly more time is spent with the game tied than either leading or trailing individually. You can also score-adjust the metrics (waronice.com), if you so choose.

This is just a couple of examples of why I will respond with some of the things I do when confronted with stats to prove a point...statistical analysis is determined by an individuals desire to prove a point.

And their ability to do so. You can't prove a point where one doesn't exist.

Even the formulas are created this way and many will even say, "when we project this or that we can see how"...No, projections are not facts, they very often do not turn out as assumed when put into practice.

None of what has been discussed is a 'projection'. Everything being discussed is how Gardiner actually performed, and what the difference between actual levels of QoC really is.

This is why there are so many failures prior to success, because projections were wrong.

Most of which were based on the 'eye-test'.

I fail to see what this has to do with the discussion.

So to completely discount the eye test for evaluation as some suggest, makes little sense to me. They should be used together, not one being completely more definitive than the other.

The statistics show what actually happened, your eyes perceive what you think happened. I'm not suggesting that trained professionals should be out of a job because of the recent influx of statistics into the sport, I'm merely suggesting that the opinion of an inherently biased individual who watches the game with particular expectation and preferences in mind (and no training properly determining what is/is not important), is significantly less indicative of actual performance than the results measured by statistics.

The only true way to know how a player would react in a different role is to put him in that role and actually play the games and see what happens.

No one's denying that. The statistics show what happened in those situations when the games were played, not what would happen.

He may do as expected, he may fall flat on his face.

Or (this is the most likely) he may have done exactly what he did, which the statistics will show (to a degree). Publicly available statistics aren't yet at a point where they can measure the entire game (SportLogIQ is, but again, not really publicly available) and how the player produced what he did, but they do a good job of determining overall results.

Stats can often contradict each other

They certainly can, which is why I (and many others) use a numbers of statistics to prove a point. When they're all giddy on a particular player, it probably means that player is pretty damn good.

and it is we who decide which is the right one, which is very subjective

Not really. If a statistic is clearly outlandish, then a peer review will quickly denounce it. If it's not, then it's one of many statistics one can use to prove a point. I haven't seen anyone use a singular stat to prove a point (in this thread at least).

no different than the eye test that way.

Very different from the eye test that way.

First the eye-test is subject to the eyes limited ability to view the entirety of the game and situation, then it's subject to the minds limited ability to process and remember it, and then (this is an oversimplification, but Nithoniniel and I have posted numerous articles on the issue) it's subject to a myriad of cognitive biases (confirmation, negativity etc). After that's all done, we choose what is and is not important, which is the only level of bias statistics share with the eye-test.

 

[RLF]

That's not what they're saying (I don't even know how you got there). They're saying that good players will have better zone-start metrics, because they'll push the play up ice, and they're mid-shift face-off numbers will look better. At the same time, bad players will have more defensive zone starts, because the ice will be slanted in the oppositions favour and, as a result, there will be more shots against (which causes whistles that create defensive zone face-offs). Even if that wasn't the case, the argument is that zone-starts have a very limited effect on performance, so this finding would refute your QoC claims, not assist them.



You mean your argument would fall apart? Their argument is that traditional zone-starts aren't a legitimate measure of zone-starts, because most changes happen 'on the fly', and that when you consider that, the actual value of zone-starts is questionable at best (over a large sample).



You do know there's a difference between QoC and Usage, right? I use usage metrics all the time.



Well, if that's the case, Gardiner does a damn good job in these situations, despite the listed attributes being his apparent shortcomings:

And he'll be put out there when trying to score too, right? If he's such a great defenseman, he'll be out on the ice in every situation.

The Predators spent 1193 minutes defending the lead, and 1158 minutes seeking the tie, this season.



Weber has spent 1133 minutes defending the lead over the past three seasons, and 1274 minutes trailing.

Try again.



Nobody is arguing that QoC doesn't matter in short-samples: it does. The argument- it's not really an argument, there's just a few people who can't accept facts- is that over the course of a season (we're using seasons worth of data) QoC's impact is negligible, because the difference in QoC between the vast majority of players is virtually non-existent.

For example, the difference between Gardiner's QoC and Phaneuf's QoC last season was less than 2 goals (barely more than 1; 1.1 or so) per 1200 minutes played (only 79 players with 1200+ MP at even-strength last season; only 7 forwards).



With a short-sample, yes.



I'd rather have Erik Karlsson. You know, the guy who so many insist is incapable of playing defense.



And they're right. What a shocker.

That's not to say Krug and Schenn should receive Suter and Weber's minutes, just that Suter and Weber clearly shouldn't be receiving the minutes they are receiving in those situations, and Krug and Schenn should be receiving more minutes.



Sure, but significantly more time is spent with the game tied than either leading or trailing individually. You can also score-adjust the metrics (waronice.com), if you so choose.



And their ability to do so. You can't prove a point where one doesn't exist.



None of what has been discussed is a 'projection'. Everything being discussed is how Gardiner actually performed, and what the difference between actual levels of QoC really is.



Most of which were based on the 'eye-test'.

I fail to see what this has to do with the discussion.



The statistics show what actually happened, your eyes perceive what you think happened. I'm not suggesting that trained professionals should be out of a job because of the recent influx of statistics into the sport, I'm merely suggesting that the opinion of an inherently biased individual who watches the game with particular expectation and preferences in mind (and no training properly determining what is/is not important), is significantly less indicative of actual performance than the results measured by statistics.



No one's denying that. The statistics show what happened in those situations when the games were played, not what would happen.



Or (this is the most likely) he may have done exactly what he did, which the statistics will show (to a degree). Publicly available statistics aren't yet at a point where they can measure the entire game (SportLogIQ is, but again, not really publicly available) and how the player produced what he did, but they do a good job of determining overall results.



They certainly can, which is why I (and many others) use a numbers of statistics to prove a point. When they're all giddy on a particular player, it probably means that player is pretty damn good.



Not really. If a statistic is clearly outlandish, then a peer review will quickly denounce it. If it's not, then it's one of many statistics one can use to prove a point. I haven't seen anyone use a singular stat to prove a point (in this thread at least).



Very different from the eye test that way.

First the eye-test is subject to the eyes limited ability to view the entirety of the game and situation, then it's subject to the minds limited ability to process and remember it, and then (this is an oversimplification, but Nithoniniel and I have posted numerous articles on the issue) it's subject to a myriad of cognitive biases (confirmation, negativity etc). After that's all done, we choose what is and is not important, which is the only level of bias statistics share with the eye-test.

1. No they do not come right out and say QOC must be considered, but talk about quality of player in question and where each type of player will likely start a faceoff. Since the other team is playing against this player and poor players spend more time in their end(poor possession) give up more shots against and vice versa for good players, (which should be no surprise) and he feels this needs to be accounted for and adjust for this. It doesn't take a big leap to the tougher the competition you play, the more likely you will play in your end and give up more shots. The weaker the competition, the more time you spend with the puck in their end. He is talking about quality of players making a difference. I know you will say small sample yes, but over a year it has little significance. Coaches deploy different styles of play, strategies etc which also affect these outcomes.
Nevermind when a team has a decent lead they are more likely to roll 4 lines more evenly to keep players fresh and are not as worried about line matches at this point, which will skew the results as well.

2. My argument falls apart?...even the author says he is not sure how accurate his formula really is.

3.What does what you do and what they did in the second article have to do with what I was saying? Gardiner's usage is also very low, and why you bring him up I don't know, because I didn't.

4. Yes Weber is put out to score as well. What does that have to do with me saying he is used more well defending and in tougher defending situations? I have no idea why you deflect to that.

5. And Schenn only spent 918 leading and Krug 849...200-300 less minutes than Weber, which was the point, not even accounting for game situations I was referring to. So what is it I need to try again on?

6. We are talking small sample here though for QOC... I was referring to only while holding a lead and I also said situational like last minute of periods and games. So going back to over the course of a season is just another deflection away from situational play to try and prove QOC has little effect on results.

7. Why even bring up Karlsson? He has nothing to do with Weber/Suter to Schenn/Krug.

8. What a shocker huh. Try and tell the Flyers, who would love to move Schenn he is being missed used and deserves more minutes, especially to hold a lead against top QOC. I wonder why they haven't listened to their own stats guys on this?

9.Significantly more time with game tied huh. Weber 1311, Schenn 893, Krug 780. I will let you compare to leading and trailing minutes. Can you show me how tied is significantly more please.

10. Can't prove a point where one doesn't exist? You mean like you saying they spend significantly more time playing tied?

11. Projection...is when someone says certain players deserve more time in certain situations or are projected as certain quality because stats suggest they should or are. Should they play those minutes and excel, then fine, stats have proven to be right. But some use what they say stat will project as fact, like the second article or some on here.

12. Statisticians are wrong on the outcomes often. Running every day was supposed to be good for you, now you will die earlier than the guy who doesn't exercise. Too much salt was bad, now we should eat more than before thought. All conclusions made by statistical research. The point is that even in hockey. all statisticians do not agree on best way to evaluate all stats. Like the first article where he says traditional way is not accurate, but also says not sure if his way is truly accurate. Stats can contradict often, so hard to use them as solid fact without discrepancy.

13 I was not talking about what they show as much as what peope will project from them. Like saying this guy would do just as well if given the opportunity, we cannot possibly know this. I am not sure why that point of view is so hard to understand.

14. My point was that stats contradict which you agree, You say you use many to determine a players value, but it is you who decides which ones to use which is my point that people use what they decide is pertinent to prove a point. I am sure you discard the ones you do not feel has as much value and draw your conclusions from what you feel is the right usage. For example, you believe the article you showed as the best way to analyze FO starts...someone else may disagree causing the two of you to have different opinions.

15. Because we choose which statistical data to use, is why I said it is subjective and so is an eye test which makes them similar in that way.

You try so hard to disprove anyone who says stats aren't the be all end all because they cannot possibly account for everything in a team game that is coached differently by each team, played with different quality of players and moves extremely fast. I am not discounting stats, but they are hardly conclusive when every minute someone is coming out with a new way to evaluate the same thing. You spend a lot of time on them and that is your right to believe all you read. It is mine to take it with a grain of salt and draw my own conclusions.

 

[theIceWookie]

lol, no. The divide isn't left and right. It's the crowd that relies on its gut and counts the number of mistakes Gardiner makes, and the crowd that sees his positive values amply illustrated by the results.
Next you'll be arguing that liberals are smarter then conservatives!!

I think you just inadvertently added proof to his argument hahaha.

Not that I support the argument, though I do find it a fairly interesting theory at face value.

 

[theIceWookie]

And that's where I'm overly done with this particular conversation. Sorry for the excessive grumpiness. I'm excessively grumpy.

I must say I've always appreciated your posts and your point of view.

But this comment, oh boy, this comment is my favorite. I like you even more now haha.

 

[Swayze*]

Wow so all of this data boils down to Jake is OK at 5 on 5 and plays lots of 5 on 5 minutes. "He's great at shot suppression" but can't be trusted on the PK and gives up a ton of goals.

Well guess what. He gets those minutes as he can't be trusted to play special teams. A defenseman who can't play on the PP or PK is pretty much unheard of.