OK, read this and realized my last comment was pointless. All the stuff you guys are saying here about team stats, about plus minus and the flaws in corsi that supposedly ruin any conclusions about QOC are seriously whack.
- JTanner

Let it go James...we're discussing and we know your point of view...you've jumped in head first and it's too late for you, we are still deciding if diving into an empty pool head first is a wise decision.

Again, that's a fair statement.

Look at any Corsi or Fenwick ranking for any year and tell me you should use those stats for anything except to laugh at the proponents.
- Aetherial

Why laugh at those? Given that like 5 or 6 of the last 12 stanley cup winners finished 1st or 2nd in Corsi for in the regular season, and I think 8 or 9 of 12 finished top 10, there is a correlation to Corsi and winningnthe stanley cup.

Why laugh at those? Given that like 5 or 6 of the last 12 stanley cup winners finished 1st or 2nd in Corsi for in the regular season, and I think 8 or 9 of 12 finished top 10, there is a correlation to Corsi and winningnthe stanley cup.
- j.boyd919

I think he's talking individual Corsi.

I think he's talking individual Corsi.
- Garnie

Whoops. Mah bad.

When people talk about "driving" possession, they mean a player whose WOWY s (with you without you) numbers are good. So if every player who plays with Taylor Hall receives a significant bump, then we know that the added value of him is that you can play him with pretty much anyone and those guys turn into top line players.
- JTanner

So no one, has ever, used Corsi or Fenwick as a proxy for "drives possession"?

... and didn't you just describe Corsi/Fenwick Relative to TM?

Why laugh at those? Given that like 5 or 6 of the last 12 stanley cup winners finished 1st or 2nd in Corsi for in the regular season, and I think 8 or 9 of 12 finished top 10, there is a correlation to Corsi and winningnthe stanley cup.
- j.boyd919

edit: nm, I just saw your most recent post.

edit: nm, I just saw your most recent post.
- Aetherial

You talking goal differential, or an aggregate +/- of all players on a team?

Obviously goal differential is the most important stat in hockey since the point of the game is to outscore your opponent.

But a lump sum +/- is equally silly as an individual +/- since it doesn't at all take into account a teams PP or PK abilities.

You have no idea. The data being collected is fine, and since huge sample sizes are the goal of all statisticians any improvements would be extremely marketable.

This one of those things cynical people say because it sounds good, but it's full on ridiculous.

Honestly MLJ, I have no idea, even to this day if you are serious or the ultimate troll.
- JTanner

The irony of this statement is almost painful.

You talking goal differential, or an aggregate +/- of all players on a team?

Obviously goal differential is the most important stat in hockey since the point of the game is to outscore your opponent.

But a lump sum +/- is equally silly as an individual +/- since it doesn't at all take into account a teams PP or PK abilities.
- smellmyfinger

Again... lots of people pointing out out the flaws of +/- (and I don't disagree)

Nobody acknowledging the flaws in the Corsi rankings which have been pointed out, and summarily ignored, for years now.

The irony of this statement is almost painful.
- Aetherial

Re: the first bolded... Lets have a little draft style pool based on goals and assist. I will pick my own players, yours can be randomly selected.

Re: the second bolded... Go look at Corsi Rel Tm, or any Corsi or Fenwick. Do the players rankings line up with the players likely to score goals? If the model fails, even once, then it should be questioned. If it epic-ly fails, multiple times, then it should be thrown away.

Re: the third bolded... Mathematical certainty? Holy crap. There is so much wrong with your last statement I don't even know where to begin. It is laughable.

You CONSTANTLY call your opinions facts, and everyone else's counter argument stupid.

You should be relieved when I call you a troll because if I accept that you are serious with a statement like that, then the alternative conclusion is much worse.

Patrick Kane... Go look at his Corsi stats last year. I can pick multiple players from ANY year that are ranked in dumb positions. I rest my case.
- Aetherial

In my opinion you are right ...that's a fact

Again... lots of people pointing out out the flaws of +/- (and I don't disagree)

Nobody acknowledging the flaws in the Corsi rankings which have been pointed out, and summarily ignored, for years now.
- Aetherial

I'm not a proponent of individual Corsi stats either. I think you have to take a look at more than just Corsi to get a full sense of a player (or teams) worth.

And any individual Corsi stats should always be relative to take into account usage, the system the player is in and the team success with and without the player.

Plus Minus is a terrible stat, I can't believe we're still doing this.

It's based off goals which are rare and random, making any conclusions based on it meaningless.

Shots are better because there is approximately 10 or 12 shots for every goal, so you can get a huge sample size rather quickly, making your conclusions much more likely to be accurate.

There are a lot of things we can argue about, this isn't one of them.

It's not my opinion that plus/minus is a bad stat: it's a mathematical certainty.
- JTanner

They really are not that random. I agree breaking it down to shots provides a better,wider scope, but if goals are so random, why do the same players lead the league in scoring each season. Simple answer, they are better than the others.

Sometimes all you need are basic stats. Some payers require no deeper digging

OK, read this and realized my last comment was pointless. All the stuff you guys are saying here about team stats, about plus minus and the flaws in corsi that supposedly ruin any conclusions about QOC are seriously whack.
- JTanner

My problem with QoC based on Corsi is that it doesn’t account for players who can actually pull the lever on shooting % consistently.

So maybe using QoC based on Corsi works for 90% of the players in the league. But it simply doesn’t work for Rielly or anyone who consistently matches up in the elite players that can significantly impact shooting %. So before making a conclusion, you need to better understand the unique underlying data for each player; i.e. do a macro and micro analysis.

I find many of the newer metrics create more and good questions. You jump to making outlandish blanket conclusions.

Again... lots of people pointing out out the flaws of +/- (and I don't disagree)

Nobody acknowledging the flaws in the Corsi rankings which have been pointed out, and summarily ignored, for years now.
- Aetherial

I believe there are flaws in every statistic when looked at all alone, but I think that looking at them alongside the other statistics it helps paint a clearer picture. But yeah like I said yesterday Corsi alone as the end all-be all of an argument is definitely flawed. It's just part of the evaluation. A small part at that.

Re: the first bolded... Lets have a little draft style pool based on goals and assist. I will pick my own players, yours can be randomly selected.

Re: the second bolded... Go look at Corsi Rel Tm, or any Corsi or Fenwick. Do the players rankings line up with the players likely to score goals? If the model fails, even once, then it should be questioned. If it epic-ly fails, multiple times, then it should be thrown away.

Re: the third bolded... Mathematical certainty? Holy crap. There is so much wrong with your last statement I don't even know where to begin. It is laughable.

You CONSTANTLY call your opinions facts, and everyone else's counter argument stupid.

You should be relieved when I call you a troll because if I accept that you are serious with a statement like that, then the alternative conclusion is much worse.

Patrick Kane... Go look at his Corsi stats last year. I can pick multiple players from ANY year that are ranked in dumb positions. I rest my case.
- Aetherial

You lost him at Patrick Kane. Tanner calls Kane a criminal while rocking his Torres jersey.

I'm not a proponent of individual Corsi stats either. I think you have to take a look at more than just Corsi to get a full sense of a player (or teams) worth.

And any individual Corsi stats should always be relative to take into account usage, the system the player is in and the team success with and without the player.
- smellmyfinger

exactly. and the guy who creates Hero Charts offers the limitations of Corsi Rel stats:

Linemates’ CF60 away from the player of interest (weighted by time on ice with the player of interest). Although this method does a better job of dealing with collinearity than traditional on-ice/off-ice relative measures, it has its limitations. The first issue is sample size. Some skaters spend a large % of ice time with a particular linemate. This means that the sample of TOI away from the player of interest for that particular linemate will be susceptible to noise – potentially skewing RelTM measures.

Another issue is the fact that talent distribution is uneven throughout a lineup. For example, a third line center’s most common winger may spend a significant portion of his ice time with the team’s elite first line centre when he is away from the player on interest. This will unfairly penalize that particular third line center’s output simply due to the fact that he plays on a team with an elite first line center. Because of this, it is more appropriate to think of RelTM metrics as a measure of a skater’s ability relative to teammates playing the same position.

https://ownthepuck.com/

exactly. and the guy who creates Hero Charts offers the limitations of Corsi Rel stats:

Linemates’ CF60 away from the player of interest (weighted by time on ice with the player of interest). Although this method does a better job of dealing with collinearity than traditional on-ice/off-ice relative measures, it has its limitations. The first issue is sample size. Some skaters spend a large % of ice time with a particular linemate. This means that the sample of TOI away from the player of interest for that particular linemate will be susceptible to noise – potentially skewing RelTM measures.

Another issue is the fact that talent distribution is uneven throughout a lineup. For example, a third line center’s most common winger may spend a significant portion of his ice time with the team’s elite first line centre when he is away from the player on interest. This will unfairly penalize that particular third line center’s output simply due to the fact that he plays on a team with an elite first line center. Because of this, it is more appropriate to think of RelTM metrics as a measure of a skater’s ability relative to teammates playing the same position.

https://ownthepuck.com/
- Steven_Stamkos

I like the stats in Hero charts that are related to how much the coach uses the players.

In this case, the coach, is filtering out all the noise, and potentially considering other factors that no stats are even close to considering.

Of course that is never 100 predictive of success either as coaches vary in their skill and attention to detail.

I think it is a safe assumption that the group of NHL coaches, as a whole, know more than the people contributing to this discussion

I like the stats in Hero charts that are related to how much the coach uses the players.

In this case, the coach, is filtering out all the noise, and potentially considering other factors that no stats are even close to considering.

Of course that is never 100 predictive of success either as coaches vary in their skill and attention to detail.

I think it is a safe assumption that the group of NHL coaches, as a whole, know more than the people contributing to this discussion
- Aetherial

I find that hero charts are fun to use and provide some useful information. I like to use them to generally help identify the worst players on bad teams and the best players on good teams.

But they are not meant to be the bible as a comparison tool for players on different teams. Anyone who concludes Hall > Kane based on a Hero Chart makes a poorly informed conclusion.

Plus Minus is a terrible stat, I can't believe we're still doing this.

It's based off goals which are rare and random, making any conclusions based on it meaningless.

Shots are better because there is approximately 10 or 12 shots for every goal, so you can get a huge sample size rather quickly, making your conclusions much more likely to be accurate.

There are a lot of things we can argue about, this isn't one of them.

It's not my opinion that plus/minus is a bad stat: it's a mathematical certainty.
- JTanner

This is where you go wrong. If you want to measure possession, than surely, shot metrics is much better. +/- shouldn't be used to measure possession anyways. It also has the same problem as Corsi does, which is a bunch of false data. The sample is hopefully over 82 games. If +/- is used for what it actually says, then it's an okay stat. It is your opinion, and +/- is as much of a mathematical certainty as Corsi is. It simply counts when a player is on the ice for a goal for, or against. Pretty simple.

You have no idea. The data being collected is fine, and since huge sample sizes are the goal of all statisticians any improvements would be extremely marketable.

This one of those things cynical people say because it sounds good, but it's full on ridiculous.

Honestly MLJ, I have no idea, even to this day if you are serious or the ultimate troll.
- JTanner

James, what would decide who the troll is, is who keeps the discussion on the actual topic, and who instead of attacking the opinion, attacks the poster.

Actually I do have an idea. It's pretty simple. It counts the shots attempts that occur when a player is on the ice, whether he has anything to do with it or not. It piles the same shot event in every players stat column. How you can not see the false data in that, is mind boggling.
I have not read a single reasonable explanation that refutes that. Not one. They just brush it under the rug normally, with statements like, it's not perfect, but it's the best we have. That may be true, but that doesn't change that it's flawed.

I publicly challenge you to refute it. Sample size does not cancel that out. It just piles up more and more false data.

When people talk about "driving" possession, they mean a player whose WOWY s (with you without you) numbers are good. So if every player who plays with Taylor Hall receives a significant bump, then we know that the added value of him is that you can play him with pretty much anyone and those guys turn into top line players.
- JTanner

WOWY's are also a flawed stat. One being corsi based, and secondly, undoubtedly when a player is moved to a different line, the sample size can be much smaller, as can the game situations the data is collected in.

I believe there are flaws in every statistic when looked at all alone, but I think that looking at them alongside the other statistics it helps paint a clearer picture. But yeah like I said yesterday Corsi alone as the end all-be all of an argument is definitely flawed. It's just part of the evaluation. A small part at that.
- j.boyd919

You know the key word in your statement there is helps. If you use them to help form an opinion, they're certainly useful. It's when problems with analysis, and you make statements that it's a fact that a player with a good corsi, means he is a good player, is where the trouble starts. They should be used as a percentage of coming to a conclusion on a player. The percentage is debatable. It's when people like Tanner, who doesn't understand them to begin with, uses them as the whole ball of wax. It is equally wrong to ignore analytics, as it is to rely on them solely.

You know the key word in your statement there is helps. If you use them to help form an opinion, they're certainly useful. It's when problems with analysis, and you make statements that it's a fact that a player with a good corsi, means he is a good player, is where the trouble starts. They should be used as a percentage of coming to a conclusion on a player. The percentage is debatable. It's when people like Tanner, who doesn't understand them to begin with, uses them as the whole ball of wax. It is equally wrong to ignore analytics, as it is to rely on them solely.
- MJL

I give up. It's like I am living in a twilight zone. The people who are the most wrong, with the worst ideas have started being condescending to me. A guy told me to but out, in my own blog. Another guy called a mathematical certainty "laughable."

And now this. Keep up the good work guys!!