But they aren't always involved in specific match-ups. If you look at how much time a player spends on the ice vs. a specific opponent, especially during the regular season, those match-ups barely exist. If a defenseman is playing 25 minutes a night and a forward he's matched up against plays 18 minutes per night, how much time do you think they actually spend on the ice together? 10 maybe?

Those match-ups really just don't happen often enough to have any real measurable impact over the long term.
- PepinoPamplemousse

Those matchups certainly do exist and 10 minutes against each other (40-60% of TOI) would be significant. If those matchups don't exist then what exactly are the coaches doing standing behind the bench?

I have say, watching McDonagh the past 5 years, I'd have to say he definitely spends most of his minutes every single game against the top stars. You get the inevitable mismatch due to icings or special teams or the inability to control the puck to get a change on the fly, but the majority of time is still matched up.

Were talking about Quality of Competition having no bearing on a players performance and shouldn't be factored into how good they are or not. But it does. If a certain player's icetime is used 90% of the time against Ovechkin or Crosby or Kane or any other top line in the league, his stats will have a different weighted outcome than a 3rd pairing dman against the 3rd or 4th line.
- xcheckmajor

But nobody does that, much as we want to pretend they do, they just don't.

But nobody does that, much as we want to pretend they do, they just don't.
- PepinoPamplemousse

But they do. You are basically saying matchups don't exist in the NHL. Every team randomly rolls their lines without regard to what the other bench is doing.

Those matchups certainly do exist and 10 minutes against each other (40-60% of TOI) would be significant. If those matchups don't exist then what exactly are the coaches doing standing behind the bench?
- Snowblind

That's in ONE game. One. Over 82 games that simply doesn't happen for a huge variety of reasons.

I have say, watching McDonagh the past 5 years, I'd have to say he definitely spends most of his minutes every single game against the top stars. You get the inevitable mismatch due to icings or special teams or the inability to control the puck to get a change on the fly, but the majority of time is still matched up.
- xcheckmajor

Yeah I really do not know what exactly this "myth" is. Unless I have been hallucinating for the past couple of decades or have fallen victim so some sort of mass psychosis, top defensemen and, to a lesser extent, top defensive centermen are deployed against top offensive players as much as possible in the regular season AND playoffs. In the playoffs, more lines and players will be matched up as the teams zero in on each other, but the broad strokes (McDonagh vs. top forwards, top defensemen vs. Tavares, Toews/Kopitar vs. top forwards) are always occurring. Some teams have Crosby AND Malkin to worry about, other teams, like the Isles, have Boychuk AND Hamonic who are the best options against different players, but matchups always are made as much as possible barring extensive garbage time.

But they do. You are basically saying matchups don't exist in the NHL. Every team randomly rolls their lines without regard to what the other bench is doing.
- xcheckmajor

No, I'm saying matchups exist, but not anywhere near to the degree that we actually think they do.

It's hard to wrap your head around, I know, but it's right.

Yeah I really do not know what exactly this "myth" is. Unless I have been hallucinating for the past couple of decades or have fallen victim so some sort of mass psychosis, top defensemen and, to a lesser extent, top defensive centermen are deployed against top offensive players as much as possible in the regular season AND playoffs. In the playoffs, more lines and players will be matched up as the teams zero in on each other, but the broad strokes (McDonagh vs. top forwards, top defensemen vs. Tavares, Toews/Kopitar vs. top forwards) are always occurring. Some teams have Crosby AND Malkin to worry about, other teams, like the Isles, have Boychuk AND Hamonic who are the best options against different players, but matchups always are made as much as possible barring extensive garbage time.
- Snowblind

Way more than you know.

That's in ONE game. One. Over 82 games that simply doesn't happen for a huge variety of reasons.
- PepinoPamplemousse

But it does. Every team has a top line that scores more than the others. Most teams routinely try to matchup their best Dmen against those top lines. I would argue that it actually is the anomaly when it DOESN'T happen. Those few minutes a game where the top D pairing is not facing the top scoring lines is either:

1. Icing and changes can't be made
2. Special teams where the makeup of the PP units can differ
3. The game situation has changed and assignments have changed
4. Coaches have changed their lines to get their top players away from the top D
5. The Dmen are having a bad game and the coaches have changed the matchups

All of which is out of the normal strategy of how a coach wanted to play.

This is why coaches make in game adjustments. The Hurricanes (Tulsky included) provide the coaching staff with reports between periods to help with in-game adjustments that sometimes pay dividends. However, in a long term picture, the impact is still pretty minimal when it comes to evaluating how you expect a player to perform.

QoC has a very real impact in the immediate evaluation of a player over a small stretch of games. Over a full season, however, it quickly moves towards being meaningless in player evaluation.
- PepinoPamplemousse

That's essentially what I was saying along with the link I posted. Not sure what your point is in saying I "clearly didn't read the article" when we're agreed on the usage of QoC...

No, I'm saying matchups exist, but not anywhere near to the degree that we actually think they do.

It's hard to wrap your head around, I know, but it's right.
- PepinoPamplemousse

Proof?

No, I'm saying matchups exist, but not anywhere near to the degree that we actually think they do.

It's hard to wrap your head around, I know, but it's right.
- PepinoPamplemousse

There are also 4 forward lines and 3 defensive pairs. Simple math will show that its impossible to play matchups 100% of the time. There are also teams with more than 1 potent line to worry about. The point is, your top defender WILL always get less minutes against the worse forwards on the other team. They will always play the majority of the game against the if not always the best, but the better part of the opponent's roster.

I am willing to bet you any amount of money that McDonagh, Keith, Doughty, Weber, Chara, etc... ALL play the majority of their minutes against the other teams' top 6.

Proof?
- Snowblind

http://nhlnumbers.com/201...of-quality-of-competition

Read it. He explains far more articulately than I can.

Way more than you know.
- PepinoPamplemousse

But putting all faith into an essentially meaningless statistic that is dubiously tabulated, completely flattening their relative quality and then deriving secondary and tertiary statistics based on those original numbers and taking it as gospel that they represent reality better actual results, eyes and common sense? That's having two feet firmly planted on the ground!

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

Read it. He explains far more articulately than I can.
- PepinoPamplemousse

There is also the idea that the game of hockey is created under laws that create scoring. There is always a disadvantage in defending versus being on the offense. Thats how scoring exists in the game. Just because the best defenders are against the best offensive players, doesn't nullify the game out.

But putting all faith into an essentially meaningless statistic that is dubiously tabulated, completely flattening their relative quality and then deriving secondary and tertiary statistics based on those original numbers and taking it as gospel that they represent reality better actual results, eyes and common sense? That's having two feet firmly planted on the ground!
- Snowblind

Correcting for the quality of competition that a player faces is surprisingly rare. One reason for this is that the analysis isn’t simple at all. A simple regression shows absolutely no relationship between a player’s Corsi or relative Corsi and his quality of competition, and multivariable analysis suggests competition has just a very small impact. This is counter-intuitive, and it is often presumed that the reason we don’t see an effect is because usage and skill are linked. If goons mostly face other goons, they’ll break even (on average) and the simple statistical look will conclude that competition doesn’t matter much – we would see that even the guys who get to play against goons all the time only break even.

So if we just ask “how does a guy do if he faces goons most of the year”, the answer will be “not so great”, because he’s probably a goon himself. We never get to run the experiment where Pavel Datsyuk plays against goons for the better part of a season. But he does get the occasional shift against them, and we can finally answer this question by looking at the data at the individual shift level instead of at the season level. We can reframe the question as “how much better does the average player do in the shifts where he faces goons than the shifts where he faces top lines” and thereby remove the influence of how often he actually faces each opponent and how good he is.

Here is a plot of how even strength results changed as a function of quality of competition in 2010-2011. It is immediately evident that quality of competition is of course very important – we can now see that if the Datsyuks of the world played against goons all year, their teams would get almost 2/3 of the shots while they were on the ice.

While competition certainly does play a big factor in determining how a player will do in any given shift, with these competition metrics we see nobody with usage extreme enough to require a major correction factor. Using the curve for the average player on the first chart, we can calculate that Nodl’s 95th percentile usage is only harsh enough to bump an average player’s Corsi down to 49.5%, while Betts’ 13th percentile usage would be soft enough to allow an average player to post a 50.6% Corsi.

The analogous situation in zone starts would be if everyone’s offensive zone start percentage were between 48% and 52%; such small corrections are scarcely worth the effort, and a person who ignored competition when evaluating players would not be wrong by much. Quality of competition is very similar to shot quality: it plays a huge role in individual shifts/shots, but over the course of a season the differences across teams and players are small enough that it can usually be neglected.

These competition metrics provide valuable insight into what a coach thinks of a player and how he tries to use them, but in practice they do not show differences large enough to have significant impact on the player’s results.

Seriously just read the (frank)in thing. The guy has worked on DNA sequencing, nanotechnology, has been involved in analytics at the NHL level for at least 5 years and has a Ph.D. in Chemistry from UC-Berkeley

I don't know why I would expect someone to read one blog and be convinced by it if they won't read another that continuously references it.

So if we just ask “how does a guy do if he faces goons most of the year”, the answer will be “not so great”, because he’s probably a goon himself. We never get to run the experiment where Pavel Datsyuk plays against goons for the better part of a season. But he does get the occasional shift against them, and we can finally answer this question by looking at the data at the individual shift level instead of at the season level.

Does that excerpt not completely contradict your point and validate those of us on this thread that disagree with it?

http://hockeyanalysis.com...3/04/12/home-vs-road-qoc/

Obviously a very short list but it does show a significant QoC variance home vs away as home teams will avoid putting out best forwards against the opposition best dmen when possible

And for sample size I changed the search criteria to 2 years but wanted to compare Drouin home vs road splits.

http://stats.hockeyanalys...p=1&sort=PCT&sortdir=DESC

http://stats.hockeyanalys...p=1&sort=PCT&sortdir=DESC

Obviously this isn't specifically accounting for QoC, however given the home vs road Fenwick splits of Drouin it would certainly support that last change at home does have a fairly dramatic effect. Road he was 51.3 vs home he was 55.1

Also found a good piece on score adjusted Fenwick that really showed how the score (and time of the game) will basically throw any QoC ratio out b/c the game situation is dictating the matchup moreso than the players on the ice. Problem is I closed the damn window and can't find it again

Seriously just read the (frank)in thing. The guy has worked on DNA sequencing, nanotechnology, has been involved in analytics at the NHL level for at least 5 years and has a Ph.D. in Chemistry from UC-Berkeley
- PepinoPamplemousse

I'd also like to see a better example than the Flyers one provided as that was a team without a clear-cut top threat as they also had a breaking-out Giroux, Carter coming off a 48 goal season and Richards not far removed from an 80-point season in addition to Briere. Thus they were more of a "pick your poison" kind of team without one obvious matchup that every coach of every team would be looking for. I'd be far more interested in how Pronger's QoC compared with the other defensemen on that particular team.

So if we just ask “how does a guy do if he faces goons most of the year”, the answer will be “not so great”, because he’s probably a goon himself. We never get to run the experiment where Pavel Datsyuk plays against goons for the better part of a season. But he does get the occasional shift against them, and we can finally answer this question by looking at the data at the individual shift level instead of at the season level.

Does that excerpt not completely contradict your point and validate those of us on this thread that disagree with it?
- Snowblind

I'm glad you decided that THAT was the piece to pull out of there, but yes, it does. I was approaching it from the wrong way, I'll admit that, but in the end result is still the same.

James, serious question for you: have you ever done a blog about the accuracy of the statistics being collected for hockey are?

I seem to recall someone from the Leafs management--perhaps it was Nonis or Burke..?--who were calling into question the accuracy of the data that is collected.

Since you're a guy who leans so heavily on statistics to make decisions/judgments, I'm curious as to whether you have blogged about or looked into this topic of the accuracy of the data being collected; how it's collected, the manner and approach that various data collectors take, etc...

I'm glad you decided that THAT was the piece to pull out of there, but yes, it does. I was approaching it from the wrong way, I'll admit that, but in the end result is still the same.
- PepinoPamplemousse

Yes, in the end the result is still the same. The coaches match up their best defensive players with other teams top offensive players. The coaches use the QoC to set their lines, so how can it not be pertinent to the conversation. A second pair D-man who does not see the top end competition is NOT better than the top pair D-man who does, just because his analytics seem slightly better on paper. This whole argument makes no sense. The players who see the highest TOI and play against the top competition are your best and most indispensable players.

For the initial point that James made, can anyone say that if Gardiner was utilized in the exact same situations Rielly is, his numbers would be anywhere close to what they are? I personally doubt it.

No, I'm saying matchups exist, but not anywhere near to the degree that we actually think they do.

It's hard to wrap your head around, I know, but it's right.
- PepinoPamplemousse

Thanks for doing my job for me!!

James, serious question for you: have you ever done a blog about the accuracy of the statistics being collected for hockey are?

I seem to recall someone from the Leafs management--perhaps it was Nonis or Burke..?--who were calling into question the accuracy of the data that is collected.

Since you're a guy who leans so heavily on statistics to make decisions/judgments, I'm curious as to whether you have blogged about or looked into this topic of the accuracy of the data being collected; how it's collected, the manner and approach that various data collectors take, etc...
- --Lamp--

The statistics are collected by the NHL and are accurate enough that the margin of error is extremely low. There are errors in every single data set. Use a big enough sample size and they don't matter.

The fact is, as was previously mentioned, the tent-pole advanced hockey stats have been sifted through by mathematicians, scientists, guys who worked for NASA, brain sugereons etc. They have all said, to a man, that even if you had sensors and 100% GUARANTEED accurate stats, the results would be no different.

Also, the entire reason Corsi is the standard is because no stat correlates with winning as accurately as score adjusted Corsi does. The only problem is that when some people see a single anomalous occurrence, they think the whole thing is shot.

The point is to analyze play, not predict it. The best you can do with regards to predicting is to come up with probability. So, if you're worse than the 16th best Corsi team, the probability is you miss the Playoffs. Having a team like LA miss the Playoffs while Calgary or Ottawa makes it has the effect of making people doubt it, but no one ever claimed 100% accuracy.

Furthermore, if you could predict 52% of results you could be an amazing gambler if you knew how to manage a bankroll.