Re-evaluating Dion
- Thread starter number17
- Start date
zeke
Well-known member
and he even admits earlier in the article that there are no real comparables for Phaneuf's usage, so clearly he tried to widen the net and let some flaws enter.
I refuse to judge the leaf players' corsis until we get a coach who can implement a decent system.
Kulemin leads the Islanders' forwards in cf% this year, and he's been the worst leaf forward in CF% the last 2yrs other than the laughable McClem/Orr/McLaren trio.
And that's even accounting for the change on Zone Start usage.
The year before randy arrived, of course, Kuly was one of the Leafs' best CF% forwards every year.
One thing the smart analytics guys are starting to realize is just how much systems can effect a player's corsi, regardless of talent level. And its not a linear correlation necessarily, or even a steady curve - instead it seems that certain systems (or non systems) coupled with certain usage can absolutely cripple a player's CF% underservedly.
Its not just that randy's system is worse than the others, but that it seems to be an outlier in today's NHL - a system intentionall completely different from what the majority of the rest of the NHL does. So when 90% of the league is doing one thing, it might be easy to compare Corsis between those teams, but the 10% doing something completely different can create bizarre outlier statistical outcomes. Outcomes which have a tough time being explained even by adjusting for zone starts or quality of competition, because those are all based on league averages in the first palce.
And I think we can see this this year already - Randy explicitly gave up (or was forced to give up) some of the tactics that made his system so unique, and the team explicitly reverted to new tactics more typical of what the majority of teams are doing out there. Randy has also eased up on his maniacal line-matching, at least a bit, meaning that player's corsis aren't as extremely distorted by qualcomp and zone starts as they may have been before. And sure it's a small sample, and the improvement has been only in fits and starts, but already we see the leafs be a more normal team in terms of possession stats....and unsurprisingly (to me) we're also seeing Phaneuf putting up very good possession numbers at the same time.
I refuse to judge the leaf players' corsis until we get a coach who can implement a decent system.
Kulemin leads the Islanders' forwards in cf% this year, and he's been the worst leaf forward in CF% the last 2yrs other than the laughable McClem/Orr/McLaren trio.
And that's even accounting for the change on Zone Start usage.
The year before randy arrived, of course, Kuly was one of the Leafs' best CF% forwards every year.
One thing the smart analytics guys are starting to realize is just how much systems can effect a player's corsi, regardless of talent level. And its not a linear correlation necessarily, or even a steady curve - instead it seems that certain systems (or non systems) coupled with certain usage can absolutely cripple a player's CF% underservedly.
Its not just that randy's system is worse than the others, but that it seems to be an outlier in today's NHL - a system intentionall completely different from what the majority of the rest of the NHL does. So when 90% of the league is doing one thing, it might be easy to compare Corsis between those teams, but the 10% doing something completely different can create bizarre outlier statistical outcomes. Outcomes which have a tough time being explained even by adjusting for zone starts or quality of competition, because those are all based on league averages in the first palce.
And I think we can see this this year already - Randy explicitly gave up (or was forced to give up) some of the tactics that made his system so unique, and the team explicitly reverted to new tactics more typical of what the majority of teams are doing out there. Randy has also eased up on his maniacal line-matching, at least a bit, meaning that player's corsis aren't as extremely distorted by qualcomp and zone starts as they may have been before. And sure it's a small sample, and the improvement has been only in fits and starts, but already we see the leafs be a more normal team in terms of possession stats....and unsurprisingly (to me) we're also seeing Phaneuf putting up very good possession numbers at the same time.
zeke
Well-known member
Fortunately, the possession stats give us far many more times the data points, which lets us deal with most of the issues of +/- ....namely quality of teammates, quality of competition, and usage. That's exactly why they are better. If we could adjust +/- for those factors, it would be a better stat too....though with only a small fraction of the data points, still not very good.
But no doubt they are far from perfect.....especially when a team plays a specifically and significantly different system than the majority of teams play in the league.
The leafs' possession numbers last year were insanely stupid. Its obvious that this collection of players have the talent to do better than that.
That being said - if we want to analyze how a team is playing (not its talent level, simply how well it is playing), then these possession numbers really are by far the best numbers to look at. It is really, really hard for a team to beat their corsi other than with an outlier in the non-corsi factor (goaltending) even in the short term, and nearly impossible over the longterm.
But no doubt they are far from perfect.....especially when a team plays a specifically and significantly different system than the majority of teams play in the league.
The leafs' possession numbers last year were insanely stupid. Its obvious that this collection of players have the talent to do better than that.
That being said - if we want to analyze how a team is playing (not its talent level, simply how well it is playing), then these possession numbers really are by far the best numbers to look at. It is really, really hard for a team to beat their corsi other than with an outlier in the non-corsi factor (goaltending) even in the short term, and nearly impossible over the longterm.
number17
Mod Squad
All of them good points.
So forget advanced stats for a sec. What IS Dion Phaneuf from objective eyes of a hockey fan?
To me, Dion is an offensive defenseman who is capable of 10-15G, 40-50 points a season offensively. He's a good PP pointman, but he's a better shot than he is as a QB, which means he's better as the one wiring up and shooting than the one distributing and shooting the puck.
Defensively Dion is over-average only, but a great shut-down defensive defenseman he is not. He is average in defensive zone coverage, 1 on 1 battle, and his physical presence is better used in open ice hits than physical battles along the board or in front of the crease.
He CAN log a lot of minutes, he CAN play in PK but you gotta wonder if he may be better off playing PK2 instead of PK1, and let someone else do the heavy lifting in ES QoC defensive duty so that he can focus on the offense. Or, at the very least, give him a capable D partner that would give him some leeway in gambling offensively, instead of constantly having to cover for his partner defensively (Holzer, Kostka, Franson, Gunnar with bad hip ... on and on and on)
So forget advanced stats for a sec. What IS Dion Phaneuf from objective eyes of a hockey fan?
To me, Dion is an offensive defenseman who is capable of 10-15G, 40-50 points a season offensively. He's a good PP pointman, but he's a better shot than he is as a QB, which means he's better as the one wiring up and shooting than the one distributing and shooting the puck.
Defensively Dion is over-average only, but a great shut-down defensive defenseman he is not. He is average in defensive zone coverage, 1 on 1 battle, and his physical presence is better used in open ice hits than physical battles along the board or in front of the crease.
He CAN log a lot of minutes, he CAN play in PK but you gotta wonder if he may be better off playing PK2 instead of PK1, and let someone else do the heavy lifting in ES QoC defensive duty so that he can focus on the offense. Or, at the very least, give him a capable D partner that would give him some leeway in gambling offensively, instead of constantly having to cover for his partner defensively (Holzer, Kostka, Franson, Gunnar with bad hip ... on and on and on)
BeLeafer
Well-known member
True and true for most stats on individual performance. It's like pointing out that Bozak's offensive numbers are closely tied to linemates, TOI and PP TOI.
I agree with this. I think they are a good measure of team performance and line-by-line performance.
MindzEye
Wayward Ditch Pig
I don't think it's really fair to say that Dion is only average defensively. Most of the **** ups that are loudly attributed to him, were him being caught out of position due to someone else's **** up.
How often do we see him abused by the star players he's matched up on? If he was the defender that we're told that he is, those star players would be abusing us a lot more often.
How often do we see him abused by the star players he's matched up on? If he was the defender that we're told that he is, those star players would be abusing us a lot more often.
zeke
Well-known member
remember that even ignoring phaneuf's unique usage and the team's unique system, phaneuf fared extremely well in other newer advanced stats - namely preventing zone entries and in efficient zone exits.
and again, i don't think its a coincidence at all that his possession numbers have gone from awful all the way up to very good this year in a more normal system with more normal usage.
it might be a blip but 10% of a season is not a tiny sample and i highly doubt his possession numbers looked anywhere near this good over any 9gm sample in the last two years.
and again, i don't think its a coincidence at all that his possession numbers have gone from awful all the way up to very good this year in a more normal system with more normal usage.
it might be a blip but 10% of a season is not a tiny sample and i highly doubt his possession numbers looked anywhere near this good over any 9gm sample in the last two years.
Wayward DP
Well-known member
I'd like to see more offensive production from Dion, but pretty sure his splits from last year showed his ES production remaining fairly consistent, just his PP production was down. Which in a year where we had a very good PP I'm not too upset about, but I think the biggest issue is his PP usage (and being paired with Franson as opposed to Reilly/Gardiner).
And yeah, the numbers otherwise speak for themselves. We've got a lot bigger problems than Dion.
And yeah, the numbers otherwise speak for themselves. We've got a lot bigger problems than Dion.
leafman101
Well-known member
Phaneuf really isn't an "offensive dman" anymore. For the most part he has scored in the 30 point range here. Obviously he has skill, but he has focused a lot on the other side of the ice.
He has had pretty insane usage the last couple years and held his own, so I don't think you can call hinm average defensively either. Most dmen in the league couldn't play 23+ minutes a game, against the other teams top line, every single shift, and not get his ass handed to him.
He's a great all around player.
He has had pretty insane usage the last couple years and held his own, so I don't think you can call hinm average defensively either. Most dmen in the league couldn't play 23+ minutes a game, against the other teams top line, every single shift, and not get his ass handed to him.
He's a great all around player.
JackBurton
Well-known member
I was a big Phanuef defender but I really dislike what I've seen of him since that collapse started last season.
Habsy
Not playing around anymore.
Funny. #1 D are supposed to make their partners look better not the other way around.
It never seems to be Dion's fault. Maybe he's just jumped the shark already.
Mike Komisarek needs to send Markov Thank you cards for the NHL career Markov gave him.
number17
Mod Squad
But are they?
Phaneuf's CF% = 51.7%, good for a very mediocre 67th in the league. If we are to look at CA/60 (which is a more defense-stat) he drops to 116 at 57.5. His FF% is 51.9, good for 105th, and his FA/60 is 42.9 at 118th in league. (all these stats for 5v5 close)
All these stats suggest Phaneuf is at best an average plus defensive defenseman. I do agree with the argument Carlyle's system will hurt any player's possession stats, but between the stats and what we can see on the ice, I just can't come to the conclusion Phaneuf is 'great' defensively.
leafman101
Well-known member
You can't ignore usage though. He is putting up those average numbers, playing against some of the toughest competition in the NHL.
So he is holding the toughest competition in the nhl to average production numbers. The average defenseman gets dominated by those guys, not neutralizes them.
zeke
Well-known member
Well, if you adjust for zone usage (which is done a little ham-fistedly but still usefully by ignoring any data points that come within 10 seconds of a faceoff), he looks much better than that at 55.7%, which ranks him 27th amongst 155 qualifiers. Franson ranks 15th at 57.9%. Rielly ranks 23rd at 56.7%.
Even if you broaden that to all 5v5 situations instead of just close situations, Dion's 54.3% ranks 46th out of 186 qualifiers. Franson ranks 61st at 52.9%.
and these numbers aren't adjusted for either quality of opponent or quality of their own team, both of which would likely drive Dion's numbers up higher.
JackBurton
Well-known member
So we're happy with Phaneuf and Franson this season because of geeky advanced stats no one cared about 5 years ago? Franson can't get shipped-out fast enough.
If we want to seriously contend, we need better than that.
If we want to seriously contend, we need better than that.