A look back at Sidney Crosby's defense over the past decade

[Flying Dego]

In no way would I try to say Ovi is better then Sid Defensively. Two different positions with roles and responsibilities but Im curious why it matters which goalies Crosby had when hes being scored on but it doesn't matter which goalies he crucified in the playoffs in his first 5 years while Ovi played Lundqvist every damn year. Crosby was torching Ottawa and NYI goalies for pts in the playoffs when Ovi played the King. It was obviously uncontrollable for Crosby and it most likely has evened out over the last 8 years or so but which goalies Crosby had backing him up shouldn't matter if everything evens out over their careers. Caps really didn't have a playoffs goalie until Holtby showed up. Do we really want to say Huet, Theodore and the rest of the journeymen and "never were" are better then Fleury was.

Kind of a funny note but the Pens always light up Lundqvist...would love it if we played him and goalie Bob every playoff.

 

[Gurglesons]

"Goals are scored randomly."

That's gotta be one of the greatest quotes in HF history.

If you don’t understand the unpredictability of goal scoring in the NHL you should not comment on the advanced statistics used in the NHL.

 

[Gurglesons]

The data absolutely doesn't say that though. You are not comprehending it. Kopitar's baseline is a better baseline defense than Crosby's - meaning he is being held to a higher standard. Same goes for Ovie and Backstrom.

Kopitar takes on a shutdown role, and penalty kills 2 minutes per game in the regular season and playoffs. Sidney Crosby has never shut anyone down, and penalty kills 1/4 as much as Kopitar in the regular season and 1/10th as much in the playoffs.

Kopitar's defense on the powerplay is also far superior to Crosby's.

You can make this argument. But, using data to make this argument doesn’t work. Which is why your original post is flawed.

You had a preconceived notion rather than wanting to view a specific stat.

Different players are used very differently. We have to normalize the data and situations. That is why 5v5 is widely used in most data studies in the NHL, because it provides an equal playground for players.

If your argument is Crosby is overrated defensively because of his usage then that is a separate argument.

 

[GreatGonzo]

The bolded is very ironic.

Ok bud

 

[GreatGonzo]

If you don’t understand the unpredictability of goal scoring in the NHL you should not comment on the advanced statistics used in the NHL.

That’s the silliest argument I’ve seen in this thread. Bravo

tell me more about how Crosby is better defensively than Kopitar

 

[GreatGonzo]

You can make this argument. But, using data to make this argument doesn’t work. Which is why your original post is flawed.

You had a preconceived notion rather than wanting to view a specific stat.

Different players are used very differently. We have to normalize the data and situations. That is why 5v5 is widely used in most data studies in the NHL, because it provides an equal playground for players.

If your argument is Crosby is overrated defensively because of his usage then that is a separate argument.

You obviously have no clue what the data says considering the conclusions you came up with. In fact you have trashed all the data that has been presented simply because it contradicts your beliefs.

It’s already been said, comparing Crosby to the best defensive forwards won’t make Crosby look strong defensively, but that’s not your argument and that’s not the “facts” that you somehow gathered from the information.

 

[GreatGonzo]

If you don’t understand the unpredictability of goal scoring in the NHL you should not comment on the advanced statistics used in the NHL.

You can’t sit there and whine about people being bias and unfair and then make a statement like that. It puts your credibility on the line. Sorry to break it to you. You also can’t make something factual just because you want it to be, and trying to suggest that goals are scored “randomly” is beyond nonsense.

 

[Empoleon8771]

To be honest, I don't know why people even bother arguing with obvious Crosby haters about Crosby. Like nothing you can say will make them change their minds, because their minds are made up because of entirely emotional reasons. In terms of RelTM xGA/60, which is the best defensive analytic we have available right now, Crosby is on par with guys like Kopitar, Toews, Kesler and Zetterberg defensively. If you look at results in terms of RelTM GA/60 from the last page, you get similar results (although I really don't like judging defensive ability based solely on goals against).

All of Crosby's actual on ice results have him as good defensively for 1Cs and on par with other good defensive 1Cs, which is where I think most reasonable people would have him. He's not on Bergeron's level or Datsyuk's level (and I don't think he's on par with ROR or Couturier, if you want more names), but no one think he is on that level (although his 18-19 was on ROR's and Couturier's level). I don't know know why this is such an issue, the stats fit around what most reasonable people would think.

 

[GreatGonzo]

To be honest, I don't know why people even bother arguing with obvious Crosby haters about Crosby. Like nothing you can say will make them change their minds, because their minds are made up because of entirely emotional reasons. In terms of RelTM xGA/60, which is the best defensive analytic we have available right now, Crosby is on par with guys like Kopitar, Toews, Kesler and Zetterberg defensively. If you look at results in terms of RelTM GA/60 from the last page, you get similar results (although I really don't like judging defensive ability based solely on goals against).

All of Crosby's actual on ice results have him as good defensively for 1Cs and on par with other good defensive 1Cs, which is where I think most reasonable people would have him. He's not on Bergeron's level or Datsyuk's level (and I don't think he's on par with ROR or Couturier, if you want more names), but no one think he is on that level (although his 18-19 was on ROR's and Couturier's level). I don't know know why this is such an issue, the stats fit around what most reasonable people would think.

You do realize how contradicting and hypocritical it sounds to call others “Crosby haters” while you go above and beyond to insist that Crosby is better than he actually is, some might call that a “homer.” It goes both ways, the difference is no one is hating Crosby outside of the OP.

enough with your spot picking stats. Crosby isn’t on par with Kopitar, Toews, or Zetterberg. It’s nonsense to continually try it suggest that and say “this is the only REAL defensive stat that matters.” You once again go above and beyond to try to prove tour point that Crosby is good defensively by saying he’s on par with some of the best. He can be good defensively without the constant need to hype it further, but he definitely isn’t on par with many, if any of those guys. What you need to realize is that isn’t a bad thing and no one should hold anything against Crosby for it.

 

[LD29]

He is not half the defense Weber is

 

[shtorm2005]

I doubt Crosby is ever used to shutdown opponent first line. Mostly he's there to outscore them.

 

[Hockey Outsider]

Well, it shouldn’t be a debate. It is raw data, the issue is the poster used on ice data and goals against which shows nothing. Goals are scored randomly.

For the record - you're absolutely right, goals are scored randomly in the NHL. I know that a lot of people don't like to hear that, because it takes a lot of the fun narratives out of the game, but the reality is goals scored (or allowed) are modeled nearly perfectly by a specific discrete probability distribution. (This hasn't ruined the game for me - when I watch, I'm there to watch hockey, not think about probability theory).

I think some people disagree because they think it means that every player is equally good - that's not at all what it means. Of course Ovechkin will score more goals than Milan Lucic - his goal-scoring is modelled by a much higher-frequency probability distribution than Lucic. But the timing, from game to game, is random.

To your point - I think you're suggesting that I should be using more of a corsi based metric (shots for vs shots against) while all these players are on the ice. My thinking was, over ten full years, any randomness should even out. We're not talking about a single game - each of those players (even McDavid) already have several hundred points of data. I agree there's probably more accuracy in using shot numbers but after a decade I don't think we're losing a ton of accuracy, for the sake of looking at actual results - which, ultimately, are what matter.

Maybe Wayne Gretzky is the greatest applied mathematician of all time?

 

[Gurglesons]

For the record - you're absolutely right, goals are scored randomly in the NHL. I know that a lot of people don't like to hear that, because it takes a lot of the fun narratives out of the game, but the reality is goals scored (or allowed) are modeled nearly perfectly by a specific discrete probability distribution. (This hasn't ruined the game for me - when I watch, I'm there to watch hockey, not think about probability theory).

I think some people disagree because they think it means that every player is equally good - that's not at all what it means. Of course Ovechkin will score more goals than Milan Lucic - his goal-scoring is modelled by a much higher-frequency probability distribution than Lucic. But the timing, from game to game, is random.

To your point - I think you're suggesting that I should be using more of a corsi based metric (shots for vs shots against) while all these players are on the ice. My thinking was, over ten full years, any randomness should even out. We're not talking about a single game - each of those players (even McDavid) already have several hundred points of data. I agree there's probably more accuracy in using shot numbers but after a decade I don't think we're losing a ton of accuracy, for the sake of looking at actual results - which, ultimately, are what matter.

Maybe Wayne Gretzky is the greatest applied mathematician of all time?

My point was Midnight Judges picked a specific data point with a specific criteria to prove in mind. His data set proves nothing in terms of predictive defensive acumen it simply shows how many goals happen with Sid on the ice.

Your work was fine and free of bias as you simply applied models to find specific players.

 

[gorangers0525]

He's good, but overrated, defensively.

 

[Hockey Outsider]

In no way would I try to say Ovi is better then Sid Defensively. Two different positions with roles and responsibilities but Im curious why it matters which goalies Crosby had when hes being scored on but it doesn't matter which goalies he crucified in the playoffs in his first 5 years while Ovi played Lundqvist every damn year. Crosby was torching Ottawa and NYI goalies for pts in the playoffs when Ovi played the King. It was obviously uncontrollable for Crosby and it most likely has evened out over the last 8 years or so but which goalies Crosby had backing him up shouldn't matter if everything evens out over their careers. Caps really didn't have a playoffs goalie until Holtby showed up. Do we really want to say Huet, Theodore and the rest of the journeymen and "never were" are better then Fleury was.

The quality of the goaltending a player is playing against is probably a valid factor to consider. I doubt it would have that much of an impact in the regular season (large number of games + reasonably balanced schedule), but it might be meaningful over the span of a specific playoff series.

A legitimate criticism of Crosby is his scoring in the playoffs gets much lower after the first round. Is that a result of playing against tougher teams, or just randomness? If someone wants to do this analysis, I'd look forward to seeing the results.

 

[Hockey Outsider]

It is interesting to compare to team average, but what does the defensive acumen of the 3rd or 4th line players have to do with top line players? Isn't that going to make up roughly a third to half the data? This could probably be more focused by using linemates + most likely D pairings (for example Crosby plays a lot with Letang, etc.).

It might also reflect the nature of the system they are playing in (Pens being geared more to offense), or the quality of depth players (Oilers) who are hardly ever on the ice with star players.

I don't think there's one right answer (that is, do we look at a player's personal GAA, or do we look at how they compare to their team) - that's why I posted both sets of numbers.

I don't think that saying that Kopitar's ES GAA is 2.40, and Crosby's is 2.69, is entirely fair to Crosby, since Kopitar has had better goaltending and (generally) played in front of better defensemen. But I don't think that saying they're both 11% ahead of their team's average is entirely fair to Kopitar - all things being equal, being 11% ahead of a great defensive team is more impressive than being 11% ahead of an average defensive team.

The question was about preventing goals at ES and with a man advantage, in the regular season and in the playoffs, and penalty killing - otherwise known as defense.

Agreed on the penalty killing. Pretty consistently throughout NHL history, top defensive forwards have played on the penalty kill. Crosby doesn't - that's a strike against him.

I'm not so sure about preventing goals on powerplay. The purpose is to score goals there. Shorthanded goals against happen, but I'd argue that's more the defenseman's fault - they're the ones (usually) at the point. Besides, there definitely seems to be some team context there. The top three players in SH goals against over the 2010's are Crosby, Malkin and Kessel with 53, 52 and 51 goals against each. It seems like more than a coincidence that these are the top three forwards on the Pens in PP ice time.

 

[Midnight Judges]

I'm not so sure about preventing goals on powerplay. The purpose is to score goals there. Shorthanded goals against happen, but I'd argue that's more the defenseman's fault - they're the ones (usually) at the point. Besides, there definitely seems to be some team context there. The top three players in SH goals against over the 2010's are Crosby, Malkin and Kessel with 53, 52 and 51 goals against each. It seems like more than a coincidence that these are the top three forwards on the Pens in PP ice time.

I think the nicest thing you could say is Crosby probably doesn't have much defensive responsibility on the PP.

Preventing short handed goals against is an objective for every power play unit. I do not know why some posters regard PP defense as off limits. It's still defense. Preventing goals is unquestionably a good thing, and SH goals count on the scoreboard just like any other.

 

[silkyjohnson50]

Are you trying to ruin daver's holidays?

Lol.

 

[wetcoast]

I think the nicest thing you could say is Crosby probably doesn't have much defensive responsibility on the PP.

Maybe they should have a seperate trophy for best defensive forward on the PP for every season then eh?

Preventing short handed goals against is an objective for every power play unit. I do not know why some posters regard PP defense as off limits. It's still defense. Preventing goals is unquestionably a good thing, and SH goals count on the scoreboard just like any other.

So that's what the PP coaching is diagraming when a team calls a timeout on the penalty.

This metric you have found is just a complete absurd notion, it's almost like saying a guy is a bad penalty killer if he isn't a SH points leader in the league or something.

I just noticed that Ovechkin is 1-0 in Faceoffs this year, can't wait till he goes 2-0 and there is a thread proclaiming his new found skill in faceoffs.

 

[Dingo]

The problem with Crosby is that our media has had a desperate desire to have him as the hands down best player in the league from before he came into it until they found the next Canadian hockey God in McDavid.

First it was that he was a winner, when Ovechkin was clearly better.

Then they werent as hard on partial seasons as usual.

Then, when he was the best player they painted the gap between him and the next best as enormous when it wasnt.

Then the selke votes and defensive talk started up when his production was no longer best.

 

[Tad Mikowsky]

Why bump a three year old thread

 

[PaulD]

This is the end of a decade so it's a good time to look back at defensive statistics over that span. Much has been made of points for this decade, but obviously goals against are also important, especially in terms of evaluating if a player has an elite defensive impact.

This is a comparison of goals against with Sidney Crosby on the ice vs elite defensive players. I threw in Ovechkin for reference:

09/10 through 18/19 Regular Season
	Crosby	Bergeron	Ovechkin	Toews	Kopitar	Datsyuk
EV GA	481	379	548	470	493	240
EV TOI	10717	10251	12218	11276	12350	8438
Total EV GA/60	2.693	2.218	2.691	2.501	2.395	1.707
PPSHGA	53	24	51	31	31	16
PP TOI	2610	1933	3271	2198	2405	1343
Total	1.218	0.745	0.935	0.846	0.773	0.715
EV+PP GA	534	403	599	501	524	256
EV+PP TOI	13327	12184	15489	13474	14755	9781
EV+PPGA/60	2.404	1.985	2.320	2.231	2.131	1.570
PK TOI	295	1417	21	1156	1493	295
PK TOI/GP	27 seconds	1:57	2 seconds	1:35	1:57	41 seconds

[TBODY] [/TBODY]

"PPSHGA" = power play short handed goals against.

And playoffs:

09/10 through 18/19 Playoffs
Player	Crosby	Bergeron	Ovechkin	Toews	Kopitar	Datsyuk
EV GA	98	58	76	70	36	27
EV TOI	1912	1789	1768	1900	1293	931
Total EV GA/60	3.075	1.945	2.579	2.211	1.671	1.740
PPSHGA	8	4	3	2	1	3
PP TOI	483	320	463	326	268	203
Total	0.994	0.750	0.389	0.368	0.224	0.887
EV+PP GA	106	62	79	72	37	30
EV+PP TOI	2395	2109	2231	2226	1561	1134
EV+PPGA/60	2.656	1.764	2.125	1.941	1.422	1.587
PK TOI	24.5 Minutes	225	4:24	197	162	56
PK TOI/GP	13 Seconds	1:54	2 seconds	1:47	2:03	0:57

[TBODY] [/TBODY]

The stat that jumps out here is Pavel Datsyuk's EV GA/60. It is positively miniscule! Easy to understand why he was a Selke favorite. He's a true shutdown forward, and these weren't even his best seasons!

Patrice Bergeron is also a standout here, preventing goals against in all situations and taking heavy penalty killing duties.

Same goes for Kopitar, and especially in the postseason, he has played as a true shutdown C.

Good job.
Datz was a beast

 

[PaulD]

To be honest, I don't know why people even bother arguing with obvious Crosby haters about Crosby. Like nothing you can say will make them change their minds, because their minds are made up because of entirely emotional reasons. In terms of RelTM xGA/60, which is the best defensive analytic we have available right now, Crosby is on par with guys like Kopitar, Toews, Kesler and Zetterberg defensively. If you look at results in terms of RelTM GA/60 from the last page, you get similar results (although I really don't like judging defensive ability based solely on goals against).

All of Crosby's actual on ice results have him as good defensively for 1Cs and on par with other good defensive 1Cs, which is where I think most reasonable people would have him. He's not on Bergeron's level or Datsyuk's level (and I don't think he's on par with ROR or Couturier, if you want more names), but no one think he is on that level (although his 18-19 was on ROR's and Couturier's level). I don't know know why this is such an issue, the stats fit around what most reasonable people would think.

 

[orby]

Haters will spend their entire lives searching for complicated reasons why elite players like Crosby and Ovechkin aren't "as good as everyone says they are." You can cherrypick advanced stats to support a predetermined conclusion all you want...but it's a pointless exercise. Sid is one of the all-time greats with a stellar all-around game and will be remembered as such. I would hazard a guess that anyone who has ever played against him would agree.

 

[Machinehead]

Why bump a three year old thread

Because I needed a list of all the people on here who don't know what "for reference" means.