Curious Statistical Case of Carolina Hurricanes

[talitintti]

Advanced stats are a good farmhand but a bad master. They are useful in the hands of personnel actually understanding the nuances of the game of hockey, but they only lead to disaster in the hands of people who don't understand the game and only understand the spreadsheet of stats.

The history of advanced stats advocating GM's and coaches is rather disastrous in the NHL.

 

[The Macho King]

It's not a shot quality issue. Most likely it's small sample size. Even over 80 games there are teams every year that have really high/low sh% and they almost always come back to league average sh% in subsequent years.

Heatmaps are stupid. Not every shot from the slot is a "good shot". Was it right along the ice? Was it a logo seeker? How close were they to the goalie? Did they actually have anything to shoot at?

 

[spintheblackcircle]

I heard a stat that the Canes have been 25th or worse in shooting percentage for something like 8 years in a row.

Full roster turnover, 3 coaching changes.....this never changes.

 

[Man Bear Pig]

And Caps are the opposite every season. Outperform metrics by a lot

I think both cases are due to talent. Caps have exceptional finishing talent while the Canes are severely lacking

I agree. I'll admit I only see a handful of Canes games each year but from what I've seen, they very clearly lack the firepower up front. You can kill it in possession numbers and shots on goal but if theres a lack of quality chances, it doesn't mean much.

 

[Stickpucker]

No finish.

We have more Finnish than most teams.

 

[Maukkis]

Heatmaps are stupid. Not every shot from the slot is a "good shot". Was it right along the ice? Was it a logo seeker? How close were they to the goalie? Did they actually have anything to shoot at?

Yet it has been established that the slot is the area from which goals are scored at the highest likelihood. The sample size behind this is way beyond fifty thousand shots, if my memory serves me, so that alone combats your point quite well.

I've been wondering about this too. The closest to an answer I have come is that the Canes are lacking in high end talent. I've seen arguments for shooting talent being a rather minuscule factor in having success, but the complete lack of finishers on that roster makes me think that it could be an important factor here.

The best shooters can comfortably outperform their xG numbers over a large sample size. It seems like absolutely no one there is committed to even being break even. Fascinating, and it also opens up the debate of whether it is in the system or not.

Also, this is why you don't trade your Skinners for nothing.

 

[gumgum]

Probably the simplest explanation would be that they lack the finishing ability of most other teams, which is it.

 

[Connor McConnor]

It's pretty simple to me. They lack super star level talent at the forward position. None of their current players are "game breakers" in my opinion. Maybe Svech can change this but right now they have nobody that can go out and be a consistent offensive threat, night in and night out. If you look at Skinner/Lindholm career years I think that is moreso about who they are now playing with.

Skinner goes to Buffalo and plays with a true superstar level talent in Eichel. Skinner has always been a consistent goal scorer but this season he has looked like a Rocket contender with Eichel.
Lindholm goes to Calgary and plays with a true superstar level talent in Gaudreau. Doesn't hurt to add a very skilled centre in Monahan to that mix either but Gaudreau is the key here. He basically makes Monahan a 30g/70+ pt guy every season, without him he probably looks like a 60 point guy (still very good).

I understand how good Aho is but he can't carry the load by himself. He is simply not as good as Gaudreau is (atleast not yet) or Eichel. He's not a goal scorer either so he needs teammates to play with.

Turns out you need superstars to win in the NHL. Just like almost every other sport.

 

[VoluntaryDom]

Actually, any "stat" or "number" involved with or named "FENWICK", will not get a positive response from me.

Like I've said, many of the "terms", "numbers", and/or "stats" involved with advanced statistics, are made up.

Ok, but xGF is not a hypothetical. It’s a number based on actual on ice events.

 

[bambamcam4ever]

Heatmaps are stupid. Not every shot from the slot is a "good shot". Was it right along the ice? Was it a logo seeker? How close were they to the goalie? Did they actually have anything to shoot at?

Yup. There is so much more involved in whether it is a "good" shot that merely location. Is there a defender all over the shooter? Is the shooter skating towards the net? Is the puck settled? This context is necessary and so many of the advanced stats proponents on HF seem to be frightened by additional context.

 

[bambamcam4ever]

Yet it has been established that the slot is the area from which goals are scored at the highest likelihood. The sample size behind this is way beyond fifty thousand shots, if my memory serves me, so that alone combats your point quite well.

I've been wondering about this too. The closest to an answer I have come is that the Canes are lacking in high end talent. I've seen arguments for shooting talent being a rather minuscule factor in having success, but the complete lack of finishers on that roster makes me think that it could be an important factor here.

The best shooters can comfortably outperform their xG numbers over a large sample size. It seems like absolutely no one there is committed to even being break even. Fascinating, and it also opens up the debate of whether it is in the system or not.

Also, this is why you don't trade your Skinners for nothing.

Just because it is true for the larger group does not mean it is so for the individual.

 

[pabst blue ribbon]

They have historically bad shooting, you'll score a lot more goals with good chance generation and good finishing as opposed to great chance generation and god awful finishing

 

[FerklundCGY]

As a team, is much, much better to be analytically great than not but just because you're team is great in most analytical categories doesn't necessarily mean you're gonna be a great team.

You still need talent.

This is one thing people who hate analytics never seem to understand. They see these analytically great players and/or teams and think that just because the results don't match the analytics that analytics have no value, which is just absolutely false.

Of course, it's much better and more valuable to your team if you're constantly out-shooting and out-chancing your opponents, but if you don't have the talent to finish your chances, then your results aren't going to match your xGF.

This is why you always see incredible shooters like Laine, Ovechkin, Monahan, etc... outperforming their ixGF and also while you always see incredible skaters like McDavid, Byron, Athanasiou, etc... outperforming their ixGF.

This is why I always get into an argument on the Calgary board with someone like Bennett, who's ixGF and ixGF/60 is usually tops on the team, yet because his production never matches his expected output, Flames fans automatically dismiss something like ixGF as a stupid analytical stat when in reality, it's because that while Bennett does create a fantastic number of scoring chances and high danger scoring chances for himself, he just doesn't have the finishing ability.

So, in short: it's great that the Hurricanes corsi, scoring chance rates & xGF analytical stats are among the very best in the NHL but they just don't have the talent on their roster to actually produce those results.

A team with 18 McDavid's and analytical stats like the Hurricanes would completely dominate the NHL and score an infinite amount of goals, whereas a team with 18 McDavid's and analytical stats like the Senators would still dominate the NHL due to the talent of the team, but would struggle mightily against the McDavid's with Hurricanes analytics if put head to head in a Best of 7.

Hope that helps.

 

[thebus88]

Yup. There is so much more involved in whether it is a "good" shot that merely location. Is there a defender all over the shooter? Is the shooter skating towards the net? Is the puck settled? This context is necessary and so many of the advanced stats proponents on HF seem to be frightened by additional context.

So many variables these "models" CAN NOT comprehend. Not everything can be given a number value.

Can you tell me the difference between a scoring chance and a "high danger" scoring chance...??

Yeah, me neither.

 

[TomasHertlsRooster]

It's not a shot quality issue. Most likely it's small sample size. Even over 80 games there are teams every year that have really high/low sh% and they almost always come back to league average sh% in subsequent years.

Shot quality can not be determined by simply location. There are other factors to consider.

Carolina probably does have bad shooters and bad goaltending, I won’t dispute that. But I reckon if you gave them exactly league average shooting and goaltending, their GF% would under-perform their xGF% by a significant margin.

 

[talitintti]

The most, most annoying part of the hockey fandom are the stats nerds, who act that they actually understand the game after looking at the stats.

It's embarassing, because the stats guys are among the 98% of fans that don't have any deeper understanding of what's going on, and are usually in the lower percentile of even that group.

 

[TomasHertlsRooster]

As a team, is much, much better to be analytically great than not but just because you're team is great in most analytical categories doesn't necessarily mean you're gonna be a great team.

You still need talent.

This is one thing people who hate analytics never seem to understand. They see these analytically great players and/or teams and think that just because the results don't match the analytics that analytics have no value, which is just absolutely false.

Of course, it's much better and more valuable to your team if you're constantly out-shooting and out-chancing your opponents, but if you don't have the talent to finish your chances, then your results aren't going to match your xGF.

This is why you always see incredible shooters like Laine, Ovechkin, Monahan, etc... outperforming their ixGF and also while you always see incredible skaters like McDavid, Byron, Athanasiou, etc... outperforming their ixGF.

This is why I always get into an argument on the Calgary board with someone like Bennett, who's ixGF and ixGF/60 is usually tops on the team, yet because his production never matches his expected output, Flames fans automatically dismiss something like ixGF as a stupid analytical stat when in reality, it's because that while Bennett does create a fantastic number of scoring chances and high danger scoring chances for himself, he just doesn't have the finishing ability.

So, in short: it's great that the Hurricanes corsi, scoring chance rates & xGF analytical stats are among the very best in the NHL but they just don't have the talent on their roster to actually produce those results.

A team with 18 McDavid's and analytical stats like the Hurricanes would completely dominate the NHL and score an infinite amount of goals, whereas a team with 18 McDavid's and analytical stats like the Senators would still dominate the NHL due to the talent of the team, but would struggle mightily against the McDavid's with Hurricanes analytics if put head to head in a Best of 7.

Hope that helps.

No, this isn’t entirely it. xGF% isn’t a stupid stat, but there is only one publicly available xGF% model and it’s not all that great; same thing with WAR and they’re both made by the same guy. Sean Tierney’s xGF%, available at Corsica.hockey really isn’t all that great.

As a Sharks fan, I can tell you this because I’ve actually been watching a pretty similar team this season. The GF% is only a little better than Carolina and their CF% and xGF% is only a little worse.

They do legitimately have poor goaltending, but I don’t really think their finishers are all that bad; they have 2 of the top-8 5V5 goal scorers in the NHL, another guy ranked 37th, and I think their finishing ability is roughly average as a whole. Their PP is ranked 9th and their PK is ranked 3rd, and their combined PP%+PK% is 107.8. That’s top-5 in the NHL and heavily suggests that they have the talent to have a PDO of at least 100, yet they are nowhere near that.

Here are a couple posts that I made that describe how the system leads to more GA than xGA:

I should preface this by saying that I am not a hockey systems expert, and that somebody else could make a better analysis on what exactly has changed. The only reason that I am so adamant about this system change is that they said they implemented it after the loss in Vegas, they clearly did implement some sort of change, and the numbers CLEARLY state that it was a bad decision.

However, from what I see:

-In the offensive zone, when the opposition has the puck and is attempting to exit the zone, the defensemen are very inclined to pinch at the blue line in order to keep the puck in the zone. This works sometimes, but when it doesn’t, it usually leads to odd-man rushes against, because the forwards usually don’t get back in the defensive zone to cover the odd-man in time. This partially explains why the Sharks give up a low to medium amount of shots against, but a high number of high danger chances and goals against.

-This works much better for a team like Vegas, partially because their forwards and defense are on the same page, and partially because their forwards are very fast and defensively responsible. That same system is not optimal for a team whose top forwards are Meier, Pavelski, Thornton, Hertl, and Couture.

-On top of that, the team doesn’t currently appear to have the level of commitment to defense necessary to fit that Vegas system. Vegas’ forwards made it a point to get back and avoid odd-man rushes. The committment to team defense isn’t there night in and night out and given the personnel they do have, even if they hunkered down and tried to do so, it likely wouldn’t be sustainable for the next 56 games and the playoffs.

-As @Lebanezer has mentioned in the past, the system, when executed properly, is very tiring for the players. They put in a ridiculous amount of effort to dominate possession generate a few measly point shots. This tires them out - perhaps because they don’t have the stamina of the Vegas personnel, and perhaps because they aren’t executing their dominance properly - and leaves the team very fatigued, both shift to shift, and period to period. This explains why their 1st periods are so strong and their 2nd periods are so poor. They expend a ton of energy to dominate possession and generate a high volume of mediocre scoring chances, score on a small amount of them, and then bleed goals when they are tired in the 2nd period and the opposition can pounce.

-The “cycle to the point, shoot it into the defender’s knees” thing that has become a meme lately was present as far back as 2016-2017, and so it isn’t directly tied to this system change, but it does appear to have been scouted and to have grown gradually worse.

I would love to hear what some other posters have to say.

This is, in my view, the biggest issue with the system. The pinching from the defense, when the opposition is breaking out. It’s similar to Vegas’ system, but our guys don’t execute it as well. This one turned out fine but it’s very easy to see why this frequently leads to odd man rushes against.

Here is a post that another HFSharks poster named Lebanezer made that describes why their xGF is so much higher than their actual GF.

Sharks take low percentage shots almost exclusively. They don’t get goalies moving side to side inside the zone, they funnel pucks to the points for low percentage shots and tip plays. Their high danger chances are off rebounds, not triangle plays and give and goes. That’s why when they get high danger chances they don’t have a high shooting percentage, a rebound shot is almost always contested. Tips are almost always contested. Hertl and Timo generate the best high danger chances because they make east west passes in the zone and drive the net. Hertl is the only guy who doesn’t pass the puck up to the point from behind the net. He curls out, uses his size and drives the net. Their system is moronic, particularly in that it’s so easy to defend. All other teams do is keep the Sharks to the outside, let them pass to the point and block the shots or retrieve the massive rebounds off the backboards, which feeds their transition. Deboer wants the players to play inside, but they can’t inside because they’re all on the outside cycling the puck back to the point. Then when they need to get to the net they’ve ceded inside position to the defender and can’t fight through to the front. The whole system is moronic and slow. They can’t play with speed because everything is bogged down by the system. Burns is often trying to get a shot through 5 people. Everything they do is low percentage and it’s systematic. It needs to be changed.

Over the past 3 full seasons, the Washington Capitals had a 5V5 GF% of like 56.5% or something ridiculous and every other team was like 3% below them, yet their xGF% was under 50%. Their PDO was 1.019 and the next best was 1.011. Everybody with a small understanding of the stats just looks at that and screams “Ovechkin!”, but even if you took out every 5V5 goal Ovechkin scored over that time frame, they still rank 3rd or 4th in 5V5 GF% and if you also took every goal from every other team’s top 5V5 goal scorer away, Washington’s 5V5 GF% blows them all out of the water.

Then, Barry Trotz leaves Washington and joins the New York Islanders. All of a sudden, despite losing their #1 center who actually does have the talent to drive PDO above 1.00, the Islanders now have a PDO if 1.022, good for 3rd best in the NHL after 30 games. They were at 1.005 last year.

TL;DR The Islanders lost John Tavares, one of the NHL’s elite finishers, and their PDO jumped by 0.017. This is because individual talent does not entirely determine PDO and GF:xGF ratio. Coaching and systems have a very large effect and maybe even a bigger effect. The Sharks are actually pretty similar to Carolina in GF% and xGF%, I watch enough of the Sharks to tell you why that is for them; it’s not due to simply finishing/goaltending and it’s probably similar for Carolina.

 

[winnipegger]

It does seem like the same story over and over. The Canes play really good 2 way hockey and can't score enough. Do you stay the course or make massive changes, who knows. The thing is, they have some assets they can move.

 

[traparatus]

As a team, is much, much better to be analytically great than not but just because you're team is great in most analytical categories doesn't necessarily mean you're gonna be a great team.

You still need talent.

This is one thing people who hate analytics never seem to understand. They see these analytically great players and/or teams and think that just because the results don't match the analytics that analytics have no value, which is just absolutely false.

Of course, it's much better and more valuable to your team if you're constantly out-shooting and out-chancing your opponents, but if you don't have the talent to finish your chances, then your results aren't going to match your xGF.

This is why you always see incredible shooters like Laine, Ovechkin, Monahan, etc... outperforming their ixGF and also while you always see incredible skaters like McDavid, Byron, Athanasiou, etc... outperforming their ixGF.

This is why I always get into an argument on the Calgary board with someone like Bennett, who's ixGF and ixGF/60 is usually tops on the team, yet because his production never matches his expected output, Flames fans automatically dismiss something like ixGF as a stupid analytical stat when in reality, it's because that while Bennett does create a fantastic number of scoring chances and high danger scoring chances for himself, he just doesn't have the finishing ability.

So, in short: it's great that the Hurricanes corsi, scoring chance rates & xGF analytical stats are among the very best in the NHL but they just don't have the talent on their roster to actually produce those results.

A team with 18 McDavid's and analytical stats like the Hurricanes would completely dominate the NHL and score an infinite amount of goals, whereas a team with 18 McDavid's and analytical stats like the Senators would still dominate the NHL due to the talent of the team, but would struggle mightily against the McDavid's with Hurricanes analytics if put head to head in a Best of 7.

Hope that helps.

This is a very likely scenario. The only alternative I see is in (lack of) pre-shot movement. None of the publicly available models account for it (as far as I know) while custom data models indicate that it's a major driver of goal production.

We need one of the eggheads with a subscription to a data store to have a lookie at what is happening just before all those slot shots are taken. That's A LOT of slot shots.

I find the notion of always practicing against awful gaoltending having an impact on shooting percentage VERY intriguing.

 

[maacoshark]

I'm sick of these advanced stats guys. They are inaccurate.

 

[Tobias Kahun]

Maybe if they stopped doing those childish group celebrations after the game, and practiced, they would score more goals.

 

[Makar Goes Fast]

yeah but its not fair that their fans get so many likes soooooo

 

[Finlandia WOAT]

They don't pass into or through the slot thereby forcing the goalie to move laterally and opening up the net before he can get set again. On the heat map all those shots are rebounds or tips. When they do, they don't have the ability to consistently bury it.

They focus on rebounds because their lack of individual skill means no forward can consistently beat a guy 1v1, forcing a defensive breakdown which, if it did happen, would be flubbed because they lack passing ability and vision to capitalize.

Basically their forwards blow.

 

[Mickey Marner]

I'm not sure how to phrase this but ill try. Is the reason their CF is so high because they can't put the puck in the net? If they could just finish the play off, those extra CF would cease and that number would (*might) not be so high. I mean you still have to be a good team to pounce on those rebounds and loose pucks and generate a subsequent shot, I guess.

But how much of their CF is due to them not finishing an offensive play or sequence off where they should have? Is there some formal name for such a phenomenon in statisitics?

Hope that makes some sense.

Yes, this infates their CF to some degree. Over the year as I've noticed this both on the team and individual level. A goal against begets no further corsi, but a shot attempt that doesn't go in might. I've never seen any attempt to study it, however. Or initial shot save/shooting percentage.

 

[cptjeff]

This is a very likely scenario. The only alternative I see is in (lack of) pre-shot movement. None of the publicly available models account for it (as far as I know) while custom data models indicate that it's a major driver of goal production.

We need one of the eggheads with a subscription to a data store to have a lookie at what is happening just before all those slot shots are taken. That's A LOT of slot shots.

I find the notion of always practicing against awful gaoltending having an impact on shooting percentage VERY intriguing.

No advanced stats measure that because it's an insanely difficult thing to quantify at scale. The more complex a system is, the harder it is to quantify. In hockey, every single statistical event is opposed, and almost always by more than one actor, and supported by more than one actor. Unlike something like baseball where the core of the game is 1 on 1- batter versus pitcher- everything in hockey but penalty shots is 6 versus 6, with the puck moving in a different but linked manner from any of the players, even when it's on a players stick- when a player is handling the puck, it's taking a different path on the ice than the player because the player handles the stick separately from the skates.

In an equation, to get a number, you can only have one unsolved variable. So you either have to get data for every single variable, or you have to eliminate variables that you don't think are relevant. The former is impossible, so advanced stat makers choose the latter, but here lies the core of the problem. They decide what variables to drop based on one of three elements: One, they understand that an element has a slight effect or none and decide that they can safely ignore it. This is perfectly valid. Two, they do not understand the game well enough to know that they should be tracking a variable. Motion before a shot falls into this category. For a long time, stat watchers argued that SV% was the perfect goalie stat and would refuse to entertain the notion that shot quality varied systemically. Now they're measuring shot distance and equating it to shot quality, unaware that there are other elements to making a quality shot that might have to be accounted for. Third, information is ignored because it's not easily quantified or quantifiable at all, and then stat proponents will claim that the factor simply does not exist because they cannot account for it numerically. The prime example of this is from baseball, where for decades stat watchers claimed that framing pitches did not exist, and that old school baseball managers were insane for paying extra for catchers who could frame pitches. Then a few years ago, they figured out how to quantify the strike rate for catchers and discovered that, guess what, framing pitches not only worked, but had significant outcomes on games and seasons- and that the decisions those antique GMs they had mocked pretty much aligned exactly with the value of catchers reflected in their new stat.

Hockey is one of the most enormously complex systems you'll find anywhere. To quantify it, you have to focus on very narrow sets of data, and you are going to lose a massive amount of information by focusing only on the few things you can actually track.

Stop watching spreadsheets and watch the games. Lots of crappy shots don't create wins. The Hurricanes take almost entirely crappy shots.