The Washington Capitals win the Stanley Cup as the 24th SAT%(possession)team in the NHL.

[SotasicA]

With the Washington Capitals 1 win away from the Stanley Cup, they will have won it as the 24th worst SAT%(possession) team in the NHL

So... 8th best?

 

[Butch 19]

nerd alert!

 

[TheDoldrums]

The title of the thread is 24th possession team in the NHL won the Stanley Cup. I am sure the stats geeks are trying to figure out how that happened?

No, not really. Again, "stats geeks" understand probability. They don't overreact to results, they focus on the process. They also understand that 5v5 Corsi is not the only stat in "analytics." Having a dominant PP is evident in the stats. So is getting great goaltending. These are "analytics." It doesn't mean Corsi is useless, it's just one small tool among many others in evaluating a team.

 

[The Winter Soldier]

The Caps have this one player that they see has very valuable. Brooks Oprik. Ask Trotz to defend Orpik and he will trash the stats because boxing an opponent away from a rebound isn't in the stats. If you ask Orpik to defend his game, you will get a similar answer.

Many Caps fans that are heavily invested in analytics thought Orpik should have been scratched all along. Orpik provides things a team needs that don't show up in the stats

Beagle too. All that matters are results. I said in my very first post, I like advanced stats as an additional layer of information.

But I would be remised if I didn't turn around a saying the advanced stats geeks like to use.

Eyes can deceive you when watching hockey.

Well advanced stats can deceive too.

Orpik, Beagle, this whole Caps are examples of this.

It all evens out in the wash.

 

[The Winter Soldier]

No, not really. Again, "stats geeks" understand probability. They don't overreact to results, they focus on the process. They also understand that 5v5 Corsi is not the only stat in "analytics." Having a dominant PP is evident in the stats. So is getting great goaltending. These are "analytics." It doesn't mean Corsi is useless, it's just one small tool among many others in evaluating a team.

Corsi stats(possession) were the genesis of the Analytics movement in hockey when LA won it. Anyone with decent hockey knowledge would know this. It is the primary reason why so many bloggers and young hockey execs were hired by NHL teams right after that. I am sorry if the Caps winning the cup as the 24th SAT% team this year or the Pens with the worst SAT% last year in the playoffs ruins your narrative. But it is what it is. Accept it for what it is. An overrated stat, where again, other measurables such as heart, willingness to block a shot, sacrificing the body, boxing out, as the Caps showed are just as important if not more that the advanced stats miss.

 

[TheDoldrums]

Corsi stats(possession) were the genesis of the Analytics movement in hockey when LA won it. Anyone with decent hockey knowledge would know this. It is the primary reason why so many bloggers and young hockey execs were hired by NHL teams right after that. I am sorry if the Caps winning the cup as the 24th SAT% team this year or the Pens with the worst SAT% last year in the playoffs ruins your narrative. But it is what it is. Accept it for what it is. An overrated stat, where again, other measurables such as heart, willingness to block a shot, sacrificing the body, boxing out, as the Caps showed are just as important if not more that the advanced stats miss.

No actually it's your narrative that Corsi = analytics. I keep saying that it's one aspect among many.

I've never said a good Corsi guarantees success or a bad one guarantees failure. Because no one who understands probability would ever say that. But I do think in general if everything else is equal, it's better to be a good Corsi team than a bad one. I guess I'm weird that way.

 

[The Winter Soldier]

No actually it's your narrative that Corsi = analytics. I keep saying that it's one aspect among many.

I've never said a good Corsi guarantees success or a bad one guarantees failure. Because no one who understands probability would ever say that. But I do think in general if everything else is equal, it's better to be a good Corsi team than a bad one. I guess I'm weird that way.

You will agree possession or a proxy for possession stats that corsi is was the genesis of the advanced stats movement and NHL hires. Atleast be honest about this point.

 

[CupsOverCash]

Well they matter until they dont. This is how I think about it. The name of the game will always be who can put the puck in the net more than the other team.

 

[TheDoldrums]

You will agree possession or a proxy for possession stats that corsi is was the genesis of the advanced stats movement and NHL hires. Atleast be honest about this point.

Okay? On-base percentage was the "genesis" of the advanced stats movement for baseball. But they've progressed beyond that now and no one would say OBP = analytics. On-base percentage is still a useful tool, but it's one of many to be used in a proper evaluation. I would describe Corsi the same way.

 

[The Winter Soldier]

Well they matter until they dont. This is how I think about it. The name of the game will always be who can put the puck in the net more than the other team.

Hockey is a copy cat league. I believe GM's that are not known for stats/spreadsheet numbers crunching will look at how Pittsburg and Washington were able to win the Cup. And this is really all that matters with less than stellar advanced stats. The main one being possession.

I do think Trotz made the most sense when he defended Orpik's numbers. What he saw and what the numbers said, told him teams that rely too much on numbers and not feel are likely going to be deceived on how effective a player truly is.

The bottom line is the Caps are the Champs. And they did not have to be the advanced stats darlings to get there this season. Other GM's would be foolish to ignore this and last year IMO.

 

[The Winter Soldier]

Okay? On-base percentage was the "genesis" of the advanced stats movement for baseball. But they've progressed beyond that now and no one would say OBP = analytics. On-base percentage is still a useful tool, but it's one of many to be used in a proper evaluation. I would describe Corsi the same way.

If this is the case, then you should have no trouble naming me 4-5 advanced stats the Caps did better than other teams this year and playoffs to win the cup. What are they?

 

[TomasHertlsRooster]

Corsi stats(possession) were the genesis of the Analytics movement in hockey when LA won it. Anyone with decent hockey knowledge would know this. It is the primary reason why so many bloggers and young hockey execs were hired by NHL teams right after that. I am sorry if the Caps winning the cup as the 24th SAT% team this year or the Pens with the worst SAT% last year in the playoffs ruins your narrative. But it is what it is. Accept it for what it is. An overrated stat, where again, other measurables such as heart, willingness to block a shot, sacrificing the body, boxing out, as the Caps showed are just as important if not more that the advanced stats miss.

Imagine thinking any of these players don’t have the willingness to block a shot or sacrifice the body

 

[Luigi Lemieux]

Corsi doesn't matter as much now that everyone knows about it. It used to be a stat that correlated with high level play and puck possession. Now if teams are playing to get a high Corsi, it doesn't necessarily translate to high level play. Correlation vs causation.

Corsi is just a proxy for puck possession anyway, it does not measure true puck possession. Caps sure seemed like a great puck possession team to me, but they're not going to take a bunch or crappy shots for the heck of it.

 

[TomasHertlsRooster]

If this is the case, then you should have no trouble naming me 4-5 advanced stats the Caps did better than other teams this year and playoffs to win the cup. What are they?

Marc-Andre Fleury in the Stanley Cup Finals: 116 saves on 136 shots. (.853 SV%)

Braden Holtby in the Stanley Cup Final: 142 saves on 155 shots (.916 SV%)

Reilly Smith, William Karlsson, and Jonathan Marchessault (3 players) in games 2-5 of the Cup Finals: 2 goals and 5 assists

Alex Ovechkin and Evgeny Kuznetsov (2 players) in games 2-5 of the Stanley Cup Finals: 4 goals and 7 assists

Vegas’ entire success was due to one line of three non-star players going on a magical run and playing as one of the NHL’s best lines, as well as one inconsistent and often maligned goaltender going on a magical run and putting up an amazing performance. Those 4 players carried Vegas in the playoffs and people wrote up this “best depth! no superstars!” narrative because it helped them write the narrative they wanted about the expansion draft.

In reality, Vegas was heavily dependent on 4 players for the entirety of the playoffs. All 4 of those players were easily outplayed by their Capitals counterparts in the Cup Finals. You need superstars playing like superstars to win the Stanley Cup. Vegas didn’t have those and it showed in round 4.

 

[TheDoldrums]

If this is the case, then you should have no trouble naming me 4-5 advanced stats the Caps did better than other teams this year and playoffs to win the cup. What are they?

...what? All I said was that analytics is more than just Corsi. I don't understand this response.

 

[MessierII]

...what? All I said was that analytics is more than just Corsi. I don't understand this response.

Find me an analytics guy who still doesn’t think Corsi is the best stat. I’ve been reading about shot attempt stats since 2009. Didn’t like it then don’t like it now.

 

[CantHaveTkachev]

Why are baseball and hockey even compared in analytics?
baseball is a game based on single events (pitcher vs batter) while hockey is a more free-flowing game with systems in place

 

[TheDoldrums]

Find me an analytics guy who still doesn’t think Corsi is the best stat.

All of them? I reject the notion of a "best stat" in hockey right now, but if anything xG (expected goals) would seem to be the leader right now. At least among publicly available ones.

 

[Filthy Dangles]

Ah so now I know why @Machinehead can't stand this team winning lord stanley

 

[MessierII]

All of them? I reject the notion of a "best stat" in hockey right now, but if anything xG (expected goals) would seem to be the leader right now. At least among publicly available ones.

Oh please. Let’s not pretend analytics guys on here and in the blogosphere still don’t look at corsi as a good player/bad player stat. It’s rampant.

 

[Peanut]

All of them? I reject the notion of a "best stat" in hockey right now, but if anything xG (expected goals) would seem to be the leader right now. At least among publicly available ones.

KO

 

[Machinehead]

Ah so now I know why @Machinehead can't stand this team winning lord stanley

I still hate you.

They made a massive turn around the last month of the season. They also were a consistent top possession team in the years leading up to this one.

Their poor start may have been one part fluke, and one part Orpik. Once they acquired Kempny and he took some of Orpik's minutes, we saw a drastic shift in their analytics.

Did they post the best metrics in the playoffs? No, but they were in the upper half and when you have a PP like their's, that's good enough.

 

[Machinehead]

Their adjusted corsi in the playoffs was 49.5%. That's fine when you're playing top possession teams 6, 7 times.

Their xG was excellent.

 

[The Winter Soldier]

Oh please. Let’s not pretend analytics guys on here and in the blogosphere still don’t look at corsi as a good player/bad player stat. It’s rampant.

You are not likely to get an answer from @TheDoldrums because the shining light for analytics ethusiasts IE. corsi/possession stats have taken a hit when Pittsburg won with the worst SAT% in the playoffs and Washington won the Cup as the 24th Sat% team in the NHL. That he has to defend them despite not being able to provide any examples how they predicted correctly of these 2 most recent Cup wins.

 

[Bounces R Way]

Corsi just measures shot attempts at even strength. The hypothesis that this is the best measure of possession and is reliable or predictive in nature regarding winning was perhaps a little underdeveloped. It presumes that all SA are created equal, when that clearly cannot be the case. The margins between winning and losing are so incredibly thin in the NHL. Yes I would agree with the statement that consistently outshooting your opponent is conducive to winning, but judging who had the run of play and were the better team based on a difference of <10 shot attempts I find to be ridiculous.