Why are people so obsessed with so called advanced statistics?

[TT1]

because I'd prefer not to watch hockey games

 

[DanZ]

This is hilarious and demonstrates the certain group pushing advanced stats. "Things like low quality shots even out across teams over a large sample size.... except when they don't even out for certain other things because I said so"!

It's not something that needs to be measured. Measuring which team is getting more shots to the goal is far more representative of which team is actually controlling the game. You seem like the type that always thinks they know what's going on but the fact remains that zone time is pointless to measure when there are better measures out there. No one will use because Corsi is better at telling us who is controlling the game.

 

[lomiller1]

You don't get it do you? I don't want to model the game, I want to analyze it..

I have no idea what you are trying to say.

I want to analyze it, to see actionable information for evaluation

This is called building a statistical model

 

[Sojourn]

It's not something that needs to be measured. Measuring which team is getting more shots to the goal is far more representative of which team is actually controlling the game. You seem like the type that always thinks they know what's going on but the fact remains that zone time is pointless to measure when there are better measures out there. No one will use because Corsi is better at telling us who is controlling the game.

With all due respect, do you understand how hilarious it is for you to say this in the context of this post?

You just did exactly what you accused him off, while dismissing a statistic that would measure how often each team has the puck, and where they have it. You can't tell me that wouldn't be useful information. No matter how much you like corsi, that's useful information. Whether you like corsi more or not, or whether corsi would be better at certain things, you're dismissing a statistic that you wouldn't have had an opportunity to really use and judge.

 

[TT1]

Hockey isnt baseball (where you can isolate variables and have very precise stats), hockey is a team sport that has A LOT of variables. Player stats vary depending on their own personal style, their team system, their linemates, their usage (defensive vs offensive players), the competition they face etc. etc. etc.

 

[aufheben]

I have no idea what you are trying to say.


This is called building a statistical model

Why would following stats disrupt analyzing it lol. I started paying wayyy more attention one I understood "advanced" stats. Especially when you realize that players can look good without actually being a net positive for their team.

 

[lomiller1]

Because for a player - a long list of things will influence those same numbers/variables and those variables also interact with each other.

The list really isn't all that long and the items on it are fairly well quantified at this point. The things of note, in decreasing order of importance, are Player Skill, Quality of Teammate, Zone starts, Quality of Competition.

Are you perhaps thinking that because there are a long list of things that can fall under the category of player skill that this is an issue? If so, it's not an issue because we don't really care how a player gets the job done we only care about there success/failure.

 

[lomiller1]

Why would following stats disrupt analyzing it lol. I started paying wayyy more attention one I understood "advanced" stats. Especially when you realize that players can look good without actually being a net positive for their team.

I'm not sure you understood what I was saying. The post I was responding to said they wanted to analyze the stats without building models, but even something a simple as points per game is a model albeit a very simple one. If you don't build models all you are doing is staring at raw numbers hoping for an epiphany.

 

[aufheben]

I'm not sure you understood what I was saying. The post I was responding to said they wanted to analyze the stats without building models, but even something a simple as points per game is a model albeit a very simple one. If you don't build models all you are doing is staring at raw numbers hoping for an epiphany.

Oh I misread it as when some people say knowing the "science" behind it makes it a less magical experience or w/e. Also applies to when people say "watch the games" but I find games more interesting to watch this way, like looking everything up during intermissions or after the game.

 

[Machinehead]

2011-12 LA Kings

14th in goal differential
29th in goals scored
17th in WINS(!!)
28th in PDO
18th on the powerplay
2nd in corsi

Playoff record: 16-4

A team that was good at one thing had one of the most dominant playoffs of all-time. There's a pretty good reason.

 

[Bjornar Moxnes]

2011-12 LA Kings

14th in goal differential
29th in goals scored
17th in WINS(!!)
28th in PDO
18th on the powerplay
2nd in corsi

Playoff record: 16-4

A team that was good at one thing had one of the most dominant playoffs of all-time. There's a pretty good reason.

Quick being an alien helped.

 

[TT1]

2011-12 LA Kings

14th in goal differential
29th in goals scored
17th in WINS(!!)
28th in PDO
18th on the powerplay
2nd in corsi

Playoff record: 16-4

A team that was good at one thing had one of the most dominant playoffs of all-time. There's a pretty good reason.

L.A is a top Corsi team every year, they've had years where they missed the playoffs aswell..
In 11-12 Quick had a .946 save %, in 13-14 they had 5 of the top 6 scorers in the playoffs (everything was clicking for their offense)

There's plenty of examples of teams winning the cup with weak corsi aswell.

 

[Passchendaele]

They annoy me beyond reason.

Ten years ago, people on here didn't talk about Corsi, Fenwick, offensive zone starts/etc. Those were the good old days.

 

[fsanford]

In O.T. the Kings corsi seems to pay big dividends.

 

[fiveonfive]

Because for a player - a long list of things will influence those same numbers/variables and those variables also interact with each other. CORSI and other stuff works OK for teams because you don't care who is influencing who/what and there isn't much change in rosters year over year. The predictive power of CORSI is strong because of that low roster turnover.

The macro model of hockey statistics is a hell of a lot more broad and stable so it can be 'simplified'.

You have a little bit of cognitive dissonance in your points. You agree that having a greater share of shot attempts works, but then you close your eyes to the fact that we have some pretty good ways to estimate which players can drive those shot attempt metrics then others by saying "well, there is so many variables". Can you show that there are any variables that would be material enough to distort the use of these statistics and their use?

My guess is that you can't, because you aren't actually familiar with any of the reasoning behind the value of these statistics.

This is hilarious and demonstrates the certain group pushing advanced stats. "Things like low quality shots even out across teams over a large sample size.... except when they don't even out for certain other things because I said so"!

Just because one stat isn't materially distorted from noise or misinterpreted from an anecdotal situation, does not mean another one is. That poster was attempting to reconcile the reason why "possession time is not as good as corsi", not provide you with a methodology of how to deal with stats or to say that only my stat works. We already know that corsi is better than time of possession. If you think this anecdote demonstrates anything about people pushing advanced stats, you are misguided and should pay more attention to the work that has been done on the topic.

With all due respect, do you understand how hilarious it is for you to say this in the context of this post?

You just did exactly what you accused him off, while dismissing a statistic that would measure how often each team has the puck, and where they have it. You can't tell me that wouldn't be useful information. No matter how much you like corsi, that's useful information. Whether you like corsi more or not, or whether corsi would be better at certain things, you're dismissing a statistic that you wouldn't have had an opportunity to really use and judge.

You are taking that post out of context. There have been multiple posters ( clearly anti advanced stat people) posting that their problem with advanced stats is that they are lazy for not tracking the real time of possession, aren't actually good proxies for time of possession, and that they don't show what they claim to show, and that therefore those stats are "bad".

However, the reality is that "possession stats" are just a name for these shot attempt metrics - a colloquial name - stemming from the fact that they were named that they track possessions that result in shot attempts, where shot attempts and prevention of opponent shot attempts is an indicator of effective play. Coris is NOT time of possession and does not seek to estimate time of possession. Corsi is also demonstrably more useful then time of possession.

That poster was dismissing the need to track the time of possession as a response to "actually this is how you should do it but you guys are dumb and are doing it wrong" with "no, we are tracking the more important thing". If you can get accurate time of puck possession statistics and derive useful indicators from them, go ahead and do it, no one is against that.

 

[Mass]

Quick being an alien helped.

Corsi creates insanely high save percentages.

 

[fiveonfive]

L.A is a top Corsi team every year, they've had years where they missed the playoffs aswell..
In 11-12 Quick had a .946 save %, in 13-14 they had 5 of the top 6 scorers in the playoffs (everything was clicking for their offense)

There's plenty of examples of teams winning the cup with weak corsi aswell.

Um, who are all these teams with weak corsi you speak of? Please provide their score adjusted corsi for their second half of their Stanley Cup wining regular season.

(hint: there actually aren't any that exist)

 

[Kakko]

L.A is a top Corsi team every year, they've had years where they missed the playoffs aswell..
In 11-12 Quick had a .946 save %, in 13-14 they had 5 of the top 6 scorers in the playoffs (everything was clicking for their offense)

There's plenty of examples of teams winning the cup with weak corsi aswell.

but... there aren't

 

[TT1]

Corsi creates insanely high save percentages.

Tell that to Thomas when he faced Vancouver in the cup finals (Boston got massacred advanced stats wise). Look at Jonathan Quick's save % in the regular season, L.A is always a top 2 corsi team.

 

[fiveonfive]

People are obsessed with the stats usually because they have a vested interest in them... like a blog, website, or twitter handle which spams the use of them ad nauseum.

One of the most hilarious "advanced" stats to come out in recent time? I believe they're calling them HDshots/HDs/HDsh or something like that....

Basically if a shot is taken from this arbitrarily partitioned area on the ice...voila... it is THIS (insert number) "dangerous".

I've seen better analysis in a grade 12 stats class.

Grade 12? Grade 6 have better analysis lol... Also yeah I absolutely agree. Danger shots are extremely skewed and dumb analysis as they're about distance not true danger. The only barring they have is that the exact same shot at the exact same speed is obviously more dangerous in the slot than the point, but the chances of those happening is almost nil. Advanced stats would have you believe that St. Louis is not even mid tier, when in reality when you watch their games they only allow a lot of HD shots ebcause they like to sit back on leads and allow the goalie to see a slow visible slot shot, aka not nearly as dangerous as a heavily screened point shot that is deflected. Oh on the opposite spectrum advanced stats would have you believe that Nashville is the best team in the league and no one comes close and that Toronto is actually a top 10 defensive team , whereas in reality the only reason they give up little slot shots is because majority of their point shots are heavily screened and deflected, or for Nashville's case Rinne tracks and freezes the puck or allows the initial shot before there can be secondary chances.

Talk about a bunch of misinformation. These two posts are the equivalent of saying " the eye test people believe that Buffalo has been the best team in the league! These people are so dumb!". Just silly stuff.

But then people wonder why those who pay attention to advanced stats aren't nice to them after they make up a bunch of fantasies and call them stupid in the first place.

 

[Machinehead]

L.A is a top Corsi team every year, they've had years where they missed the playoffs aswell..
In 11-12 Quick had a .946 save %, in 13-14 they had 5 of the top 6 scorers in the playoffs (everything was clicking for their offense)

There's plenty of examples of teams winning the cup with weak corsi aswell.

Plenty???

I'll wait.

 

[fiveonfive]

Tell that to Thomas when he faced Vancouver in the cup finals (Boston got massacred advanced stats wise). Look at Jonathan Quick's save % in the regular season, L.A is always a top 2 corsi team.

Can you quote what the advanced stats were for that series for us? Score adjusted 5v5 CF% would do. ( I don't know them my self, but I bet "massacred" is as a accurate of a description as saying a 2:1 win is a blowout)

LA is always a top 2 corsi team and they happened to have 2 Stanley cups in recent years. How you can rationalize that into corsi=wrong is beyond me, but I would love to hear your explanation.

 

[TT1]

Plenty???

I'll wait.

In recent years..

Kings had the best Corsi in the 1st round last year (55.72%) and lost 1-4 (to SJ), SJ made the finals by being the 12th best Corsi team (47.87%) and Pittsburgh was 8th (51.69%). Last year the top 4 corsi teams lost in the 1st round (L.A, Florida, NYR, Chicago).

2 years ago Chicago made it to the finals by being 6th in Corsi (51.16%) and Tampa was 11th (48.89%). STL and Vancouver were the top 2 Corsi teams in the 1st round (57.12% for STL and 55.17% for Van) and they both lost to the bottom 2 Corsi teams (Minny had 46.37% and Calgary had 44.80%).

Corsi is a useful stat but it's just 1 of the many stats that you have to factor in when trying to evaluate a player/team. There are teams that have weak Corsi based on their system. They give up perimeter shots and prioritize slot coverage/high danger areas way more, thats just them prioritizing quality of shots over quantity of shots. There are certain dmen that willingly play that style aswell. Again, hockey is a team sport.. it has A LOT of variables, context is everything.

Can you quote what the advanced stats were for that series for us? Score adjusted 5v5 CF% would do. ( I don't know them my self, but I bet "massacred" is as a accurate of a description as saying a 2:1 win is a blowout)

LA is always a top 2 corsi team and they happened to have 2 Stanley cups in recent years. How you can rationalize that into corsi=wrong is beyond me, but I would love to hear your explanation.

I said why in my post.. goaltending (11-12) and over the top production (13-14). It's as simple as that.

 

[Penske]

People have always done this with any stats. I'm not a professional scout so I don't value my "eye test" as that high (although it seems a lot here do).

I watch essentially every Rangers game and I still follow their "advanced/enhanced" stats. If I was that good at judging players solely by watching them I'd have a job in hockey.

The eye test can be flawed if you are only watching the game on TV though. You won't see everything.

 

[fiveonfive]

In recent years..

Kings had the best Corsi in the 1st round last year (55.72%) and lost 1-4 (to SJ), SJ made the finals by being the 12th best Corsi team (47.87%) and Pittsburgh was 8th (51.69%). Last year the top 4 corsi teams lost in the 1st round (L.A, Florida, NYR, Chicago).

2 years ago Chicago made it to the finals by being 6th in Corsi (51.16%) and Tampa was 11th (48.89%). STL and Vancouver were the top 2 Corsi teams in the 1st round (57.12% for STL and 55.17% for Van) and they both lost to the bottom 2 Corsi teams (Minny had 46.37% and Calgary had 44.80%).

Corsi is a useful stat but it's just 1 of the many stats that you have to factor in when trying to evaluate a player/team. There are teams that have weak Corsi based on their system. They give up perimeter shots and prioritize slot coverage/high danger areas way more, thats just them prioritizing quality of shots over quantity of shots. There are certain dman that willingly play that style aswell. Again, hockey is a team sport.. it has A LOT of variables, context is everything.

Oh, this explains a lot. You don't understand how corsi works or how to use it. It's clear because you aren't making any sense.It's the team's corsi during the entire regular season that matters (actually the last 25 games of the regular season is even better), not their corsi during the series against the team they play. Also, score adjustments is a big thing for a lot of the individual mashups you have posted, because it's a known and measured effect, and it dictates that teams that are down during a game will start taking the greater share of shot attempts, and the greater the lead of their opponent the greater is the effect. So for a lot of those, the losing teams spent a lot of time "chasing games" theresfore their raw corsi during the series looks inflated - because it is due to score effects.

Pittsbugh was number 2 in Corsi last year, not number 8. SJ was number 9 not #12. Notice other theams like Dallas, Nashviellle, and Tampa, were all top 10 as well and made it deep in the playoffs. Florida was actually 20th, NYR were 26th and Chicago were 15th. Not sure why you thought they were top 4.See link below (need to sort it for CF at 5v5)
https://stats.hockeyanalysis.com/teamstats.php?disp=1&db=201516&sit=5v5&sort=CFPCT&sortdir=DESC

Now the year before, Chicago was number 2 and Tampa Bay was number 4, not 6th and 11th like you posted.Vancouver was 19th - not a top team. St. Louis was 11th and lost to 16th Mini, but their corsi deference wasn't that big, so of course there is no surprise that the favorite didn't win. After all, it lets you forecast probability, it does not grantee a result. See for your self below:
https://stats.hockeyanalysis.com/teamstats.php?disp=1&db=201415&sit=5v5&sort=CFPCT&sortdir=DESC

And your last paragraph is strait up silly. No, no team gives up shot attempts on purpose and forfeits shot attempts due to their system - that is a myth. Same with players. There are teams who play the "limit shots to the outside" systems and still have great shot metrics and end up succeeding as predicted, and there are the reverse who have terrible metrics and end up sucking.

I said why in my post.. goaltending (11-12) and over the top production (13-14). It's as simple as that.

So you are basically trying to claim "actually the statistical model that predicts them as a likely winner is wrong, because I can come up with arbitrary reasons that they won that aren't that statistic". You can hold on to that, but it's not very convincing. Their corsi indicated that they would be a likely favorite to win during those years, and they did win twice! That confirms the model, not the reverse. How can you not see that?