Why are people so obsessed with so called advanced statistics?

[Kshahdoo]

I mean, good puck possession numbers are probably good, but if you don't have players, who can score, you can possess the puck as long as you wish, it won't help you a lot. It can take just a few seconds to score a goal for a competent opposition, and then you can keep possessing the puck.

Shots for and against isn't bad as well, but what about quality of shots? One thing, if it's Tarasenko or Ovechkin, and the other one, if it's some 4th liners, who can't even elevate the puck.

Click to read more...

 

[Analyst365]

Because they're not actually advanced, they're quite simple and much easier to grasp than abstract concepts like team play, maximizing individual strengths, players hiding injuries, line chemistry, wavering confidence, diet, new born baby sleep derivation affecting play ... stuff that real hockey experts in the game have to assess on a day-to-day basis.

 

[tarheelhockey]

Because the numbers bear out. You can predict a long-term results better from possession numbers than from anything else.

 

[WarriorOfGandhi]

because I'd prefer not to watch hockey games

 

[JaegerDice]

Because they are repeatedly proven to correlate with winning more strongly than any other stat?

Shot quality is a unicorn. Spoiler: Your goalie is only going to stop between 90 and 95% of shots over the course of a season or playoff run, or hell, playoff series, regardless of shot quality.

A bad angle shot from far out is a low quality shot. A bad angle shot that ricochets off a stick is a goal, and by definition, the highest quality shot (given that scoring goals and not looking fancy is the actual goal).

Finally, don't mix up inputs and outputs. Nobody is ignoring leadership, teamwork, chemistry, work ethic, yada yada. Those are INPUTS, just as speed, size, IQ, vision, hands, etc are INPUTS. What matters is OUTPUT. How the team performs with said player on the ice, and more specifically, their ability to reliably, repeated impact goal differential and wins.

 

[Bill Waters*]

because I'd prefer not to watch hockey games

Bingo

 

[guyzeur]

Cause a lot of people like to play with numbers?

 

[weastern bias]

Because when observed over a long sample size they basically work

 

[Analyst365]

Because they are repeatedly proven to correlate with winning more strongly than any other stat?

So you only need the standings then.

 

[GetFocht]

This message board is a haven for arguments, each individual poster wants to be right so they use statistics to prove their own agenda.

Advanced stats are and will always be supplementary yet people take it as the be all, end all of hockey.

 

[JaegerDice]

So you only need the standings then.

No, standings points are a relatively weak predictor of both playoff success and repeatable winning season over season.

 

[aufheben]

Top 5 CorsiFor% teams post-lockout:

Detroit
Chicago
Los Angeles
San Jose
Boston

Noticing a trend here? These teams account for like 10 SCF appearances.

 

[KevinRedkey]

Clinging on to advanced stats is a way to support a viewpoint/argument, without ever watching the player play (the eye test).

 

[Apotheosis]

Because it's the most consistent, observable statistic to predict if a team will be good or not. You kind of just walked into your own argument and tore it apart when you brought up Tarasenko and Ovechkin. The point is, you have the puck more, you have more chances to shoot. More chances to shoot, with better players means you will likely be scoring more than giving them up (which correlates to having the puck away from your own zone not letting opponent's shoot against you). It's no surprise that the best team of the last 10 or so years (the Hawks) are consistently a great possession team along with great players.

 

[Bill Waters*]

Because they are repeatedly proven to correlate with winning more strongly than any other stat?

Shot quality is a unicorn. Spoiler: Your goalie is only going to stop between 90 and 95% of shots over the course of a season or playoff run, or hell, playoff series, regardless of shot quality.

A bad angle shot from far out is a low quality shot. A bad angle shot that ricochets off a stick is a goal, and by definition, the highest quality shot (given that scoring goals and not looking fancy is the actual goal).

I don't think thats true. Right now the oilers have a a sub-50% corsi yet they are 2nd highest scoring team precisely because of the quality of their chances, i.e. break-aways, two-on-ones.

 

[Random Forest]

So you only need the standings then.

No. Standings describe historical information. That's great if that's what you're interested in, but not if your goal is to predict future outcomes. Enter analytics that have far better predictive value than historical data like wins, goals, points, etc.

 

[Apotheosis]

I don't think thats true. Right now the oilers have a a sub-50% corsi yet they are 2nd highest scoring team precisely because of the quality of their chances, i.e. break-aways, two-on-ones.

And by all metrics, it will be unsustainable. (At least from the stats from what I have seen). Recent wins are mostly based on Talbot standing on his head.

 

[CornKicker]

its a way that math lovers can evaluate actual athletes without having set foot in a dressing room let alone played the game. The great thing about advanced analytics and the argument the lovers will use is that the best teams have good advanced stats. of course they do, everyone with eyeballs or a inkling of hockey knowledge can also see they are good lol.

 

[Bill Waters*]

Because it's the most consistent, observable statistic to predict if a team will be good or not. You kind of just walked into your own argument and tore it apart when you brought up Tarasenko and Ovechkin. The point is, you have the puck more, you have more chances to shoot. More chances to shoot, with better players means you will likely be scoring more than giving them up (which correlates to having the puck away from your own zone not letting opponent's shoot against you). It's no surprise that the best team of the last 10 or so years (the Hawks) are consistently a great possession team along with great players.

And there's the rub. Look at last years Leafs: good corsi, yet lacks talent. Corsi does not measure talent. Good teams are built on talent/skill.

 

[aufheben]

Clinging on to advanced stats is a way to support a viewpoint/argument, without ever watching the player play (the eye test).

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.

 

[rhinoshawarma]

Theyre somewhat usefull for those who didnt watch the games. If you watch a game you usually can tell and decide for yourself who has better shifts, who has the puck more, who shoots more and has better chances. Ultimately they can give some information on how players affect the game and the swing of games but in the end the eye test trumps all, no pun intended

However some people rely too heavily on it. For example eller has great possession stats and fo% and i think hes great but it certainly doesnt mean he should be played like bergeron. Its important to get your info from many things, from the eye test, regular stats, advanced and insiders imo.

 

[JaegerDice]

I don't think thats true. Right now the oilers have a a sub-50% corsi yet they are 2nd highest scoring team precisely because of the quality of their chances, i.e. break-aways, two-on-ones.

And it's a 7 game sample.

Only 3 of the last 8 cup winners were not a top 2 possession team heading into the playoffs.

2006 Hurricanes:

Most penalties called in any season in modern history, disproportionate impact of special teams. Plus Cam Ward went on a run.


2007 Ducks:

2 HOF Dmen + a strong goalie performance.

2011 Bruins:

Tim Thomas + Canucks injuries.


Every other cup has been won by a top 2 corsi team heading into the playoffs, and most of the finals have been battles between two top 5 corsi teams, regardless of placement in the standings.

The best of those teams won 2 and 3 cups respectively.

 

[Apotheosis]

And there's the rub. Look at last years Leafs: good corsi, yet lacks talent. Corsi does not measure talent. Good teams are built on talent/skill.

That's why I put that in my post. To me, Corsi is only useful to predict when a team is competitive. Same applies to the Leafs now. Top 3 in high danger scoring chances and scoring for fun, but goaltending is letting us down. Once Andersen's numbers regress to his career mean and he adjusts to the system around him, the Corsi is predictive of the talent in front of him that is scoring for fun at a very sustainable 10.25 shooting %.

 

[Sens of Anarchy]

They're in vogue.

They can provide more in depth / additional information about a team and specific players situational effectiveness

They can be used in arguments and discussions about players

 

[PensBandwagonerNo272*]

Top 5 CorsiFor% teams post-lockout:

Detroit
Chicago
Los Angeles
San Jose
Boston

Noticing a trend here? These teams account for like 10 SCF appearances.

Pittsburgh? Washington? Tampa Bay?

Notice a trend here?

I mean if Detroit is number one, they haven't even been a contender for years really.