NHL.com will start tracking "advanced" stats

[sureLoss]

http://sports.yahoo.com/blogs/nhl-pu...medium=twitter

One of the knocks on advanced stats in the NHL is that they’re treated as official numbers by the League -- we can find goals and assists on NHL.com, but not Corsi and PDO. This allows critics to treat them like some kind of crazy nerd voodoo rather than an increasingly legitimate means through which to evaluate players, teams and games.

Well, it appears to the nerds won.

“Absolutely,” said NHL COO John Collins, at least weekend’s All-Star Game.

Collins confirmed that while the League waits to standardize the new player-tracking system that will eventually quantify puck possession stats more accurately, NHL.com is adding “enhanced stats” to its collection of player and team statistics.

The addition is scheduled for late February.

“You’re going to see a big change in the way we present our stats, in terms of the depth and the utility of how to do it. And that’s before the puck tracking [system],” said Collins.

Among the 30 or so advanced stats being added to the NHL.com stats pages:

Corsi, which estimates puck possession by totaling shots on goal, missed shots and blocked shots
Fenwick, which does the game minus the blocked shots
PDO, a combination of shooting and save percentage while on the ice at even strength
Zone starts, which designate what percentage of a player’s shifts end in which zone.
Average shot distance
Goals and assists per 20 minutes and 60 minutes
Penalties drawn and taken per 20 minutes and 60 minutes
The possession stats (Corsi, Fenwick) will also be presented in several ways, including game situations and as percentages in comparison to other players on the team.

[Jacks]

These stats would be a lot better if they somehow took into account the quality of shot. The classic argument is the oilers, they take 40 shots and give up 35 so they win the Corsi battle but 15 of the 35 shots they gave up were prime scoring opportunities while 30+ of the shots they took were from the perimeter.

Also they should take into account teams with a lead. A team that is up 5-1 isn't going to be pressing the offense as much as the other team gambling trying to catch up.

[WCW Nitro]

I wonder if video games will start including this too.

[Enoch Root]

give them real names already!

[GGG]

Originally Posted by Jacks

These stats would be a lot better if they somehow took into account the quality of shot. The classic argument is the oilers, they take 40 shots and give up 35 so they win the Corsi battle but 15 of the 35 shots they gave up were prime scoring opportunities while 30+ of the shots they took were from the perimeter.

Also they should take into account teams with a lead. A team that is up 5-1 isn't going to be pressing the offense as much as the other team gambling trying to catch up.

They are including average shot distance which at least will give some indicator of shot quality.

Also the Fenwick Close stat only counts games within 2 goals.

Good to add these stats and make them accessible for people. I can't wait to start seeing the numbers once the player tracking comes out. Depending on what gets publicised you could track the distance between an offensive player and all the defenders an measure how much a defense shifts to stop a certain player.

[Resolute 14]

Originally Posted by Jacks

These stats would be a lot better if they somehow took into account the quality of shot. The classic argument is the oilers, they take 40 shots and give up 35 so they win the Corsi battle but 15 of the 35 shots they gave up were prime scoring opportunities while 30+ of the shots they took were from the perimeter.

Also they should take into account teams with a lead. A team that is up 5-1 isn't going to be pressing the offense as much as the other team gambling trying to catch up.

Many variants do, actually. Base Corsi tracks one thing. Corsi Close, for instance, tracks that thing when the score is close. Some of the other websites go as far as Corsi relative to the quality of your competition - how well you do relative to the players you are facing. As time goes on, people will develop more metrics that tracks these things better.

See also baseball. AVG --> OBP --> OPS --> EQA

[Northendzone]

I wonder what the nhlpa's stance is on some stats like in game skating speed and shot speed being published - might dispel some notions about certain players.

Fenwick, fewick close, fen wick far, inverse fenwick - I guess my question is - how is this going to affect my hockey pools?

[sureLoss]

Originally Posted by Northendzone

I wonder what the nhlpa's stance is on some stats like in game skating speed and shot speed being published - might dispel some notions about certain players.

I think the NHLPA will be sort of happy. IIRC in the rules of arbitration, any stat the NHL officially tracks can be used as evidence. Therefore players can bamboozle arbitrators into thinking that Crosby's contract is comparable because they share Corsi numbers

[northcrunk]

Advanced stats are stupid.

[Street Pharmacist]

War-on-ice.com tracks "scoring chances" defined as shots in the slot, shots on the rush and rebounds as they were all associated with high shooting percentages. That stat is more predictive of future win/loss than any other stat so far

[formulate]

Originally Posted by Street Pharmacist

War-on-ice.com tracks "scoring chances" defined as shots in the slot, shots on the rush and rebounds as they were all associated with high shooting percentages. That stat is more predictive of future win/loss than any other stat so far

It's truly fantastic.

[MrMastodonFarm]

No RGI?

[Street Pharmacist]

Originally Posted by MrMastodonFarm

No RGI?

I'm sure someone around could keep you up to date on that one

[Roof-Daddy]

Originally Posted by Street Pharmacist

War-on-ice.com tracks "scoring chances" defined as shots in the slot, shots on the rush and rebounds as they were all associated with high shooting percentages. That stat is more predictive of future win/loss than any other stat so far

Where do the Flames rank?

[FlamesAllTheWay]

I can't imagine this will make contact negotiations and arbitration any easier.

That said, I'll always welcome anything that will paint a better overall picture of a players performance.

[driveway]

I just hope the NHL takes the opportunity to rename the stats. It's been entertaining to talk about Corsi and Fenwick, but I'd way rather something descriptive like Shot Attempts and Shot Attempts minus Blocks.

And PDO I feel obliged to dislike as it's named after the username of an Oiler fan on a messageboard. I don't really care what the NHL calls it so long as it's descriptive. "Cumulative Percentages" or "Shooting and Save Percentage."

[Street Pharmacist]

Originally Posted by Roof-Daddy

Where do the Flames rank?

Can't get war-on-ice.com to work at the moment, but it ain't pretty. Fifth last I believe. The biggest issue is not scoring chances against, it's that the Flames don't generate enough. They just score on a very high rate of their own scoring chances

[Matty81]

No issues with the stats being available but for stats that are supposed to be advanced and insightful most of these are pretty simple and stupid... Basic ratio analysis at best. Maybe helpful and illuminating in some ways but deeply flawed measures like corsi are getting pumped up way too much as a record of truth in determining how good a team is. I think what is most frustrating is when people are parroting these like they are more significant than actually watching hockey games.

It also almost seems like the advanced stat guys collectively agreed that more shots = better teams and keep coming up with more bad ratios to support the same conclusion. They're always going to be flawed. Distance away or location is the best they've got yet it's a terrible proxy for shot quality for example. Measures like shooting percentage are equated with luck which is totally counterintuitive... in other sports it's the key measure.

A team like the oilers have been decent at advanced stats this year but if you watch them, they are prone to several catastrophic defensive collapses a game that render any advantage in shots meaningless.

I love statistical analysis but some perspective is needed on these. Helpful but just another set of tools. Mostly created and promoted by tools.

[Tinordi]

Keep the names, they are homages to the early pioneers of this analysis.

[Caged Great]

Originally Posted by Street Pharmacist

War-on-ice.com tracks "scoring chances" defined as shots in the slot, shots on the rush and rebounds as they were all associated with high shooting percentages. That stat is more predictive of future win/loss than any other stat so far

That was the thing that I really disliked about the other advanced stats. There was no context. This is more useful. How many scoring chances are being generated and how they convert them is a lot more important than crappy shot attempts just because Cosri.