Assume you could quantify anything, what stat could revolutionize hockey evalutation?

[Filthy Dangles]

Hypothetical. If you could quantify and put a number on any hockey event that happens on the ice, what stat(s) could take evaluating hockey players and teams to the next level?

Hockey is such a fluid and dynamic game that it's hard to capture and quantify everything in a more static game like baseball. Use your imagination. We basically live and die by shot (attempt) metrics at the moment for anything other than goals and points.

Click to read more...

 

[ESH]

A star that tracks when players break up plays, poke the puck away, or any other good defensive play would be interesting to see. This would make someone like Drew Doughty seem even more impressive

 

[x Tame Impala]

Time of Possession. Just like in the EA NHL games. It gives a simple and accurate representation of controlling the play

 

[Filthy Dangles]

A star that tracks when players break up plays, poke the puck away, or any other good defensive play would be interesting to see. This would make someone like Drew Doughty seem even more impressive

thats a pretty good one.

Time of Possession. Just like in the EA NHL games. It gives a simple and accurate representation of controlling the play

agree. you almost forget we dont have this because Shot Attempt metrics are often touted as 'Possession metrics' which is clearly false.

offensive zone possession time probably correlates very well with winning hockey

 

[Filthy Dangles]

Time of Possession is a big obvious one that we don't have yet but he should.

Another one being Zone Entries.

To score you need to get the puck into the offensive zone. I am sure the vast majority of goals occur when you carry the puck in with possession as opposed to dumping it in and trying to get it back. So players who facilitate clean zone entries probably contribute to scoring goals. Tracking these seems pretty easy too just like how the scorers track attempted shots manually so I m not sure why we don't have these yet. Clearly very valuable in evaluating players and not too hard to obtain and track.

 

[Diamondillium]

If you could quantify 'true points' where goal, assist, and assist 2 are awarded simply to the 3 players who caused the goal rather than specific touches, as well as being able to accurately quantify 'fault points' for a goal against, you would pretty much be able to solve hockey.

 

[Rysto]

Goal expectation. Come up with a formula that for any shot, gives the chance that the average NHL goaltender allows it for a goal. This fixes a lot of problems that Corsi has, especially over shorter time periods.

This stat would allow you to disentangle the effects of team defence from goaltending. Great defenses will have a low expected goals value. Great goaltenders will consistently allow fewer goals than expected goals predicts. A team's offence can be judged by their expected goals for.

From there, you can begin to look at individual player performances by examining their expected goals for and expected goals against, just like people do now with CF%.

And it would finally shut up those people who go on about "intangibles". If intangibles really exist, it would show up as a higher expected goals for or a lower expected goals against. If a player's "intangibles" aren't actually translating into more offence or better defence, then they don't actually help the team win (although to be honest, this still wouldn't people from asserting they exist anyway...)

 

[Mr Snrub]

Goaltending is still a pretty imprecise art, most people default to SV%. I'd love to a see a goalie stat that measures something like their ability in clutch situations, or even something as simple as SV% in one-goal games or something.

 

[Beezeral]

an actually verifiable stat that quantifies how many wins a player is responsible for per season.

 

[tarheelhockey]

Quality of Play

 

[Fugazy]

Quality of Play

This.

 

[Braunbaer]

Intangibles.

 

[Filthy Dangles]

Intangibles.

"....What would you want quanitified..."

"Something that can't be quantified"

seems legit

 

[Farnham4top6]

Penalties called/Penalties commited
Would be a nice stat, it probably would look something like this 5/200

 

[jw2]

"....What would you want quanitified..."

"Something that can't be quantified"

seems legit

Asks question in OP

Gets mad when someone perfectly answers OP

Seems legit

 

[Braunbaer]

"....What would you want quanitified..."

"Something that can't be quantified"

seems legit

If you use phrases like "assume you could quantify anything" in the thread title and your first "sentence" in the OP consists of the lone word "Hypothetical" ... then yeah, my answer is 100% legit.

And it wasn't even supposed to be some stupid Toews-joke or anything like that. Just an honest answer following your premise.

 

[Goose]

Accurate WAR.

 

[Narow]

Passing stats...numbers of passes, % of completed/failed/ intercepted passes etc

A value on quality of a players shot..ingame meassurements of players skating and shooting speeds. G-force meters for acceleration, decceleration and turns. (Would be awesome for hitting aswell seeing who hits hardest and what what type of forces looks like, this could also lead to safety in preventing concussions if we find out at what type of impacts the risk is too high. Perhaps harder but safer hits could be done too)

Obviously you should be able to toggle on and off those got all players or even individual players to you liking.

Maybe its going away from stats and more into interface but i want to be able to be more involved from the couch

Oh and stats for puck on blade aka true possesion combined with giveaways. Would let us know who the best possesion players are

 

[Jot]

-Successful Zone Entrys
- # of Crossbars hit
- Successful dumps % (puck is retrieved) For Dmen
- Puck Retrieval % (from dumps) For Fwds

Just thinking of ideas off the top of my head.

 

[Aceboogie]

Time of Possession is a big obvious one that we don't have yet but he should.

Another one being Zone Entries.

To score you need to get the puck into the offensive zone. I am sure the vast majority of goals occur when you carry the puck in with possession as opposed to dumping it in and trying to get it back. So players who facilitate clean zone entries probably contribute to scoring goals. Tracking these seems pretty easy too just like how the scorers track attempted shots manually so I m not sure why we don't have these yet. Clearly very valuable in evaluating players and not too hard to obtain and track.

Time of possession is a tricky one. I hear it all the time people want it because they want ACTUAL time of possession, not a proxy via shot attempts. But ill pose you this question: Why is time of possession more important? I mean if the point of a game is to score goals which involves getting shots on net, why is time of possession more important? Say a team just holds the puck in the corners of the rink, or up high with the defenseman and gets very few shots on net, why would that be more important that a team that gets more shots on net

The only way I could see actual possession time as being important is if you could measure players efficiency at turning possession time with the puck into actual shots. Like high possession time and low shot totals is inefficient. High shot + high possession time means a dominant player

 

[Aceboogie]

-Successful Zone Entrys
- # of Crossbars hit
- Successful dumps % (puck is retrieved) For Dmen
- Puck Retrieval % (from dumps) For Fwds

Just thinking of ideas off the top of my head.

NHL.com tracks crossbars, posts hit etc. Jamie Benn led with 6

http://www.nhl.com/stats/player?aggregate=0&gameType=3&report=shottype&pos=S&reportType=season&seasonFrom=20162017&seasonTo=20162017&filter=gamesPlayed,gte,&sort=shots

 

[Aceboogie]

Goal expectation. Come up with a formula that for any shot, gives the chance that the average NHL goaltender allows it for a goal. This fixes a lot of problems that Corsi has, especially over shorter time periods.

This stat would allow you to disentangle the effects of team defence from goaltending. Great defenses will have a low expected goals value. Great goaltenders will consistently allow fewer goals than expected goals predicts. A team's offence can be judged by their expected goals for.

From there, you can begin to look at individual player performances by examining their expected goals for and expected goals against, just like people do now with CF%.

And it would finally shut up those people who go on about "intangibles". If intangibles really exist, it would show up as a higher expected goals for or a lower expected goals against. If a player's "intangibles" aren't actually translating into more offence or better defence, then they don't actually help the team win (although to be honest, this still wouldn't people from asserting they exist anyway...)

This exists, expected goals. Takes into consideration shot location, distance and type of shot. Uses historical data of goalies save % on these type of shots to come up with difficulty of different variety of shots. Its team and player seperated

Also goalies have Goal Saved Above Average which takes into consideration the number of high, mid and low danger shots and their sv% for those different shots. So a goalie facing way more high danger chances and saves a ton has higher GSAA than a goalie who faces more low danger shots

 

[s3antana5757]

Time of possession is a tricky one. I hear it all the time people want it because they want ACTUAL time of possession, not a proxy via shot attempts. But ill pose you this question: Why is time of possession more important? I mean if the point of a game is to score goals which involves getting shots on net, why is time of possession more important? Say a team just holds the puck in the corners of the rink, or up high with the defenseman and gets very few shots on net, why would that be more important that a team that gets more shots on net

The only way I could see actual possession time as being important is if you could measure players efficiency at turning possession time with the puck into actual shots. Like high possession time and low shot totals is inefficient. High shot + high possession time means a dominant player

The Bruins feed on NESN tracks zone time. Now they only track THE PUCK, not who has the puck. So if a defenseman is standing by his net with the puck on his stick, still counts toward the others team zone time. Jack Edwards is annoying with this stat, but sometimes it's a good stat.

 

[Aceboogie]

The Bruins feed on NESN tracks zone time. Now they only track THE PUCK, not who has the puck. So if a defenseman is standing by his net with the puck on his stick, still counts toward the others team zone time. Jack Edwards is annoying with this stat, but sometimes it's a good stat.

This is why I have a problem with it now as is. D that stand behind the net and let all teams get set up will crush it in this stat. Or say a D likes to move it quick and makes really good outlet passes, should be rank lower in this than a D who holds on to it in own end and lets opposing team set up a great scheme in NZ. For me the quick pass D will benefit the team more

I do think there is good uses for it, I just question why itd be more valuable than shot metrics. I think if time of possession was "created" by shot metrics were, people would be saying the opposite. IE, why do we care about time of possession when time of possession doesnt win games, goals generated by shots do. I think the argument is two fold for time of posssession supporters:
1) If your team has the puck, other team doesnt so cant score. True but again, shot metrics can measure this for shots against your team
2) More time of possession lets your team get a better scoring position. Again thats true but we have shot location data

I want time of possession data but I do think it wont be a better stat than anything we have now

 

[Five Alarm Fire]

Line Chemistry/Player Chemistry.