News Article: The Definitive Way to Rank Players, per Sportsnet (Advanced Stats etc.)

[Ghost of Ethan Hunt]

Below is an excerpt from the Sportsnet article ranking the TOP 20 C's the last 3 yrs... For those who are familiar with advanced stats & have the skill/time/desire to compile the data set for our roster, I thought the specific stats used in each category provide a comprehensive analysis tool to rank our players vs. each other and/or the league. Several of you, i.e. Henkka, are very good at this (unlike myslef).

I thought this'd be great when we have NHL data on our 4 Dmen prospects +Ras/Z2/Svech etc. throughout the year.


https://www.sportsnet.ca/hockey/nhl/definitive-ranking-nhls-top-20-centres-three-seasons/

Each season was weighted so that recent seasons account for more of a player’s total score: 22.2 per cent from 2015-16, 33.3 per cent from 2016-17, and 44.5 per cent from 2017-18. But if a player only played in two of those years, they would see a weighting of 42.8 per cent for 2016-17, and 57.8 per cent for 2017-18.

Here are the statistics used in each category:

Offence: 5-on-5 and power play goals, primary assists, secondary assists, high danger scoring chances, scoring chances, passes to the slot, scoring chances off the rush, passes off the rush, rebound recoveries, scoring chance generating plays, shot attempts, penalties drawn, on-ice goals for relative to teammates (all per 60 minutes), and offensive zone pass completion rate.

Transition: 5-on-5 outlet passes, stretch passes, controlled carries out of the defensive zone, neutral zone passes forward and east-west, controlled entries into the offensive zone (all per 60 minutes), Corsi for percentage, Corsi for percentage relative to teammates, faceoff win percentage by zone, controlled exit rate relative to teammates, controlled entry rate relative to teammates, outlet, stretch, and neutral zone pass completion rates relative to teammates.

Defence: 5-on-5 and shorthanded loose puck recoveries by zone, blocked passes, stick checks, body checks, blocked shots, puck battles won, penalties taken, on-ice goals against relative to teammates, on-ice shot attempts against relative to teammates, on-ice passes to the slot against relative to teammates, on-ice high danger chances against relative to teammates (all per 60 minutes), turnover rate by zone relative to teammates.

Degree of difficulty: Puck IQ’s competition faced percentages (2016-17 only), Game Score’s quality of forwards faced, Game Score’s quality of defencemen faced, Game Score’s quality of teammate forwards, Game Score’s quality of teammate defencemen, Corsi quality of competition, Corsi quality of teammates, time on ice quality of competition, time on ice quality of teammates, personal shooting percentage relative to career average, on-ice shooting percentage, on-ice save percentage, defensive zone starts percentage, 5-on-5 ice time, and overall ice time per game.

There’s a slight change to the charts this year as well. I’ll be showing how players rank both in percentile form, a straight up “what percentage of their peers are they ahead of” and also the percentage that the player impacts a category compared to the best player at their position, which I’m calling per cent of peak...

Click to read more...

 

[Ghost of Ethan Hunt]

Speaking of, I just found this: (links below pic)

Has anyone checked out this free iPHONE app- NHL Rank King

Creator Mason Black was a fellow participant at the recent #OTTHAC18 (Ottawa Hockey Analytics Confernce) with others such as Rob Vollman, Hockey Abstract, LA Kings; Micah Blake McCurdy, hockeyviz.com; Scott Cullen, TSN; Asmae Toumi, Hockey-Graphs; Timo Seppa, National Hockey League etc.

From his twitter feed @NHLRankKing-
"Download our FREE iPhone app to access daily fantasy, line combos, starting goalies & prospect/NHL rankings. Looking at data in a different way!

I'm excited to announce a big update to our FREE NHL Rank King app that will be available very soon. You asked for it, and we are going to deliver a fully customizable draft rankings no matter how complex your fantasy leagues are: Points Only Roto Head-to-Head Keeper Salary Cap"

For those interested, here is the poster I submitted to #OTTHAC18 If you have any questions/comments please don't hesitate to ask!

If you have an iPhone or iPad you can download it here. ‎Rank King

Included is the PNHLe.

PNHLe is a value that attempts to predict a prospect’s optimal point production at the NHL level when that player is in the prime of their career. You can think of PNHLe as the expected ceiling a player will hit if they reach their full potential.
You can read up more on the stat, where it comes from and the methodology behind how the algorithm was created here. https://www.nhlrankking.com/PNHLe.htm

https://www.nhlrankking.com/PNHLe.htm

 

[Frk It]

Nice to see Backstrom properly ranked for once.

 

[Henkka]

These are really good (realgud) ranks because they take so many things in account.

I think there are the right guys. Would have love to see old ones, how good Pavel and Z were on their prime or little after.

 

[TheMule93]

as long as these stupid stats say our players are good then im all for them

 

[silkyjohnson50]

These are really good (realgud) ranks because they take so many things in account.

I think there are the right guys. Would have love to see old ones, how good Pavel and Z were on their prime or little after.

Exactly what I was thinking. I have to imagine Datsyuk would have been near the top quite a few seasons as he was an advanced stats God and did so many things well.

 

[newfy]

The issue with this is that advanced stats will never tell the whole story in hockey, theres too much going on. How much do you weigh rebound recoveries on offense into the equation vs neutral zone loose puck recoveries for example? They obviously have a formula that accounts for these 2 stats but how much are those worth in a real like scenario?

It is nice to have in conjunction with the eye test though to really get the full picture

 

[kliq]

I love that Toews is ranked 18th. Good player, but I’m tired of hearing him in the same breathe as Crosby, Malkin, and McDavid.

 

[InjuredChoker]

the definitive part likely ain't true, for the reason @newfy already brought up.

but on the other hand, i personally don't have big issues with this list.

 

[The Zetterberg Era]

I love that Toews is ranked 18th. Good player, but I’m tired of hearing him in the same breathe as Crosby, Malkin, and McDavid.

This isn't prime Toews only a few Hawks fans aren't acknowledging his decline. Honestly that he came in at 18th (in terms of that high) surprised me. Malkin over Crosby which is nice to see is how I have felt for most of his prime anyway, but certainly true of last year.

 

[Henkka]

This isn't prime Toews only a few Hawks fans aren't acknowledging his decline. Honestly that he came in at 18th (in terms of that high) surprised me. Malkin over Crosby which is nice to see is how I have felt for most of his prime anyway, but certainly true of last year.

Malkin wins Crosby statistically, because Crosby takes the toughest matchups always.

 

[newfy]

Malkin wins Crosby statistically, because Crosby takes the toughest matchups always.

That is accounted for in these rankings. Crosby isnt Crosby of 5 years ago anymore, he hasnt even broke 90 points in 4 seasons. Malkins last season was really good and he should be ranked higher right now

 

[Flowah]

That is accounted for in these rankings. Crosby isnt Crosby of 5 years ago anymore, he hasnt even broke 90 points in 4 seasons. Malkins last season was really good and he should be ranked higher right now

Have the concussions affected him long term, in terms of playing ability? He still seems to be pretty damn productive.

The issue with this is that advanced stats will never tell the whole story in hockey, theres too much going on. How much do you weigh rebound recoveries on offense into the equation vs neutral zone loose puck recoveries for example? They obviously have a formula that accounts for these 2 stats but how much are those worth in a real like scenario?

It is nice to have in conjunction with the eye test though to really get the full picture

I think the idea is that in the long term, with enough data points, that stuff sorta evens out.

 

[Henkka]

That is accounted for in these rankings. Crosby isnt Crosby of 5 years ago anymore, he hasnt even broke 90 points in 4 seasons. Malkins last season was really good and he should be ranked higher right now

When Crosby plays against shut-down pairs or Top lines, Malkin plays against 2nd pairs. This should be easy to understand. They play in totally different environment. Malkin has always easier opponents = inflated scoring.

Looking only on points... that was maybe the way on last millennium.

 

[newfy]

When Crosby plays against shut-down pairs or Top lines, Malkin plays against 2nd pairs. This should be easy to understand. They play in totally different environment. Malkin has always easier opponents = inflated scoring.

Looking only on points... that was maybe the way on last millennium.

Apparently you didnt read the article, quality of competition is taken into account on this ranking. Malkin plays very similar competition to Crosby and its accounted for in this ranking. Malkin comes out ahead still. Crosby isnt the Crosby of old anymore

 

[Ghost of Ethan Hunt]

Apparently you didnt read the article, quality of competition is taken into account on this ranking. Malkin plays very similar competition to Crosby and its accounted for in this ranking. Malkin comes out ahead still. Crosby isnt the Crosby of old anymore

I've thought Malkin was the superior player going back to '06/'07. He is much more physically imposing in all zones, size/physicality/reach. Though I'd like to see Sportsnet add Quality of Linemates to the equation somehow. Malkin used to get Kessel & Hornqvist iirc on his wings quite a bit, whereas Crosby seemingly had less than stellar linemates. But the media darling golden boy got 3 cups, 2 of which were legitimately won....27 seconds too many men on the ice uncalled penalty/Malkin's punch to the back of the head after the game/Franzen clobbered coming off concussion #??/ repeated cross-checks on Z by Crysby uncalled/Gary Buttman scheduling the finals 2 days after WCF. The '09 cup should have an "*" after it.

 

[kliq]

This isn't prime Toews only a few Hawks fans aren't acknowledging his decline. Honestly that he came in at 18th (in terms of that high) surprised me.

Trust me, a lot more fans then you think try to sell Toews as the elite of the elite. I could post proof, but I'm guessing I can't post links to other boards.

 

[The Zetterberg Era]

Trust me, a lot more fans then you think try to sell Toews as the elite of the elite. I could post proof, but I'm guessing I can't post links to other boards.

I think that is an older narrative that hasn't happened a heck of a lot for the last 12 months.

What I was driving at was I feel like most have dropped that, it was surprising to see him even come in the top 20. Honestly for me it was a because I never felt he was top 5 I think perhaps I have been a bit harsh with his decline as a result. Perhaps his fall off isn't quite as dramatic as I thought at least not yet, because I don't think he is going to regain a whole lot of prowess though. But still to see him come in that high was notable and impressive to me at least.

 

[Invictus12]

I've thought Malkin was the superior player going back to '06/'07. He is much more physically imposing in all zones, size/physicality/reach. Though I'd like to see Sportsnet add Quality of Linemates to the equation somehow. Malkin used to get Kessel & Hornqvist iirc on his wings quite a bit, whereas Crosby seemingly had less than stellar linemates. But the media darling golden boy got 3 cups, 2 of which were legitimately won....27 seconds too many men on the ice uncalled penalty/Malkin's punch to the back of the head after the game/Franzen clobbered coming off concussion #??/ repeated cross-checks on Z by Crysby uncalled/Gary Buttman scheduling the finals 2 days after WCF. The '09 cup should have an "*" after it.

In all honestly, in game 6 and 7 of that series, the Wings only seemed to show up for the 3rd period when already down on the score sheet. Maybe the team was a bit arrogant coming off game 5 where they just completely dominated the game from a to z.

 

[Winger98]

In all honestly, in game 6 and 7 of that series, the Wings only seemed to show up for the 3rd period when already down on the score sheet. Maybe the team was a bit arrogant coming off game 5 where they just completely dominated the game from a to z.

I think they were just gassed and beat up.

 

[Invictus12]

I think they were just gassed and beat up.

They certainly showed energy in the third which makes m believe it was more of a motivation issue.

 

[Dexter Colt]

The issue with this is that advanced stats will never tell the whole story in hockey, theres too much going on. How much do you weigh rebound recoveries on offense into the equation vs neutral zone loose puck recoveries for example? They obviously have a formula that accounts for these 2 stats but how much are those worth in a real like scenario?

It is nice to have in conjunction with the eye test though to really get the full picture

I wholeheartedly agree.

I do statistics for a living, and reducing such a dynamic environment into a a statistical model scares the bejesus out of me. The use of some simple ratios for whatever purpose they might serve, I do understand. But ranking players on aspects such as "offense", which comprise of an arbitrarily chosen and weighed set of measures, is sketchy. After all, the statistics will tell you what you want, as long as you abuse them enough.