2020-2021 VHL Ratings Package

[Minnesota Wild]

Patrice Bergeron has a higher rating in every single category except slightly lower Skating and Strength and lower endurance. And sort of significantly better in most. I don't think he's getting too worried about Skinner challenging him for the Selke in real life or in VHL.

Skinner has an overall rating the same as Tyler Motte, doesn't seem like he's being overrated in any significant manner even if his defense seems high.

[Montreal Canadiens]

But you didn’t answer how skinner starts off with 0 out of 15% for PK, because he can’t get any of that % if he has zero seconds killing penalties right?
Secondly how is plus minus not used in determining 5 on 5 effectiveness lol, any advanced stat has some correlation to who you are on the ice with and against, hockey is a team game, so one guy is -16 and the other is plus 10, they both play 14 minutes a night approx.

I’m just scratching my head on how if he start at 85 - his obviously isn’t getting 70/70 for 5 on 5 play and I’m assuming blocking 13 shit in 15 games isn’t giving a whole lot of that 5% nor is he getting much of def zone starts at 44% which has to be in the bottom 25% of the league.

And how does the algo collect the other 11 rating categories? Or are those subjective?

[Utah Hockey Club]

Aug 13, 2021 18:54:16 GMT -5 Montreal Canadiens said:

But you didn’t answer how skinner starts off with 0 out of 15% for PK, because he can’t get any of that % if he has zero seconds killing penalties right?

You're right, for that 15% of the DF rating, Skinner rates out very poorly, however the lowest rating a player can get is a 57, not a 0. However, he rates our quite well in the 70% of the RAPM xGA/60

Secondly how is plus minus not used in determining 5 on 5 effectiveness lol, any advanced stat has some correlation to who you are on the ice with and against, hockey is a team game, so one guy is -16 and the other is plus 10, they both play 14 minutes a night approx.

If you think that plus minus is a credible stat, then I think the conversation between us is going to end soon. What if one of Armia's linemates goes for a lazy change and Armia steps on the ice and 2 seconds later the opponent scores. Armia gets dash 1. Is that a reflection of Armia's abilities? No. Plus minus is more of a result of team play and not individual play.

I’m just scratching my head on how if he start at 85 - his obviously isn’t getting 70/70 for 5 on 5 play and I’m assuming blocking 13 shit in 15 games isn’t giving a whole lot of that 5% nor is he getting much of def zone starts at 44% which has to be in the bottom 25% of the league.



Read my point above about how a player can get a minimum of a 57 rating, not a 0.

And how does the algo collect the other 11 rating categories? Or are those subjective?

Here's some light Friday night reading for you on the thesis and methodology behind RAPM and the statistical model. If you have any further questions maybe reach out to them. hockey-graphs.com/2019/01/14/reviving-regularized-adjusted-plus-minus-for-hockey/ 

[Montreal Canadiens]

And I guess part of what I’m asking is how is the 10% for zone starts calculated, Patrice Bergeron is deployed 37% in d zone and danault is at 63%. If your lead the league is d zone starts you get the full 10 or is it based on 100 d zone starts = 10 90% “
=9 etc.. same for the others, if 1 blocked shot per game leads the league for fwds does that give you 5% and working backwards from that?
As well as the 15% for killing penalties? Does it go by min per game on PK, Team PK efficiency?
And the 5 on 5, if your top 10% in the league do you get the full 70 pts, 20% gets you 60 working down?

I’m assuming for pass and sc ratings it’s purely assists/gp gives you a rating or do you weigh PP/SH and EV assists differently?

[Montreal Canadiens]

Any advanced stat is going to have a correlation with who you play with on the ice, same as plus minus. Is Jeff skinner plays with 4th line plugs all year he isn’t going to have good numbers, in any category, nor would anyone not in the superstar conversation, so while I ageee that plus minus isn’t the be all end all of stats, it also cannot be ignored.
We definitely see different things, you’re obviously a Kyle Dubas fan who wants to build his team strictly based on numbers and I’m more of a watch the game and tell me who is a good hockey players type of guy lol. And as a 3rd line puck possession, PK, shut down guy I’m taking Joël Armia over Jeff skinner 100% of the time, he didn’t just be a 4 x 3.5 deal for his ppg numbers.

[Utah Hockey Club]

Aug 13, 2021 19:11:05 GMT -5 Montreal Canadiens said:

And I guess part of what I’m asking is how is the 10% for zone starts calculated, Patrice Bergeron is deployed 37% in d zone and danault is at 63%. If your lead the league is d zone starts you get the full 10 or is it based on 100 d zone starts = 10 90% “
=9 etc.. same for the others, if 1 blocked shot per game leads the league for fwds does that give you 5% and working backwards from that?
As well as the 15% for killing penalties? Does it go by min per game on PK, Team PK efficiency?
And the 5 on 5, if your top 10% in the league do you get the full 70 pts, 20% gets you 60 working down?

I’m assuming for pass and sc ratings it’s purely assists/gp gives you a rating or do you weigh PP/SH and EV assists differently?

Ratings are based on z scores. Best in the league gets a 99, worst in the league gets a 57. Z scores don’t act in a linear fashion. If interested in more information about them Google it. I don’t feel like teaching you stats haha.

Assists are weighted based on EV, PP, and SH as well as primers and secondary where primary EV are weighted the heaviest and PP secondary are weighted the least.

Scoring is 90% goals/gp while 10% xGF/60.

[Utah Hockey Club]

Aug 13, 2021 19:19:25 GMT -5 Montreal Canadiens said:

Any advanced stat is going to have a correlation with who you play with on the ice, same as plus minus. Is Jeff skinner plays with 4th line plugs all year he isn’t going to have good numbers, in any category, nor would anyone not in the superstar conversation, so while I ageee that plus minus isn’t the be all end all of stats, it also cannot be ignored.
We definitely see different things, you’re obviously a Kyle Dubas fan who wants to build his team strictly based on numbers and I’m more of a watch the game and tell me who is a good hockey players type of guy lol. And as a 3rd line puck possession, PK, shut down guy I’m taking Joël Armia over Jeff skinner 100% of the time, he didn’t just be a 4 x 3.5 deal for his ppg numbers.

Ok, I’ve tried to be nice here, but I’m not trying to be Kyle Dubas. I’m trying to give this league a reliable ratings package that is devoid of bias. The fact that you think that plus minus is a reliable stat that needs to be included is a fucking joke man. The 1990’s called and they want their infinite hockey wisdom back.

I’ve spend the past 4 months working on this in my limited free time between raising a new born, working full time and waking up at 5am every morning so I can run 90km/week as I train for a marathon.

Tell me Rick, would you like to do the ratings package for next year? I’m pretty confident what I’ve made here would blow whatever Mickey Mouse package you came up with out of the fucking water.

Let me say it loud and clear. JOEL ARMIA IS NOT AS GOOD AS YOU THINK. Fuck me.

[Montreal Canadiens]

LOL now you're being funny, maybe a bit of bias is needed then if advanced stats are giving a guy with zero defensive acuity better numbers then guys who make their living filling the role. There are literally articles on how bad Jeff skinners 2 way game is.

You're making this about Armia, from what i see overall the DF rating has some work to do moving forward when Kappo Kakko and his zero PK time, and 37 %DF zone starts is rated as well as Phillip danault in that category.

Or Tanner Jeannot, ya who, having a better rating package with 15 career games played then 3/4 of the league.

Im sure you put a shit ton of hours into this, and i have no idea how to make it better, but if you dont get constructive feedback the final product wont improve over time. And a word of advice, although we really appreciate the time you put into this, if you're stretching yourself thin with a young child at home on top of your job ect... spend the time with the kid, we will manage with shitty generic ratings from someone else, at the end of the day this is a distraction, some years i spend time on it, others years life gets in the way, at the end of the day, life comes first.

[Utah Hockey Club]

Aug 13, 2021 22:07:26 GMT -5 Montreal Canadiens said:

There are literally articles on how bad Jeff skinners 2 way game is.

Or Tanner Jeannot, ya who, having a better rating package with 15 career games played then 3/4 of the league.

You know those articles you read about Jeff Skinner's terrible two-way game? They must of been from before the most recent season because he was a terribly dreadful defensive player, in fact worse than Armia was this year and this is support by the data as below is his chart for the 2019-2020 season:




and the year before in 2018-2019 when he had the 40 goal season:



So it appears that while Jeff Skinner was previously a porous defensive player, the coaching staff in Buffalo was able to teach a dog new tricks so to speak for the most recent season.

The issue here is that due to Skinner's poor play in previous seasons, you have a pre-conceived notion that he will forever be a poor defensive player and couldn't possibly be a defensively responsible player, when in fact he became that this most recent season.

This is a prime case in why bias should be be removed from ratings package because people not in tune with the day-to-day of players will rate the players on their past work instead of what they've done most recently.

I will agree that Tanner Jeannerot needs to be downrated. I'll need to put in a game multiplier to bring down guys that didn't play a whole season moving forward.

[Florida Panthers]

Aug 13, 2021 22:07:26 GMT -5 Montreal Canadiens said:

LOL now you're being funny, maybe a bit of bias is needed then if advanced stats are giving a guy with zero defensive acuity better numbers then guys who make their living filling the role. There are literally articles on how bad Jeff skinners 2 way game is.

You're making this about Armia, from what i see overall the DF rating has some work to do moving forward when Kappo Kakko and his zero PK time, and 37 %DF zone starts is rated as well as Phillip danault in that category.

Or Tanner Jeannot, ya who, having a better rating package with 15 career games played then 3/4 of the league.

Im sure you put a shit ton of hours into this, and i have no idea how to make it better, but if you dont get constructive feedback the final product wont improve over time. And a word of advice, although we really appreciate the time you put into this, if you're stretching yourself thin with a young child at home on top of your job ect... spend the time with the kid, we will manage with shitty generic ratings from someone else, at the end of the day this is a distraction, some years i spend time on it, others years life gets in the way, at the end of the day, life comes first.

Why are you getting so worked up about something that is “just a distraction”?
Armia is worse then Skinner at defence, either produce some stats that back up your argument or just deal with it lol