2020-2021 VHL Ratings Package

[Utah Hockey Club]

Aug 14, 2021 11:08:13 GMT -5 Montreal Canadiens said:

Aug 14, 2021 7:54:42 GMT -5 Edmonton Oilers said:

According to "Stats" jon merril is just as good as Ryan Mcdonnaugh in the 3 most critical categories, one is a top 4 d man on a stanley cup winning team, one was a healthy scratch and -11 on the team who they beat in the finals and the eye test would tell you he is probably even worse then that lol. Just saying, when the stat you are using is saying shitty defensive players are rating better then players deployed for their defensive skills maybe we are looking at the wrong metrics..

Well Jon Merill was in fact a solid defensive defenseman last year, maybe not so much in Montreal, but in Detroit where he played 75% of his games he was solid. McDonough was still an above average defender, but not as good as Merill. Where McDonough excelled compared to Merill is provide offensive output while Merill was a offensive black hole which is why McDonough had a higher OV rating among other reasons. Also, playoff stats have no effect on the ratings package since half the league doesn't play so it would be unfair to use an incomplete data set. 

Again, this is support by the data...

[Buffalo Sabres]

the personal shots need to be taken out of this.

I for one back up Rick in SOME of his claims.

I'm not fully convinced either on metrics of the NHL. this same bullshit argument is what makes me no longer have interest in baseball.

at the end of the day it does need to make sense. Adam did an amazing job here, but ues it's year one and there should be some criticism. Adam should also be able to handle some criticism and people should also be able to handle when he proves where the numbers come from.

If I had a vote, I would also suggest that maybe not all the metrics being used based on each category are correct, maybe the percentage breakdown of what each category isn't correct. who is to say for sure. It a year one of these ratings, it's a fantastic start. I would however like to see either more things come into play, perhaps, now this may be hard for some of you to grasp ..... Points, goals, assists, pk time, pp time, he'll even plus minus. sure call me dinosaur. I dont give a fuck. I'll just call ya a Millenial. New age isn't always right, I promise you all that. the argument of advanced stats and old school is super annoying. both sides need to see the other because the truth lies in the middle.
don't believe me, look at the NHL leafs.
advanced stats people have this high and mighty to them and I'm sorry, but it's annoying.

Anyone who makes ratings is gonna take this attack from time to time. it comes with the territory and I'm sure Adam can defend himself. I 100% appreciate all the work this was, as I fully believe Rick does as well. other chiming in with nothing to contribute but cut up Habs and people isn't necessary. healthy discussion is encouraged and needed I believe.

[Utah Hockey Club]

Aug 14, 2021 12:24:43 GMT -5 Buffalo Sabres said:

the personal shots need to be taken out of this.

I for one back up Rick in SOME of his claims.

I'm not fully convinced either on metrics of the NHL. this same bullshit argument is what makes me no longer have interest in baseball.

at the end of the day it does need to make sense. Adam did an amazing job here, but ues it's year one and there should be some criticism. Adam should also be able to handle some criticism and people should also be able to handle when he proves where the numbers come from.

If I had a vote, I would also suggest that maybe not all the metrics being used based on each category are correct, maybe the percentage breakdown of what each category isn't correct. who is to say for sure. It a year one of these ratings, it's a fantastic start. I would however like to see either more things come into play, perhaps, now this may be hard for some of you to grasp ..... Points, goals, assists, pk time, pp time, he'll even plus minus. sure call me dinosaur. I dont give a fuck. I'll just call ya a Millenial. New age isn't always right, I promise you all that. the argument of advanced stats and old school is super annoying. both sides need to see the other because the truth lies in the middle.
don't believe me, look at the NHL leafs.
advanced stats people have this high and mighty to them and I'm sorry, but it's annoying.

Anyone who makes ratings is gonna take this attack from time to time. it comes with the territory and I'm sure Adam can defend himself. I 100% appreciate all the work this was, as I fully believe Rick does as well. other chiming in with nothing to contribute but cut up Habs and people isn't necessary. healthy discussion is encouraged and needed I believe.

I agree Rick brought up some great points which I noted in my responses. A model will never be perfect, but the goal is to make it as close to perfect as possible and as you noted a key to this is feedback I which I’m thankful I got from Rick and others. Since this is v1, I will be taking some of this feedback and use it in subsequent years.

Rick asked questions and I answered them with fact based answers. He countered with his subjective thoughts on particular players.

Scott in terms of the analytics vs. old school argument, I don’t think there is one here. The model takes old school box car stats and supplements them with advanced stats where possible. DF is unique is that it’s the attempt to reduce events and so there really is no good traditional stat to account for this so that’s why an advanced stat takes the majority weight here. How does one accurately quantify the absence of something? Points, goals, assists, and pp time, are already included in other stat categories:

PH = 100% pts/gp
SC = 90% goals/gp
PA = 100% mix of EV, PP, SH and primary and secondary assists
DF = 15% PK TOI

Other basic stats such as hits and blocks are also included in other categories in one facet or another.

I will say I really enjoy the feedback. However, if someone says “why is it this way?” and I explain in great detail as to why is it that way. Please don’t tell me I’m wrong simply on your intuition. Tell me I’m wrong and back it up with facts that I can draw upon and I’ll be more happy to take the L and adjust the model to your feedback so it can become even better and make it a W for the league.

[Montreal Canadiens]

LOL ok, ask rod brindamour if he isnt still deploying Jordan staal in every key defensive assignment 110% of the time. The numbers dont line up with reality is the point.

as far as habs bias i just listed 3-4 other examples on how the metric being used to determine def ratings. At the end of the day i dont give a flying fuck but if we all just put our heads in the sand and say yay this is awesome there is no way to improve it.

John merrils advanced stats might say he was a norris player in detroit, watching him play the game of hockey tells me he is horrible, maybe luck had more to do with his numbers then reality in this case lol

[Utah Hockey Club]

Aug 14, 2021 16:30:45 GMT -5 Montreal Canadiens said:

John merrils advanced stats might say he was a norris player in detroit, watching him play the game of hockey tells me he is horrible, maybe luck had more to do with his numbers then reality in this case lol

Nobody nor any advanced stats describe John Merrill as a Norris calibre defenceman while playing in Detroit. He was an above average defender and a black hole for offence, that’s it. Nice try though.

[Calgary Flames]

Aug 14, 2021 16:51:07 GMT -5 Utah Hockey Club said:

Aug 14, 2021 16:30:45 GMT -5 Montreal Canadiens said:

John merrils advanced stats might say he was a norris player in detroit, watching him play the game of hockey tells me he is horrible, maybe luck had more to do with his numbers then reality in this case lol

Nobody nor any advanced stats describe John Merrill as a Norris calibre defenceman while playing in Detroit. He was an above average defender and a black hole for offence, that’s it. Nice try though.

What if he plays for Montreal? And then what?

[Montreal Canadiens]

Aug 14, 2021 18:09:13 GMT -5 Calgary Flames said:

Aug 14, 2021 16:51:07 GMT -5 Utah Hockey Club said:

Nobody nor any advanced stats describe John Merrill as a Norris calibre defenceman while playing in Detroit. He was an above average defender and a black hole for offence, that’s it. Nice try though.

What if he plays for Montreal? And then what?

He did, and he was awefull, what's your point?

[Anaheim Ducks]

I think maybe just let it go seems like a petty argument. Adam did a terrific job. Your points are both valid but it is what it is. 🤷

[Calgary Flames]

Aug 14, 2021 19:03:27 GMT -5 Montreal Canadiens said:

Aug 14, 2021 18:09:13 GMT -5 Calgary Flames said:

What if he plays for Montreal? And then what?

He did, and he was awefull, what's your point?

I thought he was supposed to be good tho? Saving grace for Montreal? No?

[Minnesota Wild]

So I was just thinking on this whole thing and figured I'd weigh in with a thought.

The brunt of this discussion is basically "are the player attributes reflective of real life". But honestly can we truly weigh in on that without even seeing a season of sim? We can look at a player.. let's say Armia and think "that defensive rating seems low", but are we even thinking about this correctly? Do we know how the sim incorporates "DF" into meaningful action within the sim?

What if a player is extremely effective at "defending" if the play is slow but has no skating or agility to defend against fast players? How do we put that in terms of a single category? What about the opposite a player who we would not consider great defensively but who is effective defensively because they are extremely positionally aware and fast even if they're still an offense first player?

I think we have to remember that these are player attributes that when put through a sim engine should hopefully replicate real life statistics closely. And without even running a sim we don't really know what needs to be tweaked yet. Maybe checking, skating and discipline are also important attributes as far as defensive effectiveness (all of which Armia rates quite fine in it would appear).

So it seems a bit to me like unless someone can definitively point to some stat that shows a flaw in the current ratings build's reasoning we don't really have much basis to be going around changing things yet. The question we want to focus on with ratings is about whether or not the players attributes will recreate real life stats, and we have no current basis for saying they won't as far as I can tell.

An interesting experiment would be to have all the lineups match last years teams, try and adjust the playing time and usage accordingly and then run a ratings package to see what it generates..