|
|
Post by Jon-Seattle on Jun 22, 2019 14:12:51 GMT -5
How the hell is Zach Werenski, a young talented first pairing defenseman end up with a defensive rating of 67, how?!? Brandon manning has a better defensive rating with a worse +/-, much worse corsi scores and having only played 40ish games. Werenski also played 45 seconds of short handed play a night, not great but significant because relative corsi was a positive 4 something in that time.
I get I’m always complaining about ekblads shitty defensive score but this is outrageous. Oh and Sergachevs 53.7 corsi gets him a defensive rating of 56??? Are you sure you got the right ratings because that’s garbage. And if you tell me it’s because he only played an average of 13 seconds short handed I swear I might just lose my mind because that’s atrocious, downright negligent, and on the best team in the league over the regular season he had a relative corsi of 3. Come on
|
|
|
|
Post by Glenn-Philadelphia on Jun 22, 2019 14:57:30 GMT -5
Regarding Sergachev's averaging 13 seconds of PK time a game. He is a defensive giant compared to Gostisbhere who at 5 seconds average per game came in with a whopping 51 def rating.
I see you mentioned his Corsi rating. Corsi makes up about 5-10% of the score.
I will have to relink to the post that discusses everything that goes into d ratings so you can check but how his +/- versus team +/-, giveaways/takeaways per minute, blocked shots, etc values compared to other comparable d-men.
Every year I say if someone can come up with a better solution to calculate def rating I am all ears and would be all over implementing it. Umfortunately, the usual response is "it's just wrong" which doesn't lead to much positive progress on implementing a better solution.
|
|
|
|
Post by Chris-Suffolk on Jun 22, 2019 15:26:42 GMT -5
Jon, just wait until you see his rfa asking price! Even if he only has a 67 D Rating, his 8 Stamina Rating will blast him up to the 6 Million asking zone!!! Maybe more...
|
|
|
|
Post by Jon-Seattle on Jun 22, 2019 16:26:46 GMT -5
I was going to start studying for medical school this week, instead I’m going to have to focus on fixing this god forsaken mess. Time to dust off the excel spreadsheets and neglect the fam. What else do I got to do though right?
|
|
|
|
Post by Scott-New York on Jun 22, 2019 16:54:32 GMT -5
Instead of trying to fix 'this mess', play him on your top pairing with someone who is a little more reliable in the D zone and you'll love the results. The stigma in this league of 'my defensemen are worthless unless they have 80+ D ratings' is misguided.
|
|
|
|
Post by Jon-Seattle on Jun 23, 2019 0:09:47 GMT -5
GHL defensive workbook.xlsx (17.01 KB) Attached is a small sample size of NHL defensemen, I don't have the almighty database Glenn has but I thought this might do for now just to compare a few guys. First I would like to point out that Keith Yandles stick check of 7 doesn’t pass the eye test. 27 takeaways should not equal a 7 on his card but i'm already off track. Second Pesces got the best defensive card in the game right now he is 54/30 in takeaways/giveaways, +35, 131 blocks and corsi of 53 with significant SH TOI but relative corsi is negative indicating that somethin may be amiss here. Essentially the counting stats are there for pesce but he seems to be a byproduct of his teams success and not the reason for it. Blocks are a flashy stat for defensmen but I'm not sure i wanna rest my hat on it. Upon closer examination of the leaguewide takeaways forwards seem to be more prevalent with league leader Mark stone at 122 and Barkov in 2nd at 100 even. in essence pesce had half as many takeaways as stone and barkov but we're running with defensive players only so lets compare him to the next defensive guys. We have Brett Burns at 88 (TOI 25.06) and Jacob Slavin at 86 (TOI 23.02). Both also played a full 82 game season though compared to Pesces 73 so now he's gunna get a boost in his /60 score and he played 3-5 minutes less so another bump in /60 final score, i get it but i'm not sure if i like it. I did a rough calculation for Burns (2.58) and Slavin (2.74), Pesce is close at 2.17, still not sure if its a 10 worthy but he's close as far as defensemen I'd guess. Pesce also had fewest giveaways but not sure if this is much of a defensive stat as much as offensive passing component. Blocks per 60 pesce leads the others on this list but is only 48th in total hits, maybe slightly higher in /60s though? I know I'm rambling and actually starting to make a case for Brett Fing Pesce as one of the better defensemen in our league, but he's a second pairing guy on a good team it just doesnt seem right. Thats where i think the relative corsi score comes into play. Brett Pesce has a 53 corsi, sure, but a -3.6 relative to the rest of his team, so i would argue hes being bouyed by his team, and therefor the laymen stats appear inflated. so what about something like corsi adjusted by half of relative corsi as the base defensive statistic. A modifier would then be used to add in the appropriate sub components. Subcomponents would comprise of takeaways being used to influence stickhandling, Hits remain a neutral category, Blocks/60 for shot block, +/- and PDO used as part of lane control and/or deny space. I dunno i'm just spitballing here because it seems just wrong that I have two top pairing defensemen sitting at near league worst ratings. Same with Gostisbehere, his Corsi is 52.2 and relative corsi is a whopping 5.0. By all means add in a small modifier for SH TOI because ya the best defensive defenseman will be playing those minutes but that does not mean he is your best defenseman in todays NHL. For this lists purposes i think the changes I suggest would land Jones at the top with Werenski close behind, Pesce or Sergachev would fall in line next obviously Sergachev doesnt have the raw stats which would still pull him down a bit which is completely fair. Yandle and Ekblad would fight it out for the next two spots with Savard near the back and Girard taking up the rear. Now i know these aren't all encompassing but i think that would pass the sanity check but maybe I'm wrong. I know im kind of shooting one of my own arguments down by moving an entire top pair down near the end of this list but this isn't a bunch of slouches here and I think the panthers hang Yandle and Ekblad out to dry quite a bit so again passes the sanity test. Just as another sanity check maybe look at a guy like Lovejoy, how the hell is he in the top 10 (7th) defensmen in our league, that is aweful, he plays 17 some minutes a night and corsi of 46.9 and relative of -2.3 oh but he has over a hundred blocks so that must mean he's amazing, hell no, he's horrendous. Obviously we have to separate NHL vs AHL stats as well and maybe we can add in a bit here where all AHL stats count 60% towards card values, as compared to NHL stats. unfortunately no corsi values could be found in my quick search for AHL players so we still wont be able to run a proper AHL farm system using this but maybe this gets the discussion headed in the right direction. Again we could work out the specifics here with a little fine tuning but right now these defensive rankings aren't passing the eye test and we need to fix it. Then again I'll probably just get told to fuck off and play with what I have so we'll see if this even goes anywhere.
|
|
|
|
Post by Scott-New York on Jun 23, 2019 4:53:38 GMT -5
The problem is you are only looking at defensive numbers. I agree that those defensemen are all better players than Lovejoy, but his 'defensive' stats are better. I'm sure he is on the ice more when killing penalties and obviously he blocks more shots. We can't give these guys better D scores, hence making them better defensemen in your mind, than the guys who are actually on the ice in defensive situations. Doing so would solve one problem and create another. I'll try to compare the cards for you later, right now I have to go to work.
|
|
|
|
Post by Phil-Cornwall on Jun 23, 2019 6:21:27 GMT -5
Yeah, Mark Stone's Defense is only 54 and there are at least 25 forwards better than that.
What the HELL ?? This is BULLSHIT
lol
but no, seriously, gotta like Jon's feisty attitude on this.
|
|
|
|
Post by Jon-Seattle on Jun 23, 2019 10:44:09 GMT -5
The problem is you are only looking at defensive numbers. I agree that those defensemen are all better players than Lovejoy, but his 'defensive' stats are better. I'm sure he is on the ice more when killing penalties and obviously he blocks more shots. We can't give these guys better D scores, hence making them better defensemen in your mind, than the guys who are actually on the ice in defensive situations. Doing so would solve one problem and create another. I'll try to compare the cards for you later, right now I have to go to work. (16.1 KB) That's the problem though, his numbers aren't good. literally his only number that is decent is his blocks. Everything else is average. His takeaways and giveaways are middle of the road for the guys I have listed. Hits again average. +/- of -1 not good. His Corsi is atrocious, literally lowest I've seen so far for an "every day" player. Oh but ya know what the combination of New Jersey and Dallas provided him an average of 3:23 SH TOI so he must be absolutely brilliant. two scores does not a player make but right now the scores for defensemen only seem to be predicated on blocks and SH TOI. prove me wrong. |
|
|
|
Post by Glenn-Philadelphia on Jun 23, 2019 11:11:21 GMT -5
|
|
|
|
Post by Jon-Seattle on Jun 23, 2019 12:09:28 GMT -5
I get that these are the stats you are using but what is the actual formula because I'm saying something isn't adding up. Right now Ben Lovejoys Corsi, +/-, and takeaways/giveaways should be pulling his card down much more than they are. I would need the individual game stats for Lovejoy to compare his actual +/- and same with the other guys that haven't played a full season and just am not prepared for that yet but its still not that positive. Also the blocks category is just painful to see that highly thought of around here. You're more likely to be screening your goalie than actually block the damn shot but that's a matter for another time. Updated chart with heat maps for your perusal: GHL defensive workbook.xlsx (16.99 KB) As a further aside, Ben Lovejoys stick check of 9 is completely ridiculous he has a total of 28 takeaways, no way in hell he's in the top 20% of defensemen with 28 takeaways. Block shots ok 8-9 I can see that but he doesn't actually create takeaways |
|
|
|
Post by Glenn-Philadelphia on Jun 23, 2019 12:42:51 GMT -5
I get that these are the stats you are using but what is the actual formula because I'm saying something isn't adding up. Right now Ben Lovejoys Corsi, +/-, and takeaways/giveaways should be pulling his card down much more than they are. I would need the individual game stats for Lovejoy to compare his actual +/- and same with the other guys that haven't played a full season and just am not prepared for that yet but its still not that positive. Also the blocks category is just painful to see that highly thought of around here. You're more likely to be screening your goalie than actually block the damn shot but that's a matter for another time. Updated chart with heat maps for your perusal: View AttachmentAs a further aside, Ben Lovejoys stick check of 9 is completely ridiculous he has a total of 28 takeaways, no way in hell he's in the top 20% of defensemen with 28 takeaways. Block shots ok 8-9 I can see that but he doesn't actually create takeaways Jon, Please understand that the stick check rating, and all of other non core ratings that are subset of OFF/DEF/NEUT is a complete subset of the respective composite score ratings and I don't believe have even been calculated yet. Any GM has the ability to provide input into these via the player rating page. The sum of these scores wont exceed a function of the composite score. See below video regarding how this accomplished. |
|
|
|
Post by Phil-Cornwall on Jun 23, 2019 12:55:08 GMT -5
We have no evidence that the GM rating actually results in player enhancement.
Change my mind.
|
|
|
|
Post by Scott-New York on Jun 23, 2019 14:28:33 GMT -5
Jon,
I understand your argument. However, it appears to me that you're focusing on the defensive rating and there is so much more to a player card. As Glenn pointed out, the sub scores are likely not complete yet and once they are, I'd bet Lovejoy's stick check goes down. Also, I understand that +/- goes into card creation, however, it is +/- for the player in comparison to the team which is the best we can do to make this absolutely meaningless stat meaningful. Now, let me get to my point of view on Werenski vs. Lovejoy and see if this makes sense.
Werneski:
Card 63-67-67
Offensive Aggregate (45 shot + 34 shot % + 52 pass) = 131
Stamina 8
Lovejoy:
Card 45-99-89
Offensive Aggregate 74
Stamina 6
As you can clearly see, Werenski is relied on to play big minutes, most likely top pairing 5 on 5 and top pairing powerplay, or at least that's what it can translate to in the GHL. Lovejoy is a bottom pairing 5 on 5 and top pairing penalty killer. If these two players played on the same team, an NHL coach is going to put Lovejoy out to kill penalties while Werenski will be stapled to the bench in that situation. Now I agree that maybe Werenski deserves a slightly higher defensive rating than 67 but he certainly shouldn't be rated higher than Lovejoy, just on NHL use alone. However, Werenski is still more valuable, both in the NHL and GHL. Take Werenski's card. His total card value is 197 (63 + 67 + 67). However taking stamina into account, 197 x .8 = 157.6 vs Lovejoy's rating using the same formula of 139.8. That's a 17.8 point difference in total value, not to mention the 57 point difference in offensive talent alone. The end result is that I would take Werenski over Lovejoy every day but they both have a purpose and are both valuable.
I will disagree with you on one other point you made. I doubt, without running the numbers, that Pesce is the best defenseman in total card value. I'd bet it's Mark Giordano. Yes, Norris Trophy winning, Mark Giordano!
|
|
|
|
Post by Jon-Seattle on Jun 23, 2019 14:39:34 GMT -5
I would argue that the GM input only allows for the redistribution of certain subcategories and in no way alters the overall defensive score. I could make ekblads shot block a 10 if I want to lower every other category, extreme example. Yes if used properly it helps round out abilities but not if the overall defensive score is wrong from the get go.
I know i’m Challenging the almighty status quo here and so am not likely to pick up a lot of traction but I think my small sample size is onto something. Players are getting undue boosts for having great SHTOI and blocks. Should they be rewarded, absolutely, but not when it would appear that these two stats outweigh all the others by a significant margin.
|
|
