EHM The Blue Line

Aladyyn wrote:This is the year Ristolainen breaks into Norris talks. Best player on the Sabres this season and the one guy that we can always count on.

Gotta play like a 1st pairing D-man before you can break into Norris talks. He's a -8.65 GF% RelTM, -4.56 CF% RelTM and he scores almost all his points on the PP, he's not a good D-man 5v5 at all. Stats from Puckalytics.

Peter_Doherty wrote:

Aladyyn wrote:This is the year Ristolainen breaks into Norris talks. Best player on the Sabres this season and the one guy that we can always count on.

Gotta play like a 1st pairing D-man before you can break into Norris talks. He's a -8.65 GF% RelTM, -4.56 CF% RelTM and he scores almost all his points on the PP, he's not a good D-man 5v5 at all.

Maybe you don't want to answer me, and if so that's fine, but I really do like "data" (I'm editing EHM/dealing with "numbers" pretty much every day, and have been for years); my "black & white" viewpoints are part of who I am, and while I may not always communicate in the most effective way/make the best choices I'm honestly not meaning to be trolling/offensive (IMO my perspective is actually quite helpful fromn an EHM editing perspective & I am actively trying to improve my social graces)


My questions (and they're legitimate/honest questions):
What does "-8.65 GF% RelTM" and "-4.56 CF% RelTM" actually mean?
What would good/elite numbers be?


I realize I could do some research myself online and educate myself, but I'm currently plugging away on the WHL for the 2016-17 TBL Rosters and figure I should keep that as my focus right now.....and honestly, in the same way I think I'm one of the most knowledgeable about editing EHM/how EHM works, I think you're one of the most knowledgeable (if not the most knowledgeable) TBLer about advanced stats

Peter_Doherty wrote:

Aladyyn wrote:This is the year Ristolainen breaks into Norris talks. Best player on the Sabres this season and the one guy that we can always count on.

Gotta play like a 1st pairing D-man before you can break into Norris talks. He's a -8.65 GF% RelTM, -4.56 CF% RelTM and he scores almost all his points on the PP, he's not a good D-man 5v5 at all. Stats from Puckalytics.

There was a point in the season where Josh Gorges was below 35CF% when away from Risto. He might be the literal worst defenseman in the league. When Risto has a partner that's not completely atrocious he's a fantastic possession player. Context.



I wonder where the problem is, hmmmmm

As for low ES point production, that is a team-wide problem. Lowest even strength sh% in the league, Bylsma's conservative system are factors there.

nino - I'll just let the puckalytics glossary explain it for me

"TM Statistics
TM statistics are like a composite WOWY (With or Without You) statistic. Let's take Dion Phaneuf's TMCF60 (TeamMate Corsi For per 60 minutes of ice time) as an example. In this case Phaneuf's TMCF60 is a weighted average of each players CF60 when Phaneuf isn't on the ice with them weighted by the ice time they have played with Phaneuf (for players with fewer than 100 minutes apart from Phaneuf their overall CF60 is used instead to keep sample sizes from getting too small).

One could consider TM statistics an expected value for the player if the player had no (positive or negative) influence on his team mates results when he plays with them. "

"RelTM statistics
RelTM statistics are nothing more than the players statistic minus the equivalent TM statistic. So, CF20RelTM = CF20 - TMCF20. "

So around 0 i average/OK, + is good. When your +/- is going towards 5 then it's either super good or super bad.

Aladyyn - The issue with that is that he plays almost exclusively with Gorges so the sample size on the others is really low.
I get that Gorges is an anchor but even with an anchor he should produce better numbers if he is to call himself a Norris contender or even a legit 1st pairing D-man (5v5).
Context is absolutely important but it doesn't explain everything either, bad numbers are still bad.

Sure, but you plug him for Norris, him not scoring points 5v5 is a very good argument for why he shouldn't be talked about for Norris.

I stand by what i said before the season, Ristolainen is a borderline 2nd pairing D-man 5v5 who is very good on the PP, so overall i would call him a #3.

He's a couple games away from a 50-50 split of with Gorges/without Gorges.

Tom Pyatt turns Erik Karlsson into a 40CF% player even though EK is the best defenseman in the world. The only difference, Karlsson wasn't stuck on the ice with him for an entire season to create a dumb narrative.

So if he's close to 50-50 split with/without Gorges then how do you explain his AWFUL 'possession' stats? You first say it's only because of Gorges, then you point out that he only plays with Gorges 50% of the time.

Ristolainen also is an awful PKer, his FA/60 is ridiculously high.

I said he was a couple of games away, the difference is still ~100 minutes. And going from 42% to 52% only gets you so far even on a 50/50 split.

The PK number is caused by the Sabres' awfully passive PK system I assume.

Yeah, and my point has never been that RIstolainen is awful, just that he's not even close to a #1 D-man which you are arguing (?).

I haven't fact checked this myself but i think it was Sean Tierney who tweeted out his ridiculous PK numbers, they were by far the worst on Sabres.

He's getting closer and closer to #1 status. His issue was mostly consistency so I'm not crowning him yet, he needs to keep it up at least for the whole season.

Aladyyn wrote:He's getting closer and closer to #1 status. His issue was mostly consistency so I'm not crowning him yet, he needs to keep it up at least for the whole season.

I don't even think he's close, he's bleeding shot attempts and he's not scoring 5v5. What makes him a #1?

Just looked up his 4v5 stats since i couldn't find the tweet i was referring to, he's at 141 CA/60, worst on Sabres.

Aladyyn wrote:He's getting closer and closer to #1 status. His issue was mostly consistency so I'm not crowning him yet, he needs to keep it up at least for the whole season.

If you're saying that he's only getting closer to #1 status and that he still lacks consistency, I don't think he deserves to be in the talks for a Norris just yet. I think you kind of replied to your own previous statement there

Are you going down the Josh Gorges hole again?

The Norris talks was a prediction based on my assumption that he is going to keep his play up for most of this season. You can laugh all you want but the media love guys like Risto.

Aladyyn wrote:Are you going down the Josh Gorges hole again?

The Norris talks was a prediction based on my assumption that he is going to keep his play up for most of this season. You can laugh all you want but the media love guys like Risto.

If he's so good then how is it that there's not a single analytics person who thinks he's a 1st pairing D-man? A WAR model like DTMAboutHeart who heavily accounts for anchors (likes McDonagh even though he's just as anchored as Ristolainen) doesn't like Ristolainen at all. I haven't seen a single model that likes Ristolainen.

Why are models using data from past seasons relevant to how Ristolainen is playing this season?

Also the smart analytics people don't have a proper opinion on Ristolainen yet because of his anchor. Analytics people that watch him pretty much without fail think he's 1st pairing calibre

I haven't seen a single person into analytics say that Ristolainen is a 1st pairing calibre D-man.

But whatever, beating a dead horse here, i'm done for now.

Matthew Coller and Derek Jedamski have mocked the "Ristolainen is actually bad" narrative every chance they get for the past year. Coller less so now that he covers Minnesota but he still finds the time here and there

Let me also say that I'm a big fan of analytics, I follow it all and try to educate myself as much as possible. But, watching a player 80+ times a year sure doesn't hurt when evaluating.

In other news, Rangers are hit hard by the injury bug, we have pretty much a full middle 6 out right now (Zibanejad, Buchnevich, Nash, Vesey, Grabner and Puempel). Add that J.T. Miller is really banged up aswell. Grabner is home in Austria to attend his grandmothers funeral, Vesey is only day-to-day. With this many injured Fs the issues on D are really shining through, our insane F depth was masking our defensive issues during the first month of the season, now not so much.

Aladyyn - Arguing that he's not bad is a completely different thing to arguing he's a legit 1st pairing D-man. I don't think he's bad and i've never said he's bad.

Peter_Doherty wrote:nino - I'll just let the puckalytics glossary explain it for me

Thanks for the reply Peter, much appreciated!

What's the "GF%" referring to?

They both do believe him to be a top pairing defenseman. It's easy to see him that way on the ice.

nino33 wrote:

Peter_Doherty wrote:nino - I'll just let the puckalytics glossary explain it for me

Thanks for the reply Peter, much appreciated!

What's the "GF%" referring to?

No worries, happy to help

GF% is Goals For %, how many of the goals scored when the player is on the ice is for his own team.

Peter_Doherty wrote:GF% is Goals For %, how many of the goals scored when the player is on the ice is for his own team.

So obviously anything positive is good, and negative is bad.....is there a team average that can be compared with? So "-8.65 GF% RelTM" might be good if the team average is twice as bad...or am I not understanding?

What's a good/elite GF% for D and for Forwards? or is there a relevant difference between the two?

GF% is pretty much a fancy +/- tbh

Ur misunderstanding, the RelTM stats are in comparison to his teammates. So -8.65 is REALLY bad, but ofcourse some context is in order and Gorges really is an anchor but it's still really bad.

When you look at raw GF% and not Rel-stats anything over 50% is obviously good, i like to use Rel-stats more though since it accounts for good/bad team.

I usually don't use GF% that much because i find shot metrics and scoring chance metrics more predictive but it can be good to add sometimes because it does bring another context to it.

Peter_Doherty wrote:CA/60

Is this Corsi Against per 60 minutes?

Peter_Doherty wrote:A WAR model like DTMAboutHeart

What's this?