The Advanced Stats Thread Episode VII: An Ode to the Sanity of Silverfish

[Megustaelhockey]

This is not the lounge, nor is it a place to talk about what's wrong with the rest of HFNYR.

Talk about advanced stats in here, or you will be removed.

 

[Machinehead]

This is not the lounge, nor is it a place to talk about what's wrong with the rest of HFNYR.

Talk about advanced stats in here, or you will be removed.

 

[YoSoyLalo]

If @silverfish gets banned from this thread he may as well delete his account

 

[Ghost of jas]

I know you are but what am I

Notice you are not being ignored.

 

[Machinehead]

I was reading about the Norris finalists and again, people are lobbying for an Orr Award for best offensive defenseman.

Am I the only one who doesn't see the point? First of all, you're not the best all-around defenseman if you're not one of the top scoring defensemen. It's right there in the award description. Moreover, I don't see where the Norris has been absolutely dominated by points. The finalists are all top 10 scorers, but the top 4 most productive defensemen all missed out. I think, in a vacuum, that's perfectly reasonable.

At the end of the day, what it comes down to is: we just have to admit that Erik Karlsson is good and deserved to win, and your solid #2 doesn't.

Offense is part of the job, and it's part of winning the Norris.

DISCLAIMER: This may not be "advanced stats" per se, but I value the general takes on defensemen in here

 

[GeorgeKaplan]

I was reading about the Norris finalists and again, people are lobbying for an Orr Award for best offensive defenseman.

Am I the only one who doesn't see the point? First of all, you're not the best all-around defenseman if you're not one of the top scoring defensemen. It's right there in the award description. Moreover, I don't see where the Norris has been absolutely dominated by points. The finalists are all top 10 scorers, but the top 4 most productive defensemen all missed out. I think, in a vacuum, that's perfectly reasonable.

At the end of the day, what it comes down to is: we just have to admit that Erik Karlsson is good and deserved to win, and your solid #2 doesn't.

Offense is part of the job, and it's part of winning the Norris.

DISCLAIMER: This may not be "advanced stats" per se, but I value the general takes on defensemen in here

It comes down to people not really watching players and not wanting to change their reputation (not that I think Karlsson was ever really “bad” defensively), and the good ol’ Canadian bias

 

[Hi ImHFNYR]

Ah, the next advanced stats thread is open! Time to show all the things I know about hockey...through the miracles of math!

And.... wait.

N-no. But I have so many insights. What do I do? What the F do I DO?

Nah. Go ahead.

Oh thank God. To this day the thread would have remained empty had you not said this. Thank you...for releasing us from purgatory.

 

[Irishguy42]

 

[Machinehead]

Now those are advanced stats

 

[Ail]

Finally, the best individual goaltending seasons are:

load("~/corsica_data/war_full.RData")
war %>%
  filter(position == "G") %>%
  arrange(desc(WAR_Goalie)) %>%
  select(player, season, WAR_Goalie) %>%
  head(20)

## player season WAR_Goalie
## 1 TIM.THOMAS 20102011 11.839463
## 2 HENRIK.LUNDQVIST 20112012 10.860530
## 3 SERGEI.BOBROVSKY 20162017 9.937583
## 4 JONATHAN.QUICK 20112012 9.439853
## 5 MIKE.SMITH 20112012 9.374637
## 6 TIM.THOMAS 20082009 8.165371
## 7 CAREY.PRICE 20142015 7.967805
## 8 TUUKKA.RASK 20122013 7.866080
## 9 HENRIK.LUNDQVIST 20132014 7.755855
## 10 CAM.WARD 20082009 7.252770
## 11 HENRIK.LUNDQVIST 20082009 6.938354
## 12 ROBERTO.LUONGO 20102011 6.856440
## 13 EVGENI.NABOKOV 20092010 6.717359
## 14 HENRIK.LUNDQVIST 20092010 6.494654
## 15 BRADEN.HOLTBY 20142015 6.484342
## 16 CAM.WARD 20102011 6.476188
## 17 HENRIK.LUNDQVIST 20122013 6.472801
## 18 JAROSLAV.HALAK 20092010 6.363367
## 19 JONAS.HILLER 20082009 6.331477
## 20 JONATHAN.QUICK 20172018 6.207274

Taking total WAR, we find the best individual seasons to be:

load("~/corsica_data/war_full.RData")
war %>%
  mutate(WAR = ifelse(position == "G", WAR_Goalie, WAR)) %>%
  arrange(desc(WAR)) %>%
  select(player, season, WAR) %>%
  head(40)

## player season WAR
## 1 TIM.THOMAS 20102011 11.839463
## 2 HENRIK.LUNDQVIST 20112012 10.860530
## 3 SERGEI.BOBROVSKY 20162017 9.937583
## 4 SIDNEY.CROSBY 20092010 9.722228
## 5 ALEX.OVECHKIN 20072008 9.453252
## 6 JONATHAN.QUICK 20112012 9.439853
## 7 MIKE.SMITH 20112012 9.374637
## 8 STEVEN.STAMKOS 20112012 8.387472
## 9 TIM.THOMAS 20082009 8.165371
## 10 CAREY.PRICE 20142015 7.967805
## 11 TUUKKA.RASK 20122013 7.866080
## 12 HENRIK.LUNDQVIST 20132014 7.755855
## 13 JOHN.TAVARES 20112012 7.626564
## 14 SIDNEY.CROSBY 20162017 7.573488
## 15 CAM.WARD 20082009 7.252770
## 16 CONNOR.MCDAVID 20172018 7.157724
## 17 JOE.THORNTON 20072008 7.043349
## 18 HENRIK.LUNDQVIST 20082009 6.938354
## 19 ROBERTO.LUONGO 20102011 6.856440
## 20 PAVEL.DATSYUK 20072008 6.718918
## 21 EVGENI.NABOKOV 20092010 6.717359
## 22 EVGENI.MALKIN 20072008 6.530675
## 23 HENRIK.LUNDQVIST 20092010 6.494654
## 24 BRADEN.HOLTBY 20142015 6.484342
## 25 CAM.WARD 20102011 6.476188
## 26 HENRIK.LUNDQVIST 20122013 6.472801
## 27 JAMIE.BENN 20132014 6.391783
## 28 HENRIK.ZETTERBERG 20072008 6.378891
## 29 JAROSLAV.HALAK 20092010 6.363367
## 30 JONAS.HILLER 20082009 6.331477
## 31 ALEX.OVECHKIN 20092010 6.294403
## 32 JAROMIR.JAGR 20072008 6.279444
## 33 JONATHAN.QUICK 20172018 6.207274
## 34 ERIC.STAAL 20082009 6.175381
## 35 ALEX.OVECHKIN 20082009 6.173078
## 36 JOE.PAVELSKI 20102011 6.142695
## 37 STEVEN.STAMKOS 20102011 6.117706
## 38 CAREY.PRICE 20102011 6.092893
## 39 RYAN.MILLER 20092010 6.013737
## 40 STEVEN.STAMKOS 20092010 5.886627

ONLY ONE VEZINA THO...

 

[silverfish]

Every time Manny publishes something makes me want to do less and less because I'll never get to this level. Brilliant stuff. Hire this man.

I mean, yeah, maybe you don't agree with the WAR rankings, but the ingenuity and logic and work it takes to build something like this... there are 31 NHL teams that could use that.

 

[YoSoyLalo]

Crosby having a higher WAR in 09-10 than Ovechkin in 07-08 is interesting

 

[Greg02]

If I have a few minutes soon I'd like to see how aggregate team WAR corresponds to wins/points in the standings. Does anyone have thoughts on how to handle the dirty data of traded players? Do you think proportionally splitting their WAR by games played is good enough?

 

[silverfish]

If I have a few minutes soon I'd like to see how aggregate team WAR corresponds to wins/points in the standings. Does anyone have thoughts on how to handle the dirty data of traded players? Do you think proportionally splitting their WAR by games played is good enough?

I don't think so. If you're trying to check the WAR math to total team wins, you can't proportion it. You'd need to calc the exact WAR for each team. Dude could be a -1 player on one team, and a +2 player on the other, but I don't think you can aggregate that to a 0.5 for both teams.

But this WAR stuff is where I officially bow out, because I really don't know enough.

 

[YoSoyLalo]

Also, I feel like goalie WAR could benefit from being its own stat, but I don’t know math, so don’t take that suggestion too seriously

 

[Machinehead]

I have my issues with WAR in its current iteration as I've outlined, but I do hope it succeeds.

Once we have one or two really strong metrics by which to judge players, we can move onto projection stats, which find good players who aren't good yet, and see the wall before players hit them.

I love the Yankees bringing in Brandon Drury. He doesn't make hard contact at a good rate, but he makes medium contact at an elite rate. Put him in the AL East where every stadium is a little league stadium and watch him magically go from 13 to 25 homers. Just a shame he had migraine issues and then an incredible prospect stole his job.

But either way, it would be cool to see moves like that in hockey where you project index success based on success with particular scales.

 

[Greg02]

I don't think so. If you're trying to check the WAR math to total team wins, you can't proportion it. You'd need to calc the exact WAR for each team. Dude could be a -1 player on one team, and a +2 player on the other, but I don't think you can aggregate that to a 0.5 for both teams.

But this WAR stuff is where I officially bow out, because I really don't know enough.

Yeah, you're probably right. Kind of makes things into a much bigger project, unfortunately.

Also, I feel like goalie WAR could benefit from being its own stat, but I don’t know math, so don’t take that suggestion too seriously

What does 'being its own stat' mean in this context?

-----------------------

By the way, the graph of defensemen split into DWAR and OWAR is pretty fascinating and not shaped at all like I would expect.

 

[silverfish]

Once we have one or two really strong metrics by which to judge players, we can move onto projection stats, which find good players who aren't good yet, and see the wall before players hit them.

I made one of these. It was terrible. One thing that was kind of useful about it though was the similarity score to compare players to each other. But people hate similarity scores. Because you say Brady Skjei is comparable to John Moore and people lose their damn minds.

 

[Machinehead]

I made one of these. It was terrible. One thing that was kind of useful about it though was the similarity score to compare players to each other. But people hate similarity scores. Because you say Brady Skjei is comparable to John Moore and people lose their damn minds.

See, you and me were already insane and never had a mind to lose. That's the key.

 

[silverfish]

Rick Nash 5v5 leaguewide playoff ranks via Corsica:

ixGF60: 1st
iCF60: 1st

Tell me again how he's not playing "his game" anymore because of concussions.

 

[Mac n Gs]

Rick Nash 5v5 leaguewide playoff ranks via Corsica:

ixGF60: 1st
iCF60: 1st

Tell me again how he's not playing "his game" anymore because of concussions.

Because he’s shooting like 4% and narratives are narratives.

 

[Synergy27]

Because he’s shooting like 4% and narratives are narratives.

Why is he shooting 4% now? Is it luck or something else? How do you know?

 

[Mac n Gs]

Why is he shooting 4% now? Is it luck or something else? How do you know?

I see a good hockey player that generates a lot of scoring chances, mostly around the net, when I watch Nash. Stats back that up too, and I have to boil it down to bad luck when you can consistently generate both shots and dangerous scoring chances. The consistency with how his shooting % keeps dipping does make sense to me that some of his abilities are degrading as he ages, but it's been really strange to watch. He does so much correctly in all 3-zones, but his scoring has evaporated.

Then you see shit like Dermott's wrister go in, and you wonder why can't Nash get some luck like that. It's even more baffling that he doesn't have more than 1 assist, especially since I've thought he's been really solid setting up plays this series.

 

[silverfish]

Nash has been great through two. I've been counting his shot assists and I have him at 4. He had a great set up to Kreijci early in the 2nd that was a dangerous chance.



Chara and McAvoy tho

 

[Harbour Dog]

Delete - wrong thread