-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
NHL Off-Season 2015 |OT| With the first pick the LA Kings are proud to select...
- Thread starter Kave_Man
- Start date
I just caught up to this episode.
John is such a retard, it's amazing. My favourite thing was when the judges came over and basically told him he was an idiot and we was just staring at them like a dog thinking he was doing something great.
That really was a great challenge though
zroid looks like such a goober
I actually liked him until that challenge too... so that was disappointing
by the way, LOL @ the egg challenge
I actually liked him until that challenge too... so that was disappointing
by the way, LOL @ the egg challenge
Me too. I thought he was great as a captain in that first team challenge.
The egg challenge that
they all messed up lol.
Hockey buzz has some stuff gf doesn't.
Coldnoodle
Member
Unhandled Exception
Member
I give you shit because you've even said you're well aware of the incredibly limited narrative of current advanced stats analysis. What's strange is that you're apparently still bought into that framework of analysis when it clearly only details one dimension of a multidimensional picture. I'd think that intellect you speak of would be better pointed at figuring out the rest of the picture rather than re-enforcing a limited narrative.
Zaraki_Kenpachi
Member
I really hope we drop Martin and Ehrhoff, it would clear so much cap. We'll probably resign at least one with our dumbass GM.
Trust me, I am very open to all the new information that is gathered, and actively support trying to improve the current state of analytics. However discarding the advancements that have already made is stupid. If you'd open your mind a little bit you'd probably realize that the stuff that exists is incredibly useful, and less "one-dimensional" than you perceive it to be.
Unhandled Exception
Member
It's purely possession driven which is not even close to the entire story of a game. But if you're happy with your team entirely rebuilding around a perspective that is what you've admitted to be limited more power to you. Everyone needs a little hope and faith in their life.
Yeah that was a pretty good play. Smart positioning and anticipation.
I'm not really that down on him, he's obviously very talented. I just have my doubts about how it will translate to the NHL. Projecting defence is hard. Outwitting NHL opponents is different from CHL ones.
KernelPanic
Member
It's purely possession driven because that's the best predictor of winning right now.
Which is the only thing that matters.
Kuro Madoushi
Unconfirmed Member
4th line role players are relatively easy to sign in the off season with minimal impact to the cap.
I say that and admit we didn't really have much of a 4th line two years ago and last year's was better but still not good.
So long as he anchors the PK and defensive zone and he uses his size to create some chances...he won't be a bust at least, but I wouldn't go around telling other people proudly that drafted a legit 4th liner...
Unhandled Exception
Member
My entire point is that it's not the only thing happening in a game that is responsible for the outcome. People treat it as such though, like Samyy for example, and point to corsi charts to explain why a team won or lost a game that they didn't even watch in the first place. That sort of attitude about this shit would be fine if the team with the best corsi ALWAYS won but we all know that isn't the case. Which is why we need a better understanding of things, and why pushing the current understanding so hard does such a disservice to advanced stats in general.
I'll take one.
Here's the thing though. If the bolded were true, then corsi would have almost no value as a statistic. It's valuable because it describes something different from game-to-game results. That's the sort of thing you need to wrap your head around to understand why analytics work, and why they're useful. Its predictive power is greater than its descriptive power.
Is it everything? Nope! But the problem arises when people try to make excuses for poor analytics. It doesn't work like that. You can compensate a little bit. But over the long term it always comes back to bite you in the ass unless you are at the very least competitive possession-wise. It's important to look at the big picture and accept possession's place in it, not attempt to make excuses for it. If what you see disagrees with the analytics then try to understand why rather than think "I know what I saw, this data is trash". That isn't productive or helpful.
Unhandled Exception
Member
I don't think I've ever said the data is trash, merely the way you and everyone else is interpreting it is very much flawed. You seem to think that I'm anti-advanced stats and I'm not. I'm just not sold on how they're being peddled around currently and think it's fucking silly as hell to structure an entire rebuild around something that is at best "useful but not everything". And as far as this goes:
tell it to your comrades.
Simon Belmont
Banned
IMO there is a whole lot of counting the hits and ignoring the misses in the advanced stats community. You never really see people posting the graph that doesn't explain a game or player, even though they're out there. I can't really speak authoritatively about whether that's the actual case, but it certainly seems that way to me.
That doesn't mean I don't think they're useful, but put me in the "not the whole story" category.
That doesn't mean I don't think they're useful, but put me in the "not the whole story" category.
Unhandled Exception
Member
The confirmation bias is strong with a lot of people in that circle.
If you want to do stuff like quickly see how much money Chicago already has tied up for 15-16 ($63.15 million for 13 players) then their cap site works well.
http://www.hockeybuzz.com/cap-central/team.php?team=CHI
KernelPanic
Member
Hockey is too random to predict the outcome of a game or even a season/Stanley Cup with 100% certainty.
The "Corsi movement" is about a stat that is the best predictor of long term success. No guarantees but it's been fairly accurate since they started tracking it.
And that's all there is to it.
Unhandled Exception
Member
I'm fully aware of what corsi is and isn't. Frequently corsi gets treated as The Answer to any question of "why did this team lose, why is this team doing worse now than before". While it will always certainly play a role, propping it up as something other than a useful statistic is laughable and is a lot of what I see when corsi enters the conversation.
edit: And I'm done talking about it now. I've said my piece, I'll let your regular corsi programming continue unfettered.
That's not really true though. The hard part of possession stats is putting the data into the right context. Patrick Kane is a pretty good example. He's an elite talent, no one in the world would dispute this. But his possession stats are pretty below average. It's all about understanding the value of a player in his role, and how to maximize it. One of the things I think Quenneville has done a good job of in Chicago is maximizing Kane's strengths while hiding his weaknesses. That's very valuable.
Simon Belmont
Banned
Sure, but the eyeball test pretty clearly shows he's a great offensive talent and he's not too interested in defense. I don't need Corsi for that.
I do have a question for the advanced stats community though: I don't mean this to be snarky, and I'm probably just ignorant here, but I do have something of a background in stats (sociology). What's the P value of a Corsi stat? How is sample size balanced against player development? What is considered an appropriate sample size to draw a conclusion? I'm genuinely curious as to how these are addressed!
Early warning: I haven't had to think about stats in two years, so I haven't. I might be way off base here.
I do have a question for the advanced stats community though: I don't mean this to be snarky, and I'm probably just ignorant here, but I do have something of a background in stats (sociology). What's the P value of a Corsi stat? How is sample size balanced against player development? What is considered an appropriate sample size to draw a conclusion? I'm genuinely curious as to how these are addressed!
Early warning: I haven't had to think about stats in two years, so I haven't. I might be way off base here.
Well, he's an extreme case, obviously. Just because your eyes agree with the data doesn't mean the data isn't useful.
Shea Weber is a better example where the "eye test" doesn't match up with the data. Here's the thing though: just because his numbers aren't great it doesn't mean he's a bad defenceman. He's probably just less valuable than people think. When you put his results into context you see that he's pretty good at certain things but his skillset doesn't maximize possession. It's not about the stats failing to capture the value of a player, it's (again) about context.
As for the underlying statistics, all of the original research is readily available online, I'm pretty sure. If you're genuinely interested in leaning more, then tweet at some of the analysts who work on this stuff and they'll probably help you out. @SteveBurtch, @acthomasca, @behindthenet, to name a few.
Weber and Josi got by far the highest QoC and lowest percentage of Off Zone starts on the Predators. It's not surprising that affected their possession stats.
http://www.behindthenet.ca/nhl_stat...3+5+4+6+7+8+13+14+29+30+32+33+34+45+46+63+67#
http://www.behindthenet.ca/nhl_stat...3+5+4+6+7+8+13+14+29+30+32+33+34+45+46+63+67#
:lol :lol
Watch how often his line gets pinned in their zone for long periods of time. If you're going to hide his weaknesses you don't put him with 2 other players that don't play defense either.
And if you want to maximize his strengths, give him better opposite wingers than Versteeg/Bickell.
Simon Belmont
Banned
Yeah, I was actually just reading a Burtch article about P value. I don't know man, he gave it a P value about 15 orders of magnitude lower than anything you'd see in a scientific paper. That's when I remembered why I hated stats.
You should have gone through my advanced economic forecasting class in college. I'd rather have kidney stones every day for the rest of my life than do that class again.
fair enough lol, I'm just thinking about long-term success here. I think Q has done a good job with him overall over the years, maybe recently not so much.
anyways, good talk guys.
Not really, no. Kane being very good at hockey stuff overcomes a lot of Q's stupidity.
Pretty much. I never took anything over a college level algebra class so I was fucked from the start. :lol
haha alright. well i'll defer to you on this, I'm definitely not a Q usage expert. i'll give you he definitely makes some weird decisions, although most coaches do unfortunately.
Simon Belmont
Banned
Haha, no thanks. I had no problem with Calculus, because I can (could) derive the shit out of anything, but stats was like pulling teeth. And that is why I chose the squishiest science I could.
Zaraki_Kenpachi
Member
I think you know deep down that Q has never been to blame and it scares you...
Looking at the free agents, there's literally no one else so we'll probably overpay and resign both. None of our youth that we haven't traded doesn't look promising enough either.
I had to basically teach myself some calculus just to get by. That class was so shitty though. The highest grade was a 53%. My 39% curved up to a C+ at the end. :lol
With jokes like that you could open for Jay Leno.
Simon Belmont
Banned
Lol, no shame there. I passed easier classes than that thanks to the bell curve.
I even stopped doing the homework about halfway through because I had no idea what I was doing. I got an A on a group project that was worth 20% of the final grade because I had a buddy of mine in the class who did know what he was doing and I just went along for the ride on that one. Pretty much the only reason I passed. :lol
Kuro Madoushi
Unconfirmed Member
I keep imagine Marvie would do a spit take anything positive is said about Quenneville.
He cleans up his own floor dumps?I keep imagine Marvie would do a spit take anything positive is said about Quenneville.
Sometimes.