rune74
Registered User
- Oct 10, 2008
- 9,228
- 552
GDT: July 1st - Free Agent Frenzy | II
- Thread starter Zombotron
- Start date
So their other articles where they say he's a moron are fair balanced and objective too right?
me2
Go ahead foot
Cupless44
Registered User
- Jun 25, 2014
- 7,154
- 3,298
For sure, they have written articles criticizing Benning that are fair and well deserved.
Three out of the top 10 best signings based on analytics. Not too shabby.
Last edited:
Just goes to show, when you get sheltered minutes or get 'healthy-scratched' against the best teams like Gagner, Del Zotto and Wiercioch, you can become an 'analytics star'.....analytics basically means 'squat' once you're asked to start playing higher in the lineup and log more minutes.
StrictlyCommercial
Registered User
Not really. In terms of offense you would expect increased icetime to correspond with increased output. Defensive play is where increased minutes will lead to a player being exposed (See Luca Sbisa).
In that sense, Gagner should be a good signing if he is utilized well. The other two... meh.
I've said it a few times, but it's completely amazing how people who embrace analytics still haven't figured out that depth players who get soft minutes with massive offensive zone starts (a la Gagner and Wiercioch) pretty much always produce great-looking advanced stats, and that it means nothing. Look what happened to Cody Hodgson with Nashville.
Anyone who still thinks Wiercioch is a good player because of his 70% offensive zone starts next to Erik Karlsson 3-4 years ago needs to give their head a shake.
I wouldn't call Del Zotto an analytics star. That is just being disingenuous for the sake of setting up a straw man.
He had OK numbers and may have been undervalued, which meant he was worth getting a look at, that's all. Nobody expects him to be a star.
He was basically brought in to be a PP guy but lost time to Ghosthisbehere and later Provorov. Maybe he has a decent turn around with us, maybe not. Stop putting everything in terms of the extremes. It is very frustrating.
And wiercioch has had very poor numbers. It makes no sense to call him an analytics star either. His numbers stink horribly.
If anyone is calling him good based on their poor interpretation of the numbers, that is on them.
He had OK numbers and may have been undervalued, which meant he was worth getting a look at, that's all. Nobody expects him to be a star.
He was basically brought in to be a PP guy but lost time to Ghosthisbehere and later Provorov. Maybe he has a decent turn around with us, maybe not. Stop putting everything in terms of the extremes. It is very frustrating.
And wiercioch has had very poor numbers. It makes no sense to call him an analytics star either. His numbers stink horribly.
If anyone is calling him good based on their poor interpretation of the numbers, that is on them.
Would vote B- if possible, but just like with the draft poll rounding up because the past three years has been mostly rounding down to a zero. Doesn't mean I agree with what they did, but looking at July 1st in a vaccuum, it was right around average. No steals, no absolute disaster contracts.
But to repeat what I assume is the general consensus; signings are individually average but as a whole lead to an incoherent and staggered direction, which has been par for the course since Benning came on board.
To liken it to a university trajectory; it's like a kid who stumbled through their first two years trying to get into (insert prestigious professional program) school while doing a (insert useless major here) degree. They're looking at and talking about switching into a (insert useful trades major here) program... but then get a B in a first year (useless major) class that has essentially zero impact into them getting into law/medical/etc. school, and suddenly they feel the dream is back on, even with their cumulative 2.3 GPA.
But to repeat what I assume is the general consensus; signings are individually average but as a whole lead to an incoherent and staggered direction, which has been par for the course since Benning came on board.
To liken it to a university trajectory; it's like a kid who stumbled through their first two years trying to get into (insert prestigious professional program) school while doing a (insert useless major here) degree. They're looking at and talking about switching into a (insert useful trades major here) program... but then get a B in a first year (useless major) class that has essentially zero impact into them getting into law/medical/etc. school, and suddenly they feel the dream is back on, even with their cumulative 2.3 GPA.
Except that according to Sportsnet, those two guys were among the top five FA signings in terms of 'analytics'......so it's really their term more than mine.
Addison Rae
Registered User
context is by far the most important part of advanced stats, which is why for the most part i ****ing hate those hero charts.
if people think matthieau perrault is a better forward than sean couturier or jake gardiner is a better defender than chris tanev because of analytics they don't understand them at all.
if people think matthieau perrault is a better forward than sean couturier or jake gardiner is a better defender than chris tanev because of analytics they don't understand them at all.
And unfortunately, most people don't understand them at all.
All we've done is gone from people taking +/- out of context to make bad evaluations to people taking Corsi or Rel Corsi out of context to make bad evaluations.
The worst part is that it isn't just casual fans - seeing Canucks Army list Wiercioch as a 'great advanced stats pickup' is just incredibly disappointing, since those are the sorts of guys you'd hope would actually be able to use the information properly.
This isn't baseball. There is just so much noise and context to everything, and people trying to get a full picture out of these sorts of numbers are just missing the boat badly.
Wiercioch really is the mine canary for how badly this information can be misinterpreted without proper understanding or context. The guy is just a crap player. He's been crap for years. He looks like crap on the ice. His traditional advanced stats are crap. His advanced stats once you separate out the context and zone starts are crap. But holy smoke! For a couple years he played pretty often with the best defender in the NHL and was handed 60-70% offensive zone starts in 14 minutes/game because his coach didn't trust him and he put up 'excellent' advanced stats ... so he must be an underappreciated excellent player!
Same thing happened with Linden Vey here in his first year where people were arguing he was actually a solid defensive player because his advanced stats weren't too bad in a super-soft 13 minutes/game. No, he wasn't.
And stuff like this is where the dinosaur traditionalists ridiculing the 'stat nerds' are - sadly - 100% correct.
krutovsdonut
eeyore
- Sep 25, 2016
- 16,829
- 9,491
actually we are just "correct". we don't need a meaningless stat in front of that word to make us seem more than that.
in all seriousness, the dinosaurs welcome their eventual mammalian successors as you scurry around our feet, and we have a contingency plan to eventually shrink and grow feathers to get into the trees where you can't reach our eggs, but for now we feel confident you guys need a bit more time to develop the data and analytical foundation to evolve into successful competitors.
me2
Go ahead foot
He had some good numbers when he was sheltered and getting ozone heavy starts (40.8% ozone vs 25.0% Dzone).
Good depth pickup though. Can be 7/8/AHL. If OJ shines nobody will care if Weircoch gets lost to waivers.
No, the dinosaurs are even more wrong. And that's how we've had terrible moves based on a guy like Sbisa being 'good in scrums' or Prust being acquired to 'help Dorsett fight' or Gudbranson somehow being a top-4 defender because he's big and young and 'tough to play against' even though he's ridiculously easy to play against, or Sutter being a 'foundational player' and on and on and on. Just garbage acquisition after garbage acquisition based on false ingrained myths about winning from decades ago.
The nerds have it right - the most important thing in the NHL right now is to be a quality possession team, and that's the biggest, most consistent driver of overall success as an organization. And it is easy to see the bigger picture of who is a good possession team as a whole.
What is much harder - and where the nerds go wrong - is identifying which players are actually driving that success and why. There's just too much noise and too much context in the stats and anyone who thinks they can just look at numbers to figure out who is a good player is completely deluding themselves, and are going to make errors just as stupid as the dinosaurs with their 'hard to play against' BS.
The stats are a nice tool in some cases - IF you watch the team regularly, actually understand how the sport is played in 2017, and understand most of the underlying context that might be causing or skewing the numbers. Which almost nobody does.
rune74
Registered User
- Oct 10, 2008
- 9,228
- 552
There seems to be a ton of variables that could effect a number of stats for any given player. Possession for instance could be based on team philosophy, who he is playing with, team morale(standings for instance), player morale....and on and on.
The question I have is, with all these variables that can change won't they effect the overall value of the analytics in question? Is that something they take into consideration?
The question I have is, with all these variables that can change won't they effect the overall value of the analytics in question? Is that something they take into consideration?
I pretty much agree with this.
Understanding what players are going to help you score more or help you defend better should be what teams are looking for. I think the days of trying to cobble a team together based on perceived intangibles can get you in a whole lot of trouble, especially in a capped league.
I think hockey analytics, especially those that are publically available, are very much in their infancy. I think Corsi, in isolation, is not particularly useful - at least as a player acquisition tool. There is certainly some good information out there, and some of it is fascinating even if it might not be very useful. I think data and analytics are very much here to stay, and the best organizations are hard at work on proprietary data that can give them an advantage.
I do think there is probably still a place for "intangibles" or "chemistry", but those are things that are nearly impossible to quantify, and you can't really afford to allocate cap dollars to bad players that may be "good in the room".
As much as some people want to break the game down to 1s and 0s, there is still a human element at play. These are people, that can't or won't always behave in a predictable manner.
That was a bit of a ramble, but basically I feel like that analytics are incredibly usefully, especially if you understand they're also incredibly limited.
Yeah, I agree with this. Broadly, the things analytics looks at (even strength possession + controlled zone entries/exits, special teams shot generation/suppression) matter so much more to winning than old-school platitudes about foundationality and gritensity and scrums and whatnot. Also, the mere existence of PDO acknowledges how much random **** affects the outcome of games.
However there is this fringe expectations that favorable possession in soft-minutes will carry over into roles higher in the lineup (you see this in OwnThePuck's roster quality projections) that just doesn't happen. It's weird because it's not like analytics doesn't look at that - quality of competition and zone start percentages are things that are looked at, but there is this selective blindness that occurs sometimes.
