Aside from the loss of Metcalf's site potentially, with the Leafs hiring analytical types, do you think they're going to go in, review the past couple of seasons and go "eewwwwww"? One would think they might.
2017-18 stats and underlying metrics thread [Mod: updated season]
- Thread starter garret9
- Start date
- Status
Aside from the loss of Metcalf's site potentially, with the Leafs hiring analytical types, do you think they're going to go in, review the past couple of seasons and go "eewwwwww"? One would think they might.
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
One of the guys they hired (not Metcalf) was very critical of the Leafs. They also silenced one of their critics.
Nah... they just need to keep on ramping up their save percentage and shooting percentage and they'll be fine...
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
Nah... they just need to keep on ramping up their save percentage and shooting percentage and they'll be fine...
+/- all the way
garret9
AKA#VitoCorrelationi
That's my good buddy Cam Charron. Also known as the videographer of my ALS icebucketchallenge video.
garret9
AKA#VitoCorrelationi
I wrote on story on stats not helping teams:
http://morethan140charactershockey....rsi-didnt-help-tyler-dellow-and-kyle.html?m=1
garret9
AKA#VitoCorrelationi
I know Jets have a guy and I know that the Jets have not made any recent changes.
- Jun 10, 2014
- 57,393
- 29,208
IIRC I've seen posts where you speculated a bit about which approach to analytics the Jets might be using. I've also seen posts where (I'm pretty sure it was you) the existence of 'bad' statistical analysts was mentioned. Perhaps that some of them might give analytics a bad name. Do you know who is the Jets guy? What do you think about the possibility that the team might have access to better data than what is available to the amateurs and enthusiasts?
CaptainChef
Registered User
garret9
AKA#VitoCorrelationi
I believe that all teams have the potential grab better data than amateurs, but I believe few actually end up with better analysis. If I had to guess before this summer, maybe 5. Probably higher now with some of the best minds in blogging world being taken... This raising the NHL's talent and lower the bar.
garret9
AKA#VitoCorrelationi
I've never talked to the guy but I know someone who did. He never asked directly those types of questions but from what my friend told me it seemed like the guy knew the existence of Corsi numbers but wasn't really in depth knowledgable on what's going on with the analytical blogging world and some of the stuff being done with the data.
This is sort of the feeling I got from Cheveldayoff's interview from Tait on analytics.
There's pros and cons to all types of data currently available to both the pros and amateurs. The big convenience with Corsi data is two fold:
1) League wide data. This is huge. Without it you can't really do much in terms of contextual studies. We know Stuart and Thorburn have poor Corsi numbers, but without league context it would be near impossible to accurately determine whether their numbers are okay relative to depth in position and usage. We wouldn't know that Stuart's numbers are about average for bottom 1/6th of defenders (ie: he is an okay number six in that regards) and that Thorburn's numbers are terrible even for a 4th line player.
2) Open source studies. One thing we'll start to notice as more bloggers get picked up or go dark is the slow down of discoveries. The online community creates a large think tank where new ideas were constantly tested and re-evaluated. Of course for teams the opportunity cost is huge as other teams also benefit from this type of arrangement.
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
I wonder how much things changed in this area once Marice came onboard. Puck Daddy in his sportsnet interview said that many teams began using extraskater data during the season for the same reasons we did. It's ease of use.
Maurice mentioned numbers as a reason he put Frolik with LL in his first game as coach. What numbers did he have access to as a TSN analyst? Probably extraskater was involved somehow. It's possible, even likely that whatever he was using was introduced to the organization. Now that the Leafs have exclusive use of extraskater, other teams won't have access to that info.
The thing that doesn't get talked about much is what will the teams like the Leafs and the Oilers do with their analytics approach? Can they have an immediate impact on their teams record? Over time it can likely help them with their player acquisition decisions but can it help them at the draft table?
Mathmew Purrrr Oh
#meowmeowmeowmeow
if some team could find a way to track rts in other leagues and use them to create some kind of algorithm that can project a players potential likelihood of success, well that would be the holy grail imo
CaptainChef
Registered User
Reading much into what Chevy says is difficult, but I really got the feeling when I listened to that interview that they really don't pay too much more than lip service to the various advanced stats that are being collected. If I can recall he sort of alluded to the fact that they have their own advanced stats that they use which didn't give me much confidence that they really put a lot of faith in the kinds of advanced stats used by many teams.
Hope I'm wrong, but I get a feeling that the Jets are way behind the curve when it comes to understanding analytics. How else do we rectify them playing a goalie with a .901 save % down the stretch until they were out of the playoff picture when they had a hot goaltender or two in the wings. How else do we explain using Thorbs forever in a top 6 role when there was much better options out there.
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
garret9
AKA#VitoCorrelationi
Please support Corey if your interested in his stuff:
http://www.gofundme.com/allthreezones
He's my zone entry/exit source (why I was able to stop doing it myself manually).
For example, I used his data here in an article I wrote on new Jet Perreault:
http://www.arcticicehockey.com/2014...erreaults-strong-possession-numbers-come-from
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
The fact that the same eyes are evaluating all the games gives the data more credibility. There is a consistency that you can't achieve with multiple evaluators. That's an amazing amount of time spent watching hockey, but unsustainable year after year. He can't keep that up all by himself forever can he?
garret9
AKA#VitoCorrelationi
Means better chance of high precision but worse chance of accuracy:
In cases like this though, precision is more desired, as how a player/team does relative to their peers is more important than how exact the number is.
garret9
AKA#VitoCorrelationi
I like to keep you guys in the know ahead of time, but I've gotten some sweet data...
Jets players' Corsi first 10 seconds after particular situations (ex: DZ FO loss)
Jets players' Corsi first 10 seconds after particular situations (ex: DZ FO loss)
garret9
AKA#VitoCorrelationi
Early findings of these faceoff situational numbers are extremely intriguing... It's also spurred a lot of ideas for me.
I wish I had access to powerscout/sportsvu...
I wish I had access to powerscout/sportsvu...
truck
Registered User
Unholy goalie
Registered User
- Jul 11, 2011
- 712
- 127
How does this relate to the work done by David Johnson whose work purposefully ignores the first 10 seconds of a zone start for zone adjusted numbers? Specifically, what advantages are you seeing, and what might he be missing, in those 10 seconds? Is your data only zone starts, or are the "situations" more diverse?
In any event, that's pretty cool and I look forward to your analyses.
garret9
AKA#VitoCorrelationi
While I'm not sure that removes the influence of zone starts, it seems (looking at Dellow's work) to at least remove the influence of the faceoff being a win or a loss.
Joe Hallenback
Moderator
- Mar 4, 2005
- 15,386
- 21,561
- Status
