Joe Hallenback
Moderator
- Mar 4, 2005
- 15,389
- 21,582
Just a question then
Is Tangradi a better 4th line option then Halischuk by using just the numbers then?
Is Tangradi a better 4th line option then Halischuk by using just the numbers then?
2017-18 stats and underlying metrics thread [Mod: updated season]
- Thread starter garret9
- Start date
- Status
garret9
AKA#VitoCorrelationi
They were used in almost the exact same manner in terms of matchups and zone starts, although I believe Halischuk got better linemates (with icetime with Scheifele, Frolik, et al).
Tangradi had far better Corsi.
Halischuk had far better point production per icetime.
I'd say Tangradi is probably a more effective player in terms of two way play, especially forechecking, defensive zone positioning and zone entries (from my tracking), but Halischuk is the more effective offensive player and finisher.
Personally, my eyes say the same.
There is a part two to this series which this leads into: when does a players finishing talent make up for a poorer Corsi%, if ever...
EDIT: I should add that part two isn't written yet...
Last edited:
garret9
AKA#VitoCorrelationi
Great quote from Jon Willis re: eye-test vs @garrethohl: "In specific instances where eyeballs and analytics disagree, an intelligent man asks why the discrepancy exists and investigates further."
Yes: the player in the link I posted allows you to skip ahead - it's at the 36m mark.
garret9
AKA#VitoCorrelationi
I wrote something you may have seen on a website that starts with r and rhymes with edit.
Never Trust Anything Blindly
http://morethan140charactershockey....er-trust-anything-blindly-especially.html?m=1
Never Trust Anything Blindly
http://morethan140charactershockey....er-trust-anything-blindly-especially.html?m=1
garret9
AKA#VitoCorrelationi
I wrote something on two of the new hires this summer:
http://morethan140charactershockey.blogspot.ca/2014/08/corsi-didnt-help-tyler-dellow-and-kyle.html
It's me delving into a style of writing that I have never done before.
http://morethan140charactershockey.blogspot.ca/2014/08/corsi-didnt-help-tyler-dellow-and-kyle.html
It's me delving into a style of writing that I have never done before.
Joe Hallenback
Moderator
- Mar 4, 2005
- 15,389
- 21,582
garret9
AKA#VitoCorrelationi
Did I say they are equal somewhere?
Probably should read the post to see that it's a satire piece...
Joe Hallenback
Moderator
- Mar 4, 2005
- 15,389
- 21,582
I have advanced degrees and a long career in research with a major emphasis on statistics and other quantitative analytic approaches. So I think it makes a lot of sense to introduce this into the arena of management of hockey teams. However, I think considerable caution is required because statistical methods all have considerable limitations, particularly when many of the influential and confounding variables are not incorporated into analyses. Also, any good researcher knows that it is "garbage in, garbage out" when it comes to statistical analyses. I wonder how much attention is being paid to data collection, and its accuracy and reliability.
My opinion is that just having a "numbers guy" won't necessarily push a team in the right direction. I think that there is considerable room for growth in the field and in the expertise of those within it. Being good at arguing or quick with a pithy quote is not generally what I look for in a good analyst. Teams can benefit from having a very thoughtful and skilled analytics specialist who is willing to look at new ways of organizing data and analysis. Having someone who simply takes some of the existing rather crude statistics and tries to persuade "hockey guys" that they know better is a recipe for disappointment.
My opinion is that just having a "numbers guy" won't necessarily push a team in the right direction. I think that there is considerable room for growth in the field and in the expertise of those within it. Being good at arguing or quick with a pithy quote is not generally what I look for in a good analyst. Teams can benefit from having a very thoughtful and skilled analytics specialist who is willing to look at new ways of organizing data and analysis. Having someone who simply takes some of the existing rather crude statistics and tries to persuade "hockey guys" that they know better is a recipe for disappointment.
Joe Hallenback
Moderator
- Mar 4, 2005
- 15,389
- 21,582
Let me ask you a question Whileee, do you actually make policy decisions in your current job or do you just provide information to people who make the policy decisions?
garret9
AKA#VitoCorrelationi
They are two out of five management hires with an analytical twist.
Makes sense to me if you are discussing analytics. I get that AGM job is >>>> above that of analyst, scout, or consultant, but that's about it.
Dellow's (EDM hire specifically supporting Eakins) job will be different than MacDonald + Mehta (head of analytics for FLA + NJD), which is likely different than Tulsky (apparently all is a secret), and all are different than Dubas' job...
In addition they all come from different places.
* Dellow is a B.Com (Hons) in Sports Admin + BA PoliSci from Laurenten, and JD from UofT, who then turned into an amateur statistician blogger
* Tulsky is a double BA in Chem & Physics from Harvard, and PHD in Chem from UC-Berkeley, who -like Dellow- gained exposure through blogging
* Mehta studied music and was once a part of the Oilogosphere but only during the VERY underground era, who then moved on to becoming a pro-poker player and published a few books on statistics and poker
* MacDonald has a BS in Eletrical Engineering, with masters and PHD in mathematics, is a professor at West Point University, and wrote some articles for HockeyProspectus and his own website Greater than Plus/Minus
* Dubas worked his way up through Soo organization (starting with equipment or stick boy I believe as a teenager), got his BA(hons) in Sports Management at Brock University, and also I believe was at one point the youngest NHLPA certified player agent ever. When Dubas realized he couldn't compete with different teams in the OHL when he became GM of the Soo so he needed to investigate ways into taking advantage of inefficiencies. As he says himself "the Oakland A's way". This lead him to the stats blogosphere where he became a part of the community, asking questions, discussing on twitter, and promoting our stuff. (PS the guy is super smart guy to talk to)
All 4 are taking different roles/positions and all 4 took different paths getting there. The only similarity to three of them is that they wrote stuff on the internet.
Sorry Joe, I know a lot of this is ranting and not really response to your comment, but once I started I thought it be good to have this info here for people to see in case they are interested.
Last edited:
Joe Hallenback
Moderator
- Mar 4, 2005
- 15,389
- 21,582
garret9
AKA#VitoCorrelationi
Dellow played high level hockey. Other ones though did not. I'm not even sure if Tulsky has ever played at all ha.
Still, that's pretty insignificant to me as long as you are not the stat type that doesn't understand the game and looks at it purely in a numbers perspective.
Dellow, Tulsky, Mehta, and Dubas are all not like that.
MacDonald I don't really know well enough to judge on that.
Pure number types who don't know hockey I'd never take advice from, but would only use to crunch numbers. Tulsky and Dellow have watched the game and learned from their observations just as much as the traditional types.
EDIT:
Way I look at it is that there are two stats types. We tend to split them as "academic stat types" and "hockey stat types". "Academics" would be guys I'd separate from Dubas in that sense; they take big data, run regressions, and figure out from that. Hockey stat types are more that look at the same things coaches value and put numerical values to it.
I make management and policy decisions, usually based on extensive analysis. I've been in advisory roles earlier in my career.
I'm pretty sure that the most useful quantitative analysts will turn out to be the ones that really understand the questions that need answering from coaches and GMs, and have the skills and creativity to develop appropriate analytic approaches, which would include helping to define what data are needed.
- Mar 10, 2010
- 34,897
- 31,344
Really nice twist to the thread boys. I like the application theories.
I use to run a single business and some data collection was very nessesary to understand cost controls and sales forecasting but I would make the bulk of the decisions intuitively because I was present in the operation daily basis. As our company grew in size and geographical reach we had to adapt because we couldn't be in two places at one time. Our ability to gather the field level, moment by moment, intellagence and information needed to be augmented by increased systimatization and more statistical comparison. Although I didn't love it I realized it was the only way we could scale our operations and grow. Although I am entrepreneurial in nature I have always had a passion for refining process through trial and error.
If I was running a hockey operation I would definatly invest in an analytics department. When you are all given the same allowance to spend and you are competing for the same shallow talent pool every edge is nessesary. From a defensive perspective you would need it to keep up. But more clearly with things like SportVu on the horizon this will become a critical element for talent evaluation in the future. I believe the old eye test and talent evaluation can not be replaced but I think the game may start to go through a more rapid change by people who adapt quickest to new information flow combined with the old methods. Teams (and businesses) usually adapt to best practice results from competitors but those who innovate well or adapt fastest will have a window of opprotunity edge.
Where this could have the craziest opprotunity is when SportVu becomes available in junior, NCAA, buildings and you can use it for scouting and a assign prospects.
Hey I am not a sports owner so what do I really know but my business insticts are screaming at me from following this as an ameture that I would not want to be the last guy standing looking for a chair to sit in when the music stops when it comes to hockey analytics collection and application.
I use to run a single business and some data collection was very nessesary to understand cost controls and sales forecasting but I would make the bulk of the decisions intuitively because I was present in the operation daily basis. As our company grew in size and geographical reach we had to adapt because we couldn't be in two places at one time. Our ability to gather the field level, moment by moment, intellagence and information needed to be augmented by increased systimatization and more statistical comparison. Although I didn't love it I realized it was the only way we could scale our operations and grow. Although I am entrepreneurial in nature I have always had a passion for refining process through trial and error.
If I was running a hockey operation I would definatly invest in an analytics department. When you are all given the same allowance to spend and you are competing for the same shallow talent pool every edge is nessesary. From a defensive perspective you would need it to keep up. But more clearly with things like SportVu on the horizon this will become a critical element for talent evaluation in the future. I believe the old eye test and talent evaluation can not be replaced but I think the game may start to go through a more rapid change by people who adapt quickest to new information flow combined with the old methods. Teams (and businesses) usually adapt to best practice results from competitors but those who innovate well or adapt fastest will have a window of opprotunity edge.
Where this could have the craziest opprotunity is when SportVu becomes available in junior, NCAA, buildings and you can use it for scouting and a assign prospects.
Hey I am not a sports owner so what do I really know but my business insticts are screaming at me from following this as an ameture that I would not want to be the last guy standing looking for a chair to sit in when the music stops when it comes to hockey analytics collection and application.
garret9
AKA#VitoCorrelationi
Can you go into more detail as to why Halischuk may be a better choice than Tangradi on the 4th? When you look at some of the analytics, Tangradi's are some of the best on the team.
Thanks
He seems like the logical choice to eat up minutes on the 4th line.
Thanks
He seems like the logical choice to eat up minutes on the 4th line.
Aavco Cup
"I can make you cry in this room"
- Sep 5, 2013
- 37,630
- 10,440
And a Kossy mention to boot. I wonder if the NHL shouldn't evaluate more than 1 system to see what these different companies can come up with after a whole season.
Sport Vu might be the best but maybe it's not.
- Jun 10, 2014
- 57,394
- 29,212
- Mar 10, 2010
- 34,897
- 31,344
- Status
