Whore Island 3.0

User avatar
cawbber
Registered Broad
Posts: 14869
Joined: Tue Nov 18, 2008 12:58 pm
Has given rep: 208 times
Received rep: 151 times

Post #251 by cawbber » Sat Apr 11, 2015 12:41 pm

Dynrehab wrote:What number of games does a player need to play to qualify for a pension?


Good question. That would be a good barometer.
User avatar
vf
Registered Broad
Posts: 16640
Joined: Fri Dec 11, 2009 11:31 pm
Has given rep: 402 times
Received rep: 231 times

Post #252 by vf » Sat Apr 11, 2015 12:50 pm

Dynrehab wrote:What number of games does a player need to play to qualify for a pension?


Goddamned phone. I posted this right after my last comment. Well, I thought I did anyways. Iirc the pension mark is 400 games. And ignore quality for now, as others have said that will be a ton of extra work.
User avatar
cawbber
Registered Broad
Posts: 14869
Joined: Tue Nov 18, 2008 12:58 pm
Has given rep: 208 times
Received rep: 151 times

Post #253 by cawbber » Sat Apr 11, 2015 3:22 pm

Players are eligible for NHL pensions at the 160 game mark. So that's the watermark.
User avatar
cawbber
Registered Broad
Posts: 14869
Joined: Tue Nov 18, 2008 12:58 pm
Has given rep: 208 times
Received rep: 151 times

Post #254 by cawbber » Sat Apr 11, 2015 3:24 pm

Blue wrote:similarly do you blame the scouts for JFJr/Bourke/Nonis removing 1sts and 2nds...and even when they try and replace one it's like 20 spots lower than your record should've had you picking?

The Toskala trade alone threw away a 1st--became Hans Eller (Shattenkirk next pick) 2nd, and a 4th.
We should still have Tuukka Rask, Alex Steen, Stralman, and of course Tyler Seguin/Hamilton instead of Kessel.

If we had Rask then Nonis doesn't burn the 2nd round pick on Bernier. That pick is this June and will be what 34th?

There should be a weighted value assigned to exactly how high a pick is in terms of what it should net. What the scouting department had each year in terms of picks and how high they were should be measured.

If that sort of thing is not done, then you wouldn't be judging the scouts per se, but the GMs


I completely agree. Which is why I prefer the NHL man games per # of picks available model.
User avatar
cawbber
Registered Broad
Posts: 14869
Joined: Tue Nov 18, 2008 12:58 pm
Has given rep: 208 times
Received rep: 151 times

Post #255 by cawbber » Fri Apr 17, 2015 12:02 pm

Thomas Malthus wrote:Cob, you might find this interesting for your project. I found it interesting in general: http://nhlnumbers.com/2015/4/11/the-draft-files-the-historical-cohort-based-approach-gets-its-sham-on

Again, I recommend reading all the previous work that they link in the article, it's pretty good.

And cob, if you are open to having me help on your draft project thing I'd love to pitch in whatever I can contribute.


you da man, Tommy.

Let me see how far I get, but I'm sure I will need a consult at some point.
Philthy Thrillz
Registered Broad
Posts: 26817
Joined: Sun Jul 10, 2011 1:05 pm
Location: KingsTown
Has given rep: 346 times
Received rep: 382 times

Re: Whore Island 3.0

Post #256 by Philthy Thrillz » Mon Jun 22, 2015 2:20 pm

Hey Tommy (or anyone else )

Could you post montreal players zone starts? Have a little wager going on...
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #257 by LeafOfBread » Mon Jun 22, 2015 2:43 pm

Philthy Thrillz wrote:Hey Tommy (or anyone else )

Could you post montreal players zone starts? Have a little wager going on...


Here you go PT

http://www.behindthenet.ca/nhl_statisti ... +46+63+67#
Philthy Thrillz
Registered Broad
Posts: 26817
Joined: Sun Jul 10, 2011 1:05 pm
Location: KingsTown
Has given rep: 346 times
Received rep: 382 times

Re: Whore Island 3.0

Post #258 by Philthy Thrillz » Mon Jun 22, 2015 4:38 pm

thanks lob. i'll bookmark the site. And I was write.
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #259 by Curry Rage » Thu Jun 25, 2015 1:36 am

Lemme guess, "Max Pac is a two way forward"?

Lulz
*Buy Oil Stocks!*
Dynrehab
Registered Broad
Posts: 2078
Joined: Mon Jun 08, 2015 5:05 pm
Has given rep: 97 times
Received rep: 200 times

Re: Whore Island 3.0

Post #260 by Dynrehab » Thu Jun 25, 2015 2:51 am

I noticed Phaneufs zone starts were a+4%.
Philthy Thrillz
Registered Broad
Posts: 26817
Joined: Sun Jul 10, 2011 1:05 pm
Location: KingsTown
Has given rep: 346 times
Received rep: 382 times

Re: Whore Island 3.0

Post #261 by Philthy Thrillz » Thu Jun 25, 2015 8:01 am

Courage 3.0 wrote:Lemme guess, "Max Pac is a two way forward"?

Lulz



You're good.
User avatar
Shawnathan Horcoff
Registered Broad
Posts: 37625
Joined: Sun Mar 11, 2007 7:51 am
Has given rep: 461 times
Received rep: 627 times

Re: Whore Island 3.0

Post #262 by Shawnathan Horcoff » Thu Jun 25, 2015 3:40 pm

Image
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #263 by LeafOfBread » Thu Jun 25, 2015 4:03 pm

Courage 3.0 wrote:Lemme guess, "Max Pac is a two way forward"?

Lulz

If playoff performance is indicative of anything, he's good at taking it at least two ways
User avatar
Thomas Malthus
Registered Broad
Posts: 4309
Joined: Mon Jun 08, 2015 5:08 pm
Location: Whore Island
Has given rep: 132 times
Received rep: 420 times

Re: Whore Island 3.0

Post #264 by Thomas Malthus » Sun Jul 12, 2015 9:49 am

"If you torture the data long enough, it will confess to anything." - Ronald Coase
"[...]all models are wrong, some are useful." - George E. P. Box
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #265 by LeafOfBread » Thu Jul 30, 2015 12:09 pm

Dude at boredz came up with some interesting work on quantifying teams' draft success. Leafs actually seem to do alright.

http://hfboards.hockeysfuture.com/showt ... ?t=1929165

We're the 7th round beasts, thanks Bergman. Our Euro picks also tend to be successful.
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #266 by Curry Rage » Thu Jul 30, 2015 12:17 pm

Haven't had time to really think about it, but my first thought is that this method rewards luck as much as skill. For example, when the top-ranked team (MTL) is hanging its hat on a 6th Rounder (Markov) and two 9th Round picks (Streit and Halak) aren't you into the realm of dumb luck? I'd bet even the GMs and Directors of Scouting that make those sorts of picks would agree.

SJ's success is at least based on some good 2nd Round picks in Vlasic and Carle.
*Buy Oil Stocks!*
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #267 by LeafOfBread » Thu Jul 30, 2015 12:20 pm

Courage 3.0 wrote:Haven't had time to really think about it, but my first thought is that this method rewards luck as much as skill. For example, when the top-ranked team (MTL) is hanging its hat on a 6th Rounder (Markov) and two 9th Round picks (Streit and Halak) aren't you into the realm of dumb luck? I'd bet even the GMs and Directors of Scouting that make those sorts of picks would agree.

Well luck will always be a factor no matter what, but that's just something that can't really be quantified
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #268 by Curry Rage » Thu Jul 30, 2015 12:32 pm

LeafOfBread wrote:
Courage 3.0 wrote:Haven't had time to really think about it, but my first thought is that this method rewards luck as much as skill. For example, when the top-ranked team (MTL) is hanging its hat on a 6th Rounder (Markov) and two 9th Round picks (Streit and Halak) aren't you into the realm of dumb luck? I'd bet even the GMs and Directors of Scouting that make those sorts of picks would agree.

Well luck will always be a factor no matter what, but that's just something that can't really be quantified


Hence why a metric like this is problematic.

It's great they've been the most "successful" (however that's defined) but unless you can come away with something that helps you repeat that success then it's just trivia. If something repeatable about picking Streit, Halak and Markov was measured then this thing is more useful.

Looking at the example of Toronto. I think it's nice that they've gotten some GP from their picks (and 2006 really helps) but if you're mostly getting 3rd/4th liners and 3rd-pairing D who manage to put up some GP numbers are you really being successful? Or is it possible that some way of weighting the various inputs (PPG, GP, etc.) is necessary to get a better feel for who was really nailing their drafts?
*Buy Oil Stocks!*
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #269 by Curry Rage » Thu Jul 30, 2015 12:47 pm

And thinking about it more, doesn't weighting all draft positions equally muck up the results? I mean, just getting to the NHL would be exceeding expectations for a 6th Round pick. For a high 1st round pick, exceeding expectations is going to be more difficult.
*Buy Oil Stocks!*
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #270 by LeafOfBread » Thu Jul 30, 2015 1:11 pm

Courage 3.0 wrote:
LeafOfBread wrote:
Courage 3.0 wrote:Haven't had time to really think about it, but my first thought is that this method rewards luck as much as skill. For example, when the top-ranked team (MTL) is hanging its hat on a 6th Rounder (Markov) and two 9th Round picks (Streit and Halak) aren't you into the realm of dumb luck? I'd bet even the GMs and Directors of Scouting that make those sorts of picks would agree.

Well luck will always be a factor no matter what, but that's just something that can't really be quantified


Hence why a metric like this is problematic.

It's great they've been the most "successful" (however that's defined) but unless you can come away with something that helps you repeat that success then it's just trivia. If something repeatable about picking Streit, Halak and Markov was measured then this thing is more useful.

Looking at the example of Toronto. I think it's nice that they've gotten some GP from their picks (and 2006 really helps) but if you're mostly getting 3rd/4th liners and 3rd-pairing D who manage to put up some GP numbers are you really being successful? Or is it possible that some way of weighting the various inputs (PPG, GP, etc.) is necessary to get a better feel for who was really nailing their drafts?


Just wondering but did you read the whole thing or did you just skim? Because he mentions that in his metric he didn't just use GP but a variety of stats such as PPG and TOI/G (for d-men), etc.
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #271 by Curry Rage » Thu Jul 30, 2015 1:18 pm

Skimmed. But your point (those would be the various inputs I mentioned) doesn't change my primary criticism regarding luck nor the fact that GP (and the various "quality" inputs) factored equally for all picks at all positions. Doing so favours later round picks when measured against "expected" results. Quite simply, a 1st Rounder is ALWAYS more likely to play an NHL game than a 7th Round pick. So, a team like the Leafs getting plenty of GP from players of all positions from later rounds (I know this simply because I know we haven't had many high picks) may be getting favoured for luck. Moreover, even using different "quality" measures for different positions doesn't account for this weighting problem (wouldn't a 1st Round forward be "expected" to have a higher PPG than a 9th Rounder?)

For a 3rd Overall pick to "exceed expectations" is going to be very difficult, but in the end it's the absolute production of your team - not their production relative to the expected results of their draft position - that dictates team success. I'd rather be "successful" in picking 2 top-3 guys who perform around or just a little above expectations than have 3 9th Round picks greatly exceed expectations but potentially still produce at a lower rate.
*Buy Oil Stocks!*
User avatar
cawbber
Registered Broad
Posts: 14869
Joined: Tue Nov 18, 2008 12:58 pm
Has given rep: 208 times
Received rep: 151 times

Re: Whore Island 3.0

Post #272 by cawbber » Thu Jul 30, 2015 9:15 pm

Courage 3.0 wrote:Haven't had time to really think about it, but my first thought is that this method rewards luck as much as skill. For example, when the top-ranked team (MTL) is hanging its hat on a 6th Rounder (Markov) and two 9th Round picks (Streit and Halak) aren't you into the realm of dumb luck? I'd bet even the GMs and Directors of Scouting that make those sorts of picks would agree.

SJ's success is at least based on some good 2nd Round picks in Vlasic and Carle.


I had an argument with a co-worker about this yesterday, actually. He believes that ranking team drafting prowess requires you to take the players impact in the NHL into account. I personally don't see how this is possible given all the potential variables, as any grading metric you come up will be arbitrary. It also generally favours offensive forwards over everyone else.

To me, games played should be the only measure for quantifying this sort of thing. It somewhat eliminates luck and other arbitrary measures.

TSN had any interesting take on this very thing, and they simply use games played as the defining metric. The twist is in the way they separate rounds 1-3 and 4-7. Slicing the data in this way shows teams that are making the best use of their late round selections. Its a simple but effective snapshot.

Image

http://www.tsn.ca/does-devils-draft-rec ... o-1.337262
User avatar
Shawnathan Horcoff
Registered Broad
Posts: 37625
Joined: Sun Mar 11, 2007 7:51 am
Has given rep: 461 times
Received rep: 627 times

Re: Whore Island 3.0

Post #273 by Shawnathan Horcoff » Thu Aug 06, 2015 11:07 am

You fellas will enjoy this one: Maple Leafs Traded the Wrong Big Contract Image
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #274 by LeafOfBread » Thu Aug 06, 2015 3:34 pm

edgar_dong wrote:You fellas will enjoy this one: Maple Leafs Traded the Wrong Big Contract Image


This was already discussed in the Differently Intelligenced People fred :crossarms:
User avatar
Shawnathan Horcoff
Registered Broad
Posts: 37625
Joined: Sun Mar 11, 2007 7:51 am
Has given rep: 461 times
Received rep: 627 times

Re: Whore Island 3.0

Post #275 by Shawnathan Horcoff » Thu Aug 06, 2015 3:55 pm

LeafOfBread wrote:This was already discussed in the Differently Intelligenced People fred :crossarms:


I figured you fellas hadn't missed it. :dion:

EDIT: fucking Big#D
User avatar
Thomas Malthus
Registered Broad
Posts: 4309
Joined: Mon Jun 08, 2015 5:08 pm
Location: Whore Island
Has given rep: 132 times
Received rep: 420 times

Re: Whore Island 3.0

Post #276 by Thomas Malthus » Fri Aug 07, 2015 10:24 am

edgar_dong wrote:EDIT: fucking Big#D


Make sure you wear protection.
"If you torture the data long enough, it will confess to anything." - Ronald Coase
"[...]all models are wrong, some are useful." - George E. P. Box
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #277 by Curry Rage » Mon Aug 10, 2015 6:09 am

A tiny drop of common sense in a sea of mathematically decorated dumb: http://hockeyanalysis.com/2015/08/09/sc ... on-sutter/

Score effects are a well known and well understood observation in hockey analytics. Essentially what score effects tell us is teams play differently depending on the score and in turn the resulting statistics are altered because of it. To keep this simple, in general teams that are leading give up more shots, but a smaller percentage of them end up as goals (they also take fewer shots but a higher percentage of them end up as goals).

Let me reiterate the main point here. When a team has a lead they effectively give up more shots but those shots are, on average, of lower quality and thus a lower percentage of them end up being goals. This effectively means when playing with a lead teams play in such a way that they boost their goalies save percentage.

This brings me to Brandon Sutter. Whenever I suggest that Brandon Sutter has an ability to boost his goalies save percentage there is always a backlash from a portion of the hockey analytics community or from those that believe in and use hockey analytics.


The basic point is something we've been saying here for ages: "systems" are made up of individuals doing certain things. Therefore, if the "system" has certain effects, there's a good chance that individual acts are causing those effects. Being able to stop other players scoring isn't voodoo. It's a bloody skill (or set of skills) that some players possess more than others.

A good portion of learning to play defense in hockey - as anyone who's ever tried knows - is learning how to ensure the guy you're checking ends up in a less dangerous shooting area and/or force them to take a shot they don't really want to take, alter the way in which they take the shot to render it less dangerous, or actually alter the shot itself by tipping it or blocking it. Of course this is going to show up somewhere in the counting of shots (blocked and otherwise), Save Percentage, etc. Christ on a crutch.

Johnson is right, this reminds me of "there's no such thing as shot quality", otherwise known as "if only Colton Orr would shoot more...." Whoopsy-daisy!
*Buy Oil Stocks!*
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #278 by LeafOfBread » Mon Aug 10, 2015 9:08 am

Courage 3.0 wrote:A tiny drop of common sense in a sea of mathematically decorated dumb: http://hockeyanalysis.com/2015/08/09/sc ... on-sutter/

Score effects are a well known and well understood observation in hockey analytics. Essentially what score effects tell us is teams play differently depending on the score and in turn the resulting statistics are altered because of it. To keep this simple, in general teams that are leading give up more shots, but a smaller percentage of them end up as goals (they also take fewer shots but a higher percentage of them end up as goals).

Let me reiterate the main point here. When a team has a lead they effectively give up more shots but those shots are, on average, of lower quality and thus a lower percentage of them end up being goals. This effectively means when playing with a lead teams play in such a way that they boost their goalies save percentage.

This brings me to Brandon Sutter. Whenever I suggest that Brandon Sutter has an ability to boost his goalies save percentage there is always a backlash from a portion of the hockey analytics community or from those that believe in and use hockey analytics.


The basic point is something we've been saying here for ages: "systems" are made up of individuals doing certain things. Therefore, if the "system" has certain effects, there's a good chance that individual acts are causing those effects. Being able to stop other players scoring isn't voodoo. It's a bloody skill (or set of skills) that some players possess more than others.

A good portion of learning to play defense in hockey - as anyone who's ever tried knows - is learning how to ensure the guy you're checking ends up in a less dangerous shooting area and/or force them to take a shot they don't really want to take, alter the way in which they take the shot to render it less dangerous, or actually alter the shot itself by tipping it or blocking it. Of course this is going to show up somewhere in the counting of shots (blocked and otherwise), Save Percentage, etc. Christ on a crutch.

Johnson is right, this reminds me of "there's no such thing as shot quality", otherwise known as "if only Colton Orr would shoot more...." Whoopsy-daisy!


Interesting article. Winnik being in the top 15 while Bozak being in the bottom 15 doesn't surprise me. I believe there was some information of a similar nature written not too long ago about how Dion Phaneuf elevates his goalies' sv%. It will be nice when shot quality will be able to be effectively tracked, hopefully not too long from now.
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #279 by Curry Rage » Thu Aug 13, 2015 4:18 pm

One of the better things I've read at PPP in a long while: http://www.pensionplanpuppets.com/2015/ ... -more-data

I've been saying this for a while, and I like the sound of my own voice so I'll repeat it here: The more data that teams have available to them, the more they'll need people who are good at theory, not numbers, to get any value out of the mountain of numbers that tracking technologies make available. Someone who can figure out what is actually worth looking at will be worth a lot more to a team than someone who can do a lot of calculations.


I've also been saying this for a very long while.
*Buy Oil Stocks!*
User avatar
tbm
Registered Broad
Posts: 1143
Joined: Tue Jun 09, 2015 1:02 pm
Location: in your neighbourhood
Has given rep: 56 times
Received rep: 63 times

Re: Whore Island 3.0

Post #280 by tbm » Tue Sep 22, 2015 10:38 am

The perch above the Toronto Maple Leafs practice rink was once a spot where executives could stand and watch what was going on below.

Not now.

When training camp shifted back home from Halifax, the balcony was occupied by a laptop and some high-tech equipment designed to give the organization an entirely different view of what is happening on the ice.

The Leafs are one of the NHL’s early adopters of the Catapult Sports tracking technology, which sees GPS units placed inside or below shoulder pads that transmit approximately 300 pieces of data per second — per player — in real time.

There are a whole host of different ways to use that mountain of information and the Leafs naturally declined to elaborate on what exactly they have planned. However, the mere fact they’ve invested in the pricey technology suggests that biomechanics is now an area in which the organization hopes to use its financial might to gain an edge.

Its arrival here follows the off-season hiring of Dr. Jeremy Bettle to fill the newly-created role of director of sports science and performance. When that move was announced in June, Leafs assistant GM Kyle Dubas said:

“Our sole objective is to be able to have our players know that as a program we are doing everything we can in all aspects to optimize their performance and maximize their potential as a hockey player, regardless of what stage of their career they’re at.”

The Philadelphia Flyers and Buffalo Sabres are the only other NHL teams known to use Catapult — Ben Peterson, the company’s sports performance manager, hinted in an interview with Sportsnet that there are a couple others — but GPS tracking is already commonplace in virtually every other major professional sport.

The biggest benefit it has in hockey, according to Peterson, is reducing the number of repetitive-use injuries by quantifying the
workload of each player throughout the season.

For example, when a skater’s force output starts to drop in one leg and not the other it’s usually a sign that some rest is required.

“A lot of times with players that eventually get a groin injury it starts out as just a very small shift or discrepancy; a very small little tweak of something that’s not normal,” said Peterson. “And if you can track athletes over time you can start to see those things change long before it manifests as an actual injury.”

The use of Catapult by NHL teams is currently restricted to practices because it hasn’t yet been approved by the NHL Players’ Association for games. However, Peterson said that GPS tracking is permitted in the American Hockey League and it’s believed that Toronto plans to have the AHL Marlies outfitted throughout the coming season.

In addition to establishing a log for the organization’s minor-league players, that would also provide an indicator of how much intensity and workload is required during a specific game situation like the penalty kill. That information could be useful in helping the Leafs coaching staff ensure it gets the most out of practice.

“Organizations have been really smart just to put monitors on guys in games in the AHL, which are running the same system as their big club,” said Peterson. “[They can] take that information to the big league club and they’ll compare it to when they’re running a penalty kill in practice and say: ‘Are we hitting the same workloads and creating the same type of real-type of environment that these guys are going to have to perform at during a game?”‘

One rival executive predicts that GPS tracking will eventually be used by every NHL team. He added that the only thing preventing his organization from jumping on board right now is the price.

Peterson, who made a presentation at last year’s MIT Sloan Sports Analytics Conference in Boston, doesn’t hesitate when asked why it’s a worthy investment.

“Catapult will admit that it’s by no means the cheapest thing in the world, but that’s because it’s a very complicated integrated device,” he said. “When you compare the price of a Catapult system to keeping your best player on the ice for an extra 10 games in a season, then it really becomes a small expense for what you’re getting for it.”

The potential for player tracking in hockey is only now being realized. Already well-rooted in ground-based sports like rugby, soccer, basketball and football, the Catapult system took some time to be adapted to the ice.

“This year really marks that first year where we really feel that we’ve kind of nailed it,” said Peterson.

For simplicity, teams have the ability to filter the millions of data points gathered during a practice into one printed sheet within 20 minutes. That could prompt coaches to change the following day’s schedule or alter an individual player’s training and recovery protocol.

However, as new applications for the raw data continue to be developed and competition among teams gets fiercer, we’re unlikely to hear too many specifics about how it is being used.

“We have a non-disclosure contract with all of our clients,” said Peterson. “We work with some organizations that really view all of this data as a competitive advantage, which in a lot of a sense it usually is.”

You can certainly put the Leafs in that category. That much we know for sure.
User avatar
LeafOfBread
Registered Broad
Posts: 7563
Joined: Wed Apr 16, 2014 12:37 am
Has given rep: 468 times
Received rep: 481 times

Re: Whore Island 3.0

Post #281 by LeafOfBread » Tue Sep 22, 2015 10:40 am

Flex that monstrous financial cock, took too long.
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #282 by Curry Rage » Tue Sep 22, 2015 10:42 am

Just coming in to link that article.

Video killed the Corsi star.
*Buy Oil Stocks!*
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #283 by Curry Rage » Wed Oct 07, 2015 8:40 am

What's that? Shots don't tell us anything about quality of possession or real possession time? Why has no one suggested this before?

http://www.thehockeynews.com/blog/can-d ... e-players/
*Buy Oil Stocks!*
User avatar
vf
Registered Broad
Posts: 16640
Joined: Fri Dec 11, 2009 11:31 pm
Has given rep: 402 times
Received rep: 231 times

Re: Whore Island 3.0

Post #284 by vf » Wed Oct 07, 2015 2:52 pm

Courage 3.0 wrote:What's that? Shots don't tell us anything about quality of possession or real possession time? Why has no one sugggested this before?

http://www.thehockeynews.com/blog/can-d ... e-players/


Whoa.

Image
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #285 by Curry Rage » Tue Oct 20, 2015 3:25 pm

Friedman:

Personally, a lot of my thinking on analytics is changing. I’m learning much of the stuff we know publicly is irrelevant, and there is much deeper work to be discovered.


Silly rabbit, Corsi's for kids!
*Buy Oil Stocks!*
User avatar
LeafErikson
Registered Broad
Posts: 30496
Joined: Tue Nov 18, 2008 6:46 pm
Location: Woodstock Onterrible
Has given rep: 81 times
Received rep: 137 times

Re: Whore Island 3.0

Post #286 by LeafErikson » Tue Nov 03, 2015 12:59 am

I've asked this before (I think) and I'm not sure I really got a great answer, but here I go again anyway.

Is there a metric that tracks positive and negative touches of the puck? By that I mean, play x has the puck, moves it out of the zone, and creates a scoring chance. This is measured as a positive touch. The next shift, same player turn the puck over due to a poor decision when being pressured by a fore-checker. I got this from a scout when he was talking about how he tracks players in games. He uses this method, rather than rely on stats sheets. And then I read someone else mention doing something similar. I guess my point is, with all the math magic that's out there, this very simple collection of data seems to be a much more accurate way of evaluating individuals.

Corsi for me works, but more as a long term predictor for teams, not players. For me, the whole area of study is convoluted, and full of guys simply trying to justify their existence. Simple is better.
User avatar
clawfirst
Registered Broad
Posts: 40009
Joined: Sat May 10, 2008 1:45 pm
Has given rep: 270 times
Received rep: 451 times

Re: Whore Island 3.0

Post #287 by clawfirst » Tue Nov 03, 2015 2:09 am

Zone exits,entries sorta does. But fuck if I trust who's counting.
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #288 by Curry Rage » Tue Nov 03, 2015 10:51 am

LeafErikson wrote:I've asked this before (I think) and I'm not sure I really got a great answer, but here I go again anyway.

Is there a metric that tracks positive and negative touches of the puck? By that I mean, play x has the puck, moves it out of the zone, and creates a scoring chance. This is measured as a positive touch. The next shift, same player turn the puck over due to a poor decision when being pressured by a fore-checker. I got this from a scout when he was talking about how he tracks players in games. He uses this method, rather than rely on stats sheets. And then I read someone else mention doing something similar. I guess my point is, with all the math magic that's out there, this very simple collection of data seems to be a much more accurate way of evaluating individuals.

Corsi for me works, but more as a long term predictor for teams, not players. For me, the whole area of study is convoluted, and full of guys simply trying to justify their existence. Simple is better.


I recall reading that DET used to keep their own internal stats on positive and negative touches. Each time a guy had the puck, they'd track whether the play he made ended in retained possession, another pass or whatever. Pretty sure Bowman got the idea from the Soviets under Tarasov, who used to track completed passes, if memory serves.

Anyway, I'm sure teams have those stats but most of the online guys are scraping data from elsewhere because you simply can't watch that many games. I've never seen anyone compile that sort of thing anywhere. Seem to recall even Mirtle (or someone like that) saying something earlier this year that the more he understands how teams actually track data, the more he realised the publicly available stuff was basically horseshit in comparison.
*Buy Oil Stocks!*
User avatar
MonkeyWrench
Registered Broad
Posts: 8333
Joined: Fri Jan 18, 2013 6:34 pm
Has given rep: 193 times
Received rep: 215 times

Re: Whore Island 3.0

Post #289 by MonkeyWrench » Fri Nov 06, 2015 11:13 pm

What do the fancy stats say about Harington up to this point?
User avatar
Thomas Malthus
Registered Broad
Posts: 4309
Joined: Mon Jun 08, 2015 5:08 pm
Location: Whore Island
Has given rep: 132 times
Received rep: 420 times

Re: Whore Island 3.0

Post #290 by Thomas Malthus » Sat Nov 14, 2015 6:36 pm

MonkeyWrench wrote:What do the fancy stats say about Harington up to this point?


  • High danger scoring chance for 54.39%
  • High danger scoring chance rel 3.51
  • scoring chance for 57.89%
  • Scoring chance rel 5.97
  • Corsi for 54.65%
  • Corsi fo rel 5.97
  • Zone start rel 11.36

Numbers tick down when adjusting for score effects. So basically, it appears that he's being successful but in easier minutes. Among our defenders he has the third best high danger scoring chance rel, second best corsi for rel, and best scoring chance for rel (these all score effects adjusted rankings).
"If you torture the data long enough, it will confess to anything." - Ronald Coase
"[...]all models are wrong, some are useful." - George E. P. Box
Curry Rage
Registered Broad
Posts: 4815
Joined: Tue Jun 09, 2015 3:32 am
Has given rep: 491 times
Received rep: 370 times

Re: Whore Island 3.0

Post #291 by Curry Rage » Tue Dec 15, 2015 6:07 am

Paging Comrade Tarasov to the red courtesy phone...Paging Comrade Tarasov to the red courtesy phone....

http://hockey-graphs.com/2015/12/14/tor ... trics-101/

As an aside, it's fun to see these guys say things like "this is a refined measure of possession based on our evidence last season that teams shot at a higher percentage from passing shots than non-passing shots."

Shot quality. Perhaps even "pass quality".
*Buy Oil Stocks!*
User avatar
Thomas Malthus
Registered Broad
Posts: 4309
Joined: Mon Jun 08, 2015 5:08 pm
Location: Whore Island
Has given rep: 132 times
Received rep: 420 times

Re: Whore Island 3.0

Post #292 by Thomas Malthus » Fri Jan 15, 2016 6:09 pm

Friend in my program is interested in hockey. We have a term paper to do. Decided to do it on the impact of coaches.

Initial stages hut likely the way to go is try and create an estimate for wins via players and use the regression residual to account for coaches contribution. Obviously there I'll he unobservables so the interpretation will he muddy but as of right now that's the approach. Direct modelling seems unlikely.

We dis cussed it over pints so I'm kind of tipsy.

Let me know of concerns that you have. It will help us troubleshoot and specify our model.
"If you torture the data long enough, it will confess to anything." - Ronald Coase
"[...]all models are wrong, some are useful." - George E. P. Box

Return to “The Auston Matthews show”

Who is online

Users browsing this forum: WTF and 4 guests