I think a lot of teams had stat guru people before, it just wasn't as advertised as much as it is right now. This isn't a mickey mouse operation so if a team can find a way to leverage something over others they will and they sure as hell wouldn't advertise to the rest of the teams. Sports in general is a copy cat league so if someone finds something that works others will surely give it a go.
I think it should be mentioned that this could be a watershed year for the advanced stats guys. For the first time in the history of the nhl, all games will be monitored with player tracking technology. One of the main failings of things like Corsi is that it's what statisticians call a "noisy" stat. Teammates, opponents and zone starts all greatly affect a players Corsi, but even the better statisticians out there don't even try to decipher things like a players average shot quality. Even though it's obvious to anyone that watches hockey that some chances are better than others, Corsi blindly ranks a one timer from the slot and a lazy shot into the pads as equal quality events.
Player tracking should give statisticians access to immense amounts of data for the first time in history. Instead of being limited to just tracking shots they can start tracking things like "what percentage of the time does a defenceman successfully touch the puck and clear it into the neutral zone." There is an opportunity this year for massive leaps in the statistical assessment of hockey players. If someone rises to the task, what Bill James did for baseball someone else could do for hockey this season.
I, for one, am excited.
__________________
Always Earned, Never Given
Last edited by TheDebaser; 08-25-2014 at 04:59 PM.
The Following User Says Thank You to TheDebaser For This Useful Post:
Is there a website where some advance stats guy has gone over previous era's to show how players from that era rank under advance statistics? It would be interesting to see if Steve Konroyd was the best dman under certain advance stats on the 84/85 team or whether it was Paul Reinhart etc. It would be interesting to see if there if anyone has ever done that type of analysis.
Wouldn't be surprised if this entire "hire an analytics guy" thing is a passing fad that will eventually phase out once "advanced" stats are all readily and easily available to everyone and tracked on nhl.com
Once all these new stats are universally tracked GMs and execs will begin to interpret and understand the information themselves and use it to make decisions, rendering these "analytics" guys superfluous, unless they can stay ahead of the game by coming up with new categories to track.
This is just a thing right now because stuff like zone entries has to be tracked manually, so the guys that track them are a step ahead and possess information that others don't have. Once everyone has that info then what?
Eventually what is considered "advanced" stats right now will end up being "normal" stats as people become more familiar with them. It will be interesting to see what the stats experts do when their 15 minutes of fame are over and they will have to evolve in order to continue drawing a salary from a NHL team.
__________________
Tyger! Tyger! burning bright
In the forests of the night,
What immortal hand or eye
Could frame thy fearful symmetry?
I kind of feel like I've missed my calling. I should have immersed myself in hockey stats and analysis over the last 15 years. Time to reevaluate my life.
As I see it now advanced stats should be used to augment decisions, not make them. Also I think some of the people who use them rely on one or two solely and don't look at all the details, even by their own standards and stats. IE they use them when it benefits their case, not fairly across the board, or fairly with the player they are pumping up.
How is that any worse then some people who just watch the game?
At least if someone is supporting their argument with stats then it is possible to use those same stats more fairly in debunking a biased argument.
If a person that just goes by watching the game judges a player a certain way and it is disputed then it always just devolves into he said, she said.
Sometimes I will read different scouting reports on a prospect and the reports will read like it is 3 different players. A lot of times, 3 independent scouts can't agree if a prospect is a good skater or not so clearly just watching them play is fraught with all kinds of pitfalls as well.
I think if you take two hockey managers that are identical in terms of hockey knowledge and give one manager access to every stat that exists and the other just gets goals and assists the first one will outperform the second nearly every time.
If that statement seems reasonable then the only logical conclusion is to embrace the stats as much as possible.
Wouldn't be surprised if this entire "hire an analytics guy" thing is a passing fad that will eventually phase out once "advanced" stats are all readily and easily available to everyone and tracked on nhl.com
Once all these new stats are universally tracked GMs and execs will begin to interpret and understand the information themselves and use it to make decisions, rendering these "analytics" guys superfluous, unless they can stay ahead of the game by coming up with new categories to track.
This is just a thing right now because stuff like zone entries has to be tracked manually, so the guys that track them are a step ahead and possess information that others don't have. Once everyone has that info then what?
Eventually what is considered "advanced" stats right now will end up being "normal" stats as people become more familiar with them. It will be interesting to see what the stats experts do when their 15 minutes of fame are over and they will have to evolve in order to continue drawing a salary from a NHL team.
Agreed . If there is someone that is given a job to do analytics their first job would be to decide what is meaningful statistics and then gather them.
Right now the quality of the public domain (NHL.com) statistics is very poor. The hits and blocked shots and Takeway giveaway stats are very suspect.
CORSI and the possession stats are based on very shakey shots directed at goal and blocked shots stats and whether you are on the ice when this event happens.
There are shots that are directed at the net that are scoring chances and there are shots directed at the net that are dump ins and giving up possession of the puck ... these are counted the same in the general NHL stats.
Even totally taken for grants stats like assists are qualitative as well as quantitative.
Does banging of the boards in your own end that due to another players excellent play and leads to a goal merit the same positive stat as being part of a three way passing play that results in a goal ???
the individual teams that are spending $$$ on analytics should have considerable effort in breaking down game film and extracting meaningful statistics like:
Actual individual time with puck possession.
completed passes.. bad passes
Individual scoring chances.
Quality shots on goal and shots that result in easy saves.
possession changing hits.
Punishing hits
important shot blocks and shot blocks of easy goalie saves.
Entry to the offensive zone: Going off side, carrying the puck in, passing the puck in dumping the puck in and the scoring chances that result from each of these events.
Defensemen standing up at the blue line or backing in, gone around for a scoring chance
Forwards getting back in time or late
picking out the correct player to check.
Teams that create this data would not be sharing it. I do not see how advanced stats would ever become public domain.
I do not see how advanced stats would ever become public domain.
It's going to become public domain the instant SportsNet launches their "Individual Skater" app, where you can track a single player the whole time they are on the ice using whichever kind of technology the NHL decides to outfit their rinks in.
This will absolutely happen during the current broadcasting rights contract. The technology currently exists and is being installed and there's no friggin' way the NHL is going to be able to keep that kind of access out of SportsNet's hands.
The fans are going to demand it, every other team sport is trending that direction in their broadcasts (If they can tell me how far Messi runs in a game, why can't I know how far Crosby skates?), and SportsNet just paid five billion dollars for the broadcast rights.
It's coming, and it's coming FAST.
As soon as that technology is publicly available, you, and me, and every other nerd who's way, way better at math than I am, are going to start pouring through the information and you're going to see, in a very short time, a massive amount of data generated and analyzed.
Right now there is a brief window for individual teams to get ahead of the curve. They can start gathering and mining their own data, which will provide them with an advantage, even if that advantage is to simply figure out which data to ignore.
I hope to hell the Flames have been on this for a couple of years.
The Following 3 Users Say Thank You to driveway For This Useful Post:
This afternoon on @Sportsnet960 Chris Snow of the #Flames joins us to talk about the "Summer of Analytics" also...
__________________
The Quest stands upon the edge of a knife. Stray but a little, and it will fail, to the ruin of all. Yet hope remains while the Company is true. Go Flames Go!
Agreed . If there is someone that is given a job to do analytics their first job would be to decide what is meaningful statistics and then gather them.
Right now the quality of the public domain (NHL.com) statistics is very poor. The hits and blocked shots and Takeway giveaway stats are very suspect.
CORSI and the possession stats are based on very shakey shots directed at goal and blocked shots stats and whether you are on the ice when this event happens.
There are shots that are directed at the net that are scoring chances and there are shots directed at the net that are dump ins and giving up possession of the puck ... these are counted the same in the general NHL stats.
Even totally taken for grants stats like assists are qualitative as well as quantitative.
Does banging of the boards in your own end that due to another players excellent play and leads to a goal merit the same positive stat as being part of a three way passing play that results in a goal ???
the individual teams that are spending $$$ on analytics should have considerable effort in breaking down game film and extracting meaningful statistics like:
Actual individual time with puck possession.
completed passes.. bad passes
Individual scoring chances.
Quality shots on goal and shots that result in easy saves.
possession changing hits.
Punishing hits
important shot blocks and shot blocks of easy goalie saves.
Entry to the offensive zone: Going off side, carrying the puck in, passing the puck in dumping the puck in and the scoring chances that result from each of these events.
Defensemen standing up at the blue line or backing in, gone around for a scoring chance
Forwards getting back in time or late
picking out the correct player to check.
Teams that create this data would not be sharing it. I do not see how advanced stats would ever become public domain.
thats what SportVU will be able to do
most Basketball arenas already have it and the league is adding it or something like it to every other arena within 2 years I believe
good piece on it from a basketball perspective last year on Grantland, many of the same things it finds will be applicable to Hockey
I wonder what these sorts of analytics will do to player values. I have a sense that a lot of veteran players are overrated not only by fans, but by managers. Once advanced analytics are incorporated into asset assessment, will the salary structure flatten? If they factor more into salary negotiations, will players start gearing their game towards hitting hitherto obscure targets?
__________________
Quote:
Originally Posted by fotze
If this day gets you riled up, you obviously aren't numb to the disappointment yet to be a real fan.
“@mirtle: Lot of rumours the Flames have been looking to make an analytics hire. Or already have. Secrecy as usual.”
__________________
The Quest stands upon the edge of a knife. Stray but a little, and it will fail, to the ruin of all. Yet hope remains while the Company is true. Go Flames Go!
^^^^ Maybe the analytics guy didn't battle hard enough to beat out Setoguchi either. Maybe he was dispatched to Addy this morning along with Baertschi, Jooris and Ortio.
The Following 2 Users Say Thank You to Lanny_McDonald For This Useful Post:
“We are all looking for an edge,” said Chris Snow, the Flames director of video and statistical analysis. “How do we scout better? How do we negotiate better contracts? How do we develop players better? Analytics is just another area where teams can gain a bit of an edge on each other. Teams are looking for good original thinkers and the Internet has provided that this summer.”
Quote:
Snow tries to provide as much context as possible of a player — who he played with, what situations, time on ice, plus/minus, how they were used — and then use it to tell a story or narrative on a player’s career or recent season.
“So we can at least have a better discussion as a starting point,” Snow said. “And then start to look at what parts of his performance are repeatable. What can we predict or feel confident of a player improving, being the same, or not likely.
“It’s not a perfect science . . . but it’s taking information and turn it into recommendations that ultimately helps or doesn’t help.”
Quote:
When general manager Brad Treliving came on scene this spring, he quickly started the discussion of their approach to analytics.
“As far as us, I’m a big believer in that you don’t leave any stone unturned,” he said. “I wouldn’t say that ours is a reaction to it. I would tell you that Calgary had been using a form of analytics in the past. I would say that we’ve advanced that, in terms of the type of things that we’re doing, the type of commitment that we’ve made to it, the type of resources we’ve dedicated to it. So I feel very comfortable that we’re very progressive in that area.
“But it’s like Oz — you don’t want to talk about what’s behind the curtain.”
Quote:
But as for their methods?
Very hush-hush.
“It’s a very competitive league, and if you have something you feel is of value and gives you a competitive edge or some type competitive advantage, you don’t necessarily want to talk a whole lot about it,” Treliving said. “I would say that there’s been a form of advanced stats or analysis or analytics or statistical data — or however you want to refer to it — in a lot of shapes or forms ... teams have been doing for longer than (summer) when it’s become more prevalent, more talked about in the media.”
In other words, the Flames aren’t just responding to the summer’s trend.
“It’s no different than a lot of things,” Treliving said. “Like 20 years ago it was odd for people to have strength and conditioning coaches. So as our game continues to evolve, you’re always looking for more information. There’s never too much information. What you use from that, what you pull and decipher from it ... all sorts of things that can give you a different way to view and analyze your team, the league, your players, other players, trends, I think is very valuable.”
__________________
The Quest stands upon the edge of a knife. Stray but a little, and it will fail, to the ruin of all. Yet hope remains while the Company is true. Go Flames Go!
Chris Snow was I think. Pretty sure they adjust SV% based on where shots come from. Pretty sure they judge scoring chances based on where shots come from and whether it was a rebound (rebounds have a very high % chance of scoring vs 1st shots.) Not sure what else they do and they probably wouldn't tell you.
