Ryan Wilson: A Complete Review of the RIT Hockey Analytics Conference A Complete Review of the RIT Hockey Analytics Conference

first!!!!!!!!!!!!!!

it seems like Ryan realised his blogs had to little stats in his last couple of blogs

Nobody claims Corsi to be the god particle of hockey analysis.

Actually some do....they even write here on HB

Ryan Wilson: A Complete Review of the RIT Hockey Analytics Conference
A Complete Review of the RIT Hockey Analytics Conference
- Ryan_Wilson

Good (frank)ing God, this is actually a thing? A whole (frank)ing conference???

Meteorologists still struggle to predict the weather accurately more than about 24 hours in advance and they have been studying the atmosphere since at least the 1400's.

I don't think we need to worry about Hockey Analytics Experts being able to predict outcomes much better than they can today - anytime soon at least.

Good (frank)ing God, this is actually a thing? A whole (frank)ing conference???
- jmatchett383

Was it sanctioned by the NHL or any USA hockey or any official hockey program?

These stats are useful for picking fantasy teams. Id rather watch other people ice fish on TV.

Haven't had a chance to digest the entire blog yet but a couple things that stand out to me early.

At the present moment we do a better job at saying what happened, not what is going to happen.

There's a shock and here's a hint, that isn't going to change. There are entirely too many intangible variables that go into goaltending that no one will ever be able to effectively predict goaltending performance. The mental aspect alone that goes into playing the position will always be impossible to predict.

A shot assist is any pass that leads directly to a shot attempt.

Another absurd attempt to view all hockey plays as equal in a vacuum. Not all passes are equal and definitely not all shots are equal so why on worth should every pass that leads to a shot be viewed as equal? I was a big proponent of advanced analytics when they 1st started becoming relevant and I still like there value as a piece of the puzzle that gives some clarity to what's going on but it's really starting to feel like a theoretical physics where the people doing the research really don't understand what's going on so they create realms of guessing to explain what they don't understand.

Meteorologists still struggle to predict the weather accurately more than about 24 hours in advance and they have been studying the atmosphere since at least the 1400's.

I don't think we need to worry about Hockey Analytics Experts being able to predict outcomes much better than they can today - anytime soon at least.
- icedog97

No reason to even play the games, just calculate the winner.

Edit: From Pete and Pete: Inspector 34

"All the while, 34’s been inspiring those around him—the Wrigleys, mainly, as well as the neighborhood kids—to seek perfection in their own lives. Big Pete has a falling out with Ellen. The kids quit playing Kick The Can, instead choosing to calculate the probabilities of who would win each game."

This show was meant to be a comedy about people who do over-the-top ridiculous things. However, this is exactly the kind of thing that advanced analytics leads to.

"Although the Penguins lost 8-1, they possession metrics show that they were clearly the better team last night."

No reason to even play the games, just calculate the winner.

Edit: From Pete and Pete: Inspector 34

"All the while, 34’s been inspiring those around him—the Wrigleys, mainly, as well as the neighborhood kids—to seek perfection in their own lives. Big Pete has a falling out with Ellen. The kids quit playing Kick The Can, instead choosing to calculate the probabilities of who would win each game."

This show was meant to be a comedy about people who do over-the-top ridiculous things. However, this is exactly the kind of thing that advanced analytics leads to.

"Although the Penguins lost 8-1, they possession metrics show that they were clearly the better team last night."
- jmatchett383

- j.boyd919

ok

No reason to even play the games, just calculate the winner.

Edit: From Pete and Pete: Inspector 34

"All the while, 34’s been inspiring those around him—the Wrigleys, mainly, as well as the neighborhood kids—to seek perfection in their own lives. Big Pete has a falling out with Ellen. The kids quit playing Kick The Can, instead choosing to calculate the probabilities of who would win each game."

This show was meant to be a comedy about people who do over-the-top ridiculous things. However, this is exactly the kind of thing that advanced analytics leads to.

"Although the Penguins lost 8-1, they possession metrics show that they were clearly the better team last night."
- jmatchett383

I think we're blowing advanced stats out of proportion. They're here to mitigate the eye test which is inherently subjective and to make sure that small sample sizes aren't taken to heart. It is entirely possible that the better team can lose a game 8-1 since there is so much chance at work in a game like hockey. Analytics aren't here to replace the game. They're here to help make informed decisions. Pretty much everyone that subscribes to analytics besides James Tanner believes this. The violent reaction from the anti-analytics crowed has always perplexed me.

I think the biggest issue people have with advanced stats is that they aren't perfect. That's impossible. They don't have perfect predictability power and because of that people want to throw the baby out with the bath water. The form of analytics we have today are, in general, tools that help people create models that help them understand what players and teams will have the most long run success. They're not going to tell you what team is going to win on any given night or who will over or under achieve in any given small sample size. If you want something that perfect you're simply just not going to find it.

I think we're blowing advanced stats out of proportion. They're here to mitigate the eye test which is inherently subjective and to make sure that small sample sizes aren't taken to heart. It is entirely possible that the better team can lose a game 8-1 since there is so much chance at work in a game like hockey. Analytics aren't here to replace the game. They're here to help make informed decisions. Pretty much everyone that subscribes to analytics besides James Tanner believes this. The violent reaction from the anti-analytics crowed has always perplexed me.

I think the biggest issue people have with advanced stats is that they aren't perfect. That's impossible. They don't have perfect predictability power and because of that people want to throw the baby out with the bath water. The form of analytics we have today are, in general, tools that help people create models that help them understand what players and teams will have the most long run success. They're not going to tell you what team is going to win on any given night or who will over or under achieve in any given small sample size. If you want something that perfect you're simply just not going to find it.
- Victoro311

My issue with them is that they don't tell you anything about what will happen from one game/period/shift to another. They're a trend, a prediction of the future based on the past in a vacuum. Also, their major short-coming, to me, is that they fail to realize that every shot and every pass, no matter what kind of category you want to put it in, is unique. There are variables that can't be accounted for (fatigue level, pace of play, ice conditions, placement of other players, game situation, stick fatigue, etc.). They can be used as a general guide to what you can expect over the long term if things stay the same, but as a predictive tool, they're below adequate, to me.

BORING!

BORING!
- Dcoms

You're a boring.

My issue with them is that they don't tell you anything about what will happen from one game/period/shift to another. They're a trend, a prediction of the future based on the past in a vacuum. Also, their major short-coming, to me, is that they fail to realize that every shot and every pass, no matter what kind of category you want to put it in, is unique. There are variables that can't be accounted for (fatigue level, pace of play, ice conditions, placement of other players, game situation, stick fatigue, etc.). They can be used as a general guide to what you can expect over the long term if things stay the same, but as a predictive tool, they're below adequate, to me.
- jmatchett383

I mean, that's certainly a valid opinion but the way I see it is what other predictive tools do we have at or disposal in regards to sports in general and hockey specifically? Looking at it relatively I do believe analytics is the best we have. You're more than welcome to come into every game cold turkey saying "today is a new day and the past has no bearing on it" and take that sort of aproach if that's what floats your boat. At least to me, I find it fun to analize the trends when hockey isn't physically on and try to make inferences about what will probably happen in the future. My eye test is too subjective and I don't have the time to watch every single NHL game let alone digest what I think of every single player to be able to do it that way.

I hope they atleast have free food and booze at this thing.

I love how Hockeybuzz boards have devolved into Statistics denial.

This is really starting to feel like a cult.

Thanks for the write up Ryan Wilson.

I hope they atleast have free food and booze at this thing.
- Grinder47

Seems like the type of crowd that only drinks IPA's they brewed at home because all other craft brews are too corporate now.

I mean, that's certainly a valid opinion but the way I see it is what other predictive tools do we have at or disposal in regards to sports in general and hockey specifically? Looking at it relatively I do believe analytics is the best we have. You're more than welcome to come into every game cold turkey saying "today is a new day and the past has no bearing on it" and take that sort of aproach if that's what floats your boat. At least to me, I find it fun to analize the trends when hockey isn't physically on and try to make inferences about what will probably happen in the future. My eye test is too subjective and I don't have the time to watch every single NHL game let alone digest what I think of every single player to be able to do it that way.
- Victoro311

The test for analytics if you gamble against an "old school" eye test guy and have him play against an analytics guy will the guy using analytics win a significant amount of money.

I'm taking the side of the guy using analytics. And the daily fantasy games seem to have proven that out as guys using black box algos are consistently profitable. Pretty much to the extent that they ruined the product.

The test for analytics if you gamble against an "old school" eye test guy and have him play against an analytics guy will the guy using analytics win a significant amount of money.

I'm taking the side of the guy using analytics. And the daily fantasy games seem to have proven that out as guys using black box algos are consistently profitable. Pretty much to the extent that they ruined the product.
- sditulli

Don't get me started on the (frank) faces that make algorithms for daily fantasy games. It ruined what would other wise have been a boatload of fun. Now it's a sure fire bet that you'll lose money unless you have one of those algorithms and I just don't have the brain power for that. Algo users should be banned. It's the same concept of counting cards but on steroids. If counting cards is taboo in gambling algorithms should be straight up illegal.

I mean, that's certainly a valid opinion but the way I see it is what other predictive tools do we have at or disposal in regards to sports in general and hockey specifically? Looking at it relatively I do believe analytics is the best we have. You're more than welcome to come into every game cold turkey saying "today is a new day and the past has no bearing on it" and take that sort of aproach if that's what floats your boat. At least to me, I find it fun to analize the trends when hockey isn't physically on and try to make inferences about what will probably happen in the future. My eye test is too subjective and I don't have the time to watch every single NHL game let alone digest what I think of every single player to be able to do it that way.
- Victoro311

No, you can't go cold turkey into a game and say the past has no bearing. For instance, if you had the Maple Leafs and the Penguins playing a best-of-7 last year, the Penguins are going to win it 9 out of 10 times, and I can tell you that without having to look at players stats. The Penguins are a better team.

Real-life example: In 2014, the Flyers lost to the Rangers in a harder-than-expected 7 games. It was clear that, although it went 7 games, the Rangers were flat-out better. After the series, my friend sent me a bunch of texts showcasing all of the fancy stats showing why the Rangers were better. However, anyone who can grasp the fundamentals of hockey and make a reasonable inference of what good hockey is could have told you that, I didn't need a bunch of excel graphs to show me that.

Again, if you can take personal bias out, I feel that I am well-informed enough about hockey to analyze a game by watching it myself instead of looking at a table or a graph to tell me which team is doing better than the other.

Edit: The #fancystats may help me quantify the difference between two teams or players.

No, you can't go cold turkey into a game and say the past has no bearing. For instance, if you had the Maple Leafs and the Penguins playing a best-of-7 last year, the Penguins are going to win it 9 out of 10 times, and I can tell you that without having to look at players stats. The Penguins are a better team.

Real-life example: In 2014, the Flyers lost to the Rangers in a harder-than-expected 7 games. It was clear that, although it went 7 games, the Rangers were flat-out better. After the series, my friend sent me a bunch of texts showcasing all of the fancy stats showing why the Rangers were better. However, anyone who can grasp the fundamentals of hockey and make a reasonable inference of what good hockey is could have told you that, I didn't need a bunch of excel graphs to show me that.

Again, if you can take personal bias out, I feel that I am well-informed enough about hockey to analyze a game by watching it myself instead of looking at a table or a graph to tell me which team is doing better than the other.

Edit: The #fancystats may help me quantify the difference between two teams or players.
- jmatchett383

Spending time being concerned with #'s such as shot attempt assist, Zone exits, dangerous shots against and whatever else is the stat of the day just seems like a massive waste of time. Im lucky to get the 2 hrs to watch a game and most of the time i record it and blast through it. Ive played long enough to not need fancy #'s to tell me that 87 is the best player and 58 is a damm good Dman. Watching a game is an escape to drink beer, not spend more mind numbing hours worrying about #'s.

Don't get me started on the (frank) faces that make algorithms for daily fantasy games. It ruined what would other wise have been a boatload of fun. Now it's a sure fire bet that you'll lose money unless you have one of those algorithms and I just don't have the brain power for that. Algo users should be banned. It's the same concept of counting cards but on steroids. If counting cards is taboo in gambling algorithms should be straight up illegal.
- Victoro311

I disagree with illegal. But think similarly. The guys creating those games destroyed their own product. The job as the casino isn't to help others make money....the job of the casino is to create a strong enough game that punters can atleast have a fiction in their head that they can come out ahead and thus they play more. Casino wants to take the money of the punters; they don't want the card sharks taking the moey.

They can tweak certain house rules that would favor punters over professionals. Would prefer the house changing the rules rather than congress.

From what I head they were actually giving out trips and stuff to the guys who played the most (professionals) when you actually want to shift the freebies to the guys losing money.

Spending time being concerned with #'s such as shot attempt assist, Zone exits, dangerous shots against and whatever else is the stat of the day just seems like a massive waste of time. Im lucky to get the 2 hrs to watch a game and most of the time i record it and blast through it. Ive played long enough to not need fancy #'s to tell me that 87 is the best player and 58 is a damm good Dman. Watching a game is an escape to drink beer, not spend more mind numbing hours worrying about #'s.
- sammy87

Looking at numbers is something I enjoy doing on my downtime when I'm bored as a follow up to the games. Never do I ever think about corsis during actual games. I just don't find them mind numbing to be honest.