Jason Lewis: When Analytics Don't Seem to Matter

Analytics matter

Jason Lewis: When Analytics Don't Seem to Matter
- Jason_Lewis

They stop mattering when your team is winning despite getting outshot at 5v5

Analytics matter
- Feds91Stammer

They do. But sometimes they don't.

They do. But sometimes they don't.
- Jason_Lewis

1 game is a small sample. Statistics don't apply to small samples.

GO LEAFS GO

As far as numbers capturing the entire game:
http://sbcpuckstorming.bl...of-players-impact-on.html

1 game is a small sample. Statistics don't apply to small samples.
- Feds91Stammer

Tell that to Colorado, Calgary, Ottawa, Edmonton etc. etc.


I am a big proponent of advanced stats, don't get me wrong. I use them ALL the time. Sometimes they can be misleading in both small and larger sample sizes though.

Tell that to Colorado, Calgary, Ottawa, Edmonton etc. etc.


I am a big proponent of advanced stats, don't get me wrong. I use them ALL the time. Sometimes they can be misleading in both small and larger sample sizes though.
- Jason_Lewis

Statistics are a guide to a result not a definition. You have probabilities, standard deviation, etc. etc. that would explain the success of Colorado (last year) and Calgary (this year).

Statistics are a guide to a result not a definition. You have probabilities, standard deviation, etc. etc. that would explain the success of Colorado (last year) and Calgary (this year).
- Feds91Stammer

You are saying exactly what I'm saying. It is a guide, not a definition. It can have unexpected deviations in both the micro and macro sense. It is incredibly useful, but we shouldn't rely SOLELY on it. We should use it in conjunction with more traditional tried and true methods.

You are saying exactly what I'm saying. It is a guide, not a definition. It can have unexpected deviations in both the micro and macro sense. It is incredibly useful, but we shouldn't rely SOLELY on it. We should use it in conjunction with more traditional tried and true methods.
- Jason_Lewis

I think it generally helps us predict things over the longterm. If you replay that LA/PHI game 100 times with the exact same possession spit, how many wins does Philadelphia get?

EDIT - But in short sample sizes it really is meaningless, as you've noted. The Columbus/Tampa game is another prime example. It didn't take into the fact that nearly every Tampa attempt was a long-range, perimeter opportunity.

You are saying exactly what I'm saying. It is a guide, not a definition. It can have unexpected deviations in both the micro and macro sense. It is incredibly useful, but we shouldn't rely SOLELY on it. We should use it in conjunction with more traditional tried and true methods.
- Jason_Lewis

And since no one in the analytics community suggests you should rely "SOLELY" on analytics either...is there a point?

You are saying exactly what I'm saying. It is a guide, not a definition. It can have unexpected deviations in both the micro and macro sense. It is incredibly useful, but we shouldn't rely SOLELY on it. We should use it in conjunction with more traditional tried and true methods.
- Jason_Lewis

Yeah I understand, my original comment was that stats don't apply at all whatsoever in a 1 game sample. Over say a 7 game series where you would have a lower probability of a player/goalie perform above their statistical norms, statistics would most likely provide you the winner of the series.

Yeah I understand, my original comment was that stats don't apply at all whatsoever in a 1 game sample. Over say a 7 game series where you would have a lower probability of a player/goalie perform above their statistical norms, statistics would most likely provide you the winner of the series.
- Feds91Stammer

Avalanche fans would reluctantly agree I'm sure.

And since no one in the analytics community suggests you should rely "SOLELY" on analytics either...is there a point?
- glgbill

There are those who rely too heavily on it, that's for certain.

There are those who rely too heavily on it, that's for certain.
- Jason_Lewis

Not among people who understand them. And you wrote following:

"...what I'm saying is that if there was a clear example that advanced stats are not the be-all end-all of hockey analysis, yesterday was it."

Which sort of raises the question whether you understand sample sizes (and more broadly, analytics), or if instead you're simply attempting to make a point against something for another purpose.

Not among people who understand them. And you wrote following:

"...what I'm saying is that if there was a clear example that advanced stats are not the be-all end-all of hockey analysis, yesterday was it."

Which sort of raises the question whether you understand sample sizes (and more broadly, analytics), or if instead you're simply attempting to make a point against something for another purpose.
- glgbill

You're missing my point, but sure, we'll go with the bolded. Hey thanks for reading.

I wouldn't say the flyers are unstructured. They are just outclassed and out talented.

They do. But sometimes they don't.
- Jason_Lewis

They don't, but sometimes they do.

They don't, but sometimes they do.
- flyerdude17

The Kings are a better hockey club than the Flyers - and it's not close. The Flyers are not absolutely unstructured, but certainly are not the team they were last season. They also are in the middle of a pretty long slump where they've won just 2 of 12 or something (and from which they are just starting to rebound). We may see an improved Flyers team in the coming weeks but make no mistake, this is not looking like a playoff team.

Having said that, and noting your correct observation that they were destroyed in the possession game, saying Mason "made the saves he had to", does not properly characterize what happened last night. He stole the game plain and simple - it can happen to the best of teams. The only relevant stat was Mason's .974 save pct. whatever your opinion of him, I'm not sure you put enough emphasis on that.

Besides that, your blog was good and I enjoy reading it (I actually like the Kings and route for them a lot).

Jason Lewis: When Analytics Don't Seem to Matter
- Jason_Lewis

I feel like you and Ryan Wilson are best friends with the advance stats and graphs

im so excited for kopitar to get out of the slump hes in, hes going to be unstoppable come playoff time.

The Kings are a better hockey club than the Flyers - and it's not close. The Flyers are not absolutely unstructured, but certainly are not the team they were last season. They also are in the middle of a pretty long slump where they've won just 2 of 12 or something (and from which they are just starting to rebound). We may see an improved Flyers team in the coming weeks but make no mistake, this is not looking like a playoff team.

Having said that, and noting your correct observation that they were destroyed in the possession game, saying Mason "made the saves he had to", does not properly characterize what happened last night. He stole the game plain and simple - it can happen to the best of teams. The only relevant stat was Mason's .974 save pct. whatever your opinion of him, I'm not sure you put enough emphasis on that.

Besides that, your blog was good and I enjoy reading it (I actually like the Kings and route for them a lot).
- BOSS_TWEED

When I said the Flyers were unstructured I figured it would be a bit contentious.

They were definitely so on the penalty kill, and I felt they were a touch on their neutral zone breakouts. Offensively you can tell the Flyers have a couple of players who really command the puck though. That aspect of the game isn't as I said.

Maybe I didn't give Mason enough credit. I didn't think he had to make too many outstanding saves, but he certainly made all the ones he had to and then some.

I feel like you and Ryan Wilson are best friends with the advance stats and graphs
- Guile

Pshhhht that guy is such a hack