The ****ing Off Season thread

[MindzEye]

Hockey isn't baseball or football. It's a seamless transition back and forth between offense and defense.

Having some stat geek come in won't do much.

Basketball shares a lot of the same dynamic nature of hockey and proprietary stats have changed the way that GM's build teams.

It has already started doing the same in hockey, we just don't know exactly how yet because the teams involved in the analytics, don't publish what methods they're using.

 

[hockeylover]

You'd better hope investing in analytics doesn't make a difference - a team being outspent by the Florida Panthers by $10 million probably can't afford to.

 

[hockeylover]

Acting like Dubas is just a stats geek is foolish anyway, IMO. He's not a Yale Economics grad, he's a hockey guy.

 

[MindzEye]

If Sens fans don't like this move, I love it.

 

[hockeylover]

Part 1 of an interview with Dave Morrison. Fairly lengthy, not done reading yet.

http://mapleleafshotstove.com/2014/07/23/mlhs-in-conversation-with-dave-morrison-part-1/

 

[Montana]

Wunderkid will provide virtually no tangible improvements to the club. Guaranteed.

Thanks Art.

 

[corksens]

If Sens fans don't like this move, I love it.

I don't not like it - you guys are just expecting too much.

The sport isn't conducive to this type of number crunching having a significant impact like other sports.

 

[leafman101]

Everyy good organization in the NHL already has an analytics department that crunches numbers with success. The Leafs were behind the eight ball here, not on the cutting edge.

 

[LeafGm]

I don't not like it - you guys are just expecting too much.

The sport isn't conducive to this type of number crunching having a significant impact like other sports.

I don't entirely disagree with you. I think it is a lot easier to mathematically quantify a relatively static sport like baseball than it is to do the same with a dynamic sport like hockey, where there's no real separation between offense and defense.

At the same time, I don't see why an organization with almost unlimited resources like MLSE shouldn't try and get the most information they possibly can, and try and get a leg up on broke-ass franchises that aren't willing to spend that money. You could even look to the amateur research that Mindz did a few years back, when he tracked the correlation between junior performance and the likelihood of a player scoring at a first-line level in the NHL. If someone posting on an internet message board could put together useful information like that in his spare time, I'd like to see what kind of proprietary statistical information our new assistant GM and a few analysts working full-time could put together.

 

[UWHabs]

Yup - Baseball you can almost entirely model, and the model tends to be solid. There's not a whole lot of interaction between players on the field, and with the new Field-Fx stuff coming out and all the Pitch-Fx work recently, the games is almost to the point of playing on a spreadsheet.

Hockey is a ton more fluid, with way way way more interaction between players. Even to the point of "do I play my 2 best players together, or do I put them on separate lines", different play-styles, etc... It will never get as analytic as Baseball. But that doesn't mean you shouldn't try. Even this corsi junk looks like it's helping teams get away from useless players and move to slightly more high skill guys. At the very least some of the analysis can help to see if the coach is playing the right guys at the right times.

 

[MindzEye]

I don't not like it - you guys are just expecting too much.

The sport isn't conducive to this type of number crunching having a significant impact like other sports.

Yeah it is.

There is fantastic analytics work being done in soccer, and in basketball. Both very similar sports in nature that put a premium on the maintenance of possession and have an extremely dynamic nature. You're confusing the speed of the game with the complexity necessary in tracking the individual interactions. When you have the budget to have a team of analysts sit down and dissect the game piece by piece, and then develop metrics based on what they deem to be important interactions.

If you were shopping on the free agent market, and had a dozen bargain basement guys available, wouldn't you want to know which ones had the highest passing completion %? How about the % of board battles they win, or the % of time a winger is able to clear the zone when the defender pinches down on him at the half boards?

How about if it's contract time for a high paid player (Bobby Ryan...?), wouldn't you want to have more data about him to help determine if he's worth the 7 million dollars he's going to ask for?

 

[MindzEye]

Yup - Baseball you can almost entirely model, and the model tends to be solid. There's not a whole lot of interaction between players on the field, and with the new Field-Fx stuff coming out and all the Pitch-Fx work recently, the games is almost to the point of playing on a spreadsheet.

Hockey is a ton more fluid, with way way way more interaction between players. Even to the point of "do I play my 2 best players together, or do I put them on separate lines", different play-styles, etc... It will never get as analytic as Baseball. But that doesn't mean you shouldn't try. Even this corsi junk looks like it's helping teams get away from useless players and move to slightly more high skill guys. At the very least some of the analysis can help to see if the coach is playing the right guys at the right times.

Yeah.

Hockey will never be figured out to the point baseball is due to the difference in natures of the respective sports, but to say that analytics can't have a large, very positive influence on how GM's put their teams together is flat out wrong.

 

[leafman101]

The biggest problem in hockey stats has been manually observing and calculating all of the situations that occur. But with the new technology that is out there there the possibilities of stat tracking has improved greatly.

Like the stuff the Raps have been using with the advanced cameras that capture everything and the programs they have developed to process all of that information.

Teams in the NHL are way ahead of what we use on behindthenet or extraskater. At least the ones that are smart enough. Thankfully the Leafs just joined that group.

 

[Artnes]

Acting like Dubas is just a stats geek is foolish anyway, IMO. He's not a Yale Economics grad, he's a hockey guy.

This. The vast majority of people know nothing more about him then what the media has protrayed him as in the past few days.

He will be labled as the fancy stats guy until he steps in a GM roll and proves otherwise.

 

[number17]

Is Dubas just a 'stats geek' now?

Cause that's what many here are portraying him to be, which, cannot be further from the truth. In his own PC he kept stressing advanced stats, or stats in generals, is just 1 of many ways to analyze a team.

 

[BG]

Not sure what people are expecting of analytics.. the team still has to win 16 play-off games to win the Cup. The additional analytics will just help focus management on the strengths/weaknesses of the group, add another component to any decision making, and occassionally discover/exploit minor inefficiencies in drafting/free agency.

 

[zeke]

Its the organizational attitude change this signals which is most important here.

For all the SMH versions of "culture change" we've had to witness, this is finally the real thing.

 

[leafman101]

Not sure what people are expecting of analytics.. the team still has to win 16 play-off games to win the Cup. The additional analytics will just help focus management on the strengths/weaknesses of the group, add another component to any decision making, and occassionally discover/exploit minor inefficiencies in drafting/free agency.

Those are some pretty huge improvements if you ask me.

 

[zeke]

Yup - Baseball you can almost entirely model, and the model tends to be solid. There's not a whole lot of interaction between players on the field, and with the new Field-Fx stuff coming out and all the Pitch-Fx work recently, the games is almost to the point of playing on a spreadsheet.

Hockey is a ton more fluid, with way way way more interaction between players. Even to the point of "do I play my 2 best players together, or do I put them on separate lines", different play-styles, etc... It will never get as analytic as Baseball. But that doesn't mean you shouldn't try. Even this corsi junk looks like it's helping teams get away from useless players and move to slightly more high skill guys. At the very least some of the analysis can help to see if the coach is playing the right guys at the right times.

See, but this is where the wall street model "the extra two percent" comes in, as Keri's book on the Rays describes.


Just that tiny 2% improvement in analysis can make all the difference.

 

[leafman101]

Dubas might not be an MIT geek, but he's part of the new generation in sports that has an appreciation for analytics. Thats why he is a rising star and that is why he was brought in. Statistics have a very important part in everything, including hockey. Nothing new around these boards, but the dinosaurs running the show had resisted using all the information available to them for too long.