Eric Tulsky at MIT/Sloan Sports Conference

[Navin R Slavin]

I posted about this in another thread, but I think it deserves its own thread.

Everyone here who is even vaguely interested in analytics -- which is basically all of us -- should watch this talk right now.

No Stone Unturned: Keeping up with the state of the art in the public domain - MIT Sloan Analytics Conference

Click to read more...

 

[BrindAmours Army]

I posted about this in another thread, but I think it deserves its own thread.

Everyone here who is even vaguely interested in analytics -- which is basically all of us -- should watch this talk right now.

No Stone Unturned: Keeping up with the state of the art in the public domain - MIT Sloan Analytics Conference

Can those of us that can't watch right now get a tl;dr?

 

[Navin R Slavin]

Can those of us that can't watch right now get a tl;dr?

I wouldn't do it justice. Go watch it when you have time.

The very basics: the public is good at tracking some things, the company is better at tracking other things, the public is working with incomplete data but has the advantage of horsepower. It's really a fascinating insight into how the organization actually thinks -- the best insights I've seen.

 

[BrindAmours Army]

I wouldn't do it justice. Go watch it when you have time.

Fair enough! I'll leave the tab open on my phone for later

 

[Anton Dubinchuk]

Just watched it. Time well spent. His reasoning for internal bias is one I hadn’t considered before; over the course of years you’ve created a roster made up of players that you more or less liked better than anyone else did at the time, so of course you’ll end up liking your team more than most. That’s not even a bias that necessarily trickles down to fans of the team - it’s not like Rutherford, Francis, and Waddell got our sign-off before making the moves that created this team. It’s not necessarily pride getting in the way of objectivity either; it’s simply that from the beginning our front office inherently likes every move and every player, otherwise they wouldn’t have made the move or acquired the player. That explanation really got my mind working on stuff.

 

[Navin R Slavin]

Like, I feel like we ought to basically make a transcript of it and post it as a sticky thread.

 

[bleedgreen]

I hate stats and I’d be interested. Hopefully I’ll have time to watch it today.

 

[Cardiac Jerks]

Fair enough! I'll leave the tab open on my phone for later

I’ve done that before and now have over twenty tabs waiting for me to get back to later.

 

[emptyNedder]

I figured I would put this here instead of on the general thread. There fans were asked to rank the D. Pesce is ranked third by some.
I made this argument earlier in the season and it bears repeating now: Pesce is having a better year than Slavin. I can buy the argument that based on the past 4 years Slavin is still considered slightly better, but this year Pesce is the team's best D by a not insignificant amount. Here are some of the analytics:

Hockey-Reference: Pesce 9.5 Point Share, Slavin 8.8
Corsica: Slavin 76.91 Player Rating, Pesce 76.19
Hockeyviz Threat shots against: Pesce 7% better than the team, Slavin 3% worse (level of competition is nearly identical)
Evolving Hockey--Pesce 2.0 WAR, Slavin 0.3 WAR; Pesce 9.6 Even strength GAR, Slavin -2.6.

Three out of four analytic views either favor or strongly favor Pesce. One favors (maybe slightly favors) Slavin.

There is also the basic statistics view: Even Strength Goals for per 60 and Even strength Goals Against per 60
Pesce GF/60=3.3 GA/60=1.9
Slavin GV/60=2.5 GA/60=2.7

None of this is to detract from Slavin--I would take him all day over almost any d-man in the league.
However, Pesce might be having the best year of any d-man in the league with less than 50 points. Even Caniacs paying attention seem to be missing how well Pesce is playing.

 

[Unsustainable]

I’m working 12-14 hours a day. I don’t have time, but the subject is interesting to me.

 

[Canes]

I figured I would put this here instead of on the general thread. There fans were asked to rank the D. Pesce is ranked third by some.
I made this argument earlier in the season and it bears repeating now: Pesce is having a better year than Slavin. I can buy the argument that based on the past 4 years Slavin is still considered slightly better, but this year Pesce is the team's best D by a not insignificant amount. Here are some of the analytics:

Hockey-Reference: Pesce 9.5 Point Share, Slavin 8.8
Corsica: Slavin 76.91 Player Rating, Pesce 76.19
Hockeyviz Threat shots against: Pesce 7% better than the team, Slavin 3% worse (level of competition is nearly identical)
Evolving Hockey--Pesce 2.0 WAR, Slavin 0.3 WAR; Pesce 9.6 Even strength GAR, Slavin -2.6.

Three out of four analytic views either favor or strongly favor Pesce. One favors (maybe slightly favors) Slavin.

There is also the basic statistics view: Even Strength Goals for per 60 and Even strength Goals Against per 60
Pesce GF/60=3.3 GA/60=1.9
Slavin GV/60=2.5 GA/60=2.7

None of this is to detract from Slavin--I would take him all day over almost any d-man in the league.
However, Pesce might be having the best year of any d-man in the league with less than 50 points. Even Caniacs paying attention seem to be missing how well Pesce is playing.

Even before this stretch of play by Pesce and when we were crap and couldn't score goals, I was pretty adamant about not trading Pesce. Yes, we could still use a guy like Nylander (at least Nylander before this year anyway) but I'm not trading Pesce for him. I don't care how much depth we have with RHD. Great d-man on insane deal, not even close to being easily replaceable.

 

[Anton Dubinchuk]

I’m all aboard the Pesce train, and have been since halfway through the season. He’s super good.

 

[bleedgreen]

I figured I would put this here instead of on the general thread. There fans were asked to rank the D. Pesce is ranked third by some.
I made this argument earlier in the season and it bears repeating now: Pesce is having a better year than Slavin. I can buy the argument that based on the past 4 years Slavin is still considered slightly better, but this year Pesce is the team's best D by a not insignificant amount. Here are some of the analytics:

Hockey-Reference: Pesce 9.5 Point Share, Slavin 8.8
Corsica: Slavin 76.91 Player Rating, Pesce 76.19
Hockeyviz Threat shots against: Pesce 7% better than the team, Slavin 3% worse (level of competition is nearly identical)
Evolving Hockey--Pesce 2.0 WAR, Slavin 0.3 WAR; Pesce 9.6 Even strength GAR, Slavin -2.6.

Three out of four analytic views either favor or strongly favor Pesce. One favors (maybe slightly favors) Slavin.

There is also the basic statistics view: Even Strength Goals for per 60 and Even strength Goals Against per 60
Pesce GF/60=3.3 GA/60=1.9
Slavin GV/60=2.5 GA/60=2.7

None of this is to detract from Slavin--I would take him all day over almost any d-man in the league.
However, Pesce might be having the best year of any d-man in the league with less than 50 points. Even Caniacs paying attention seem to be missing how well Pesce is playing.

The first full year they played, where they both ended up over +20 and everyone else was minus - I think we had legit arguments going on here that Pesce was better at times.

I would wonder bout how Slavin's fancy stats were affected by the painful first attempt at the Dougie/Slavin pairing that seemed to hurt both players earlier on. Did the first time end in December? They're much better now together, but it was borderline awful at times the first half of the year.

Faulk has "won" the who's getting traded discussion. I would argue that this incarnation of the team began with the pillars of Slavin and Pesce. Those two and Aho are the guys we build around.

 

[NotOpie]

I argued vociferously on the main trade Nylander/Pesce threads that the real reason no deal happened is that when the teams came down to figuring out who would have to add, Donny Wads told the whippersnapper that any deal for Brett Pesce would require a significant add on the Leaf's part....if Nylander was the main piece coming back.

I still hold that opinion.

 

[Boom Boom Apathy]

I argued vociferously on the main trade Nylander/Pesce threads that the real reason no deal happened is that when the teams came down to figuring out who would have to add, Donny Wads told the whippersnapper that any deal for Brett Pesce would require a significant add on the Leaf's part....if Nylander was the main piece coming back.

I still hold that opinion.

The nice thing about trades that never happen, is that we can hold whatever opinion we so choose for the rest of time and nobody can prove us wrong.

 

[SlavinAway]

I would wonder bout how Slavin's fancy stats were affected by the painful first attempt at the Dougie/Slavin pairing that seemed to hurt both players earlier on. Did the first time end in December? They're much better now together, but it was borderline awful at times the first half of the year.

Faulk has "won" the who's getting traded discussion. I would argue that this incarnation of the team began with the pillars of Slavin and Pesce. Those two and Aho are the guys we build around.

I was wondering the same thing. How would the comparison between Slavin and Pesce stack up if looked at over the period of time since Slavin and Hamilton were put back together and looked like a legit 1st pairing?

Completely agree with the second piece. Slavin, Pesce, Aho, and Svech are the pillars we build around.

 

[emptyNedder]

I was wondering the same thing. How would the comparison between Slavin and Pesce stack up if looked at over the period of time since Slavin and Hamilton were put back together and looked like a legit 1st pairing?

I am not sure I can find the comparison numbers. But I do know from hockeyviz that Pesce/Faulk took most of the shifts against opponents top lines starting in late January. That included the game where McDavid didn't score, the shutout of the Pens, and the first shutout of the Stars. Also, Faulk has gone from a minus 4 to plus 5 since Jan. 20 when he was paired with Pesce. Pesce has actually been +18 during that period so he has helped other D-men be plus 9 as well as helping Faulk be +9.

Since Jan. 20: Slavin 10 points; +11. Hamilton 15 points; +12;
Faulk 12 points; +9. Pesce 14 points; +18.

 

[NotOpie]

The nice thing about trades that never happen, is that we can hold whatever opinion we so choose for the rest of time and nobody can prove us wrong.

You know what they say, "....opinions are like assholes....everybody has one and nobody thinks theirs stinks".

 

[Elsker]

Thanks for sharing that link. Video is about 20 minutes long for those that have it on a tab.

Eric Tulsky was/is widely respected as an originator, not just a copycat, in the advanced stats universe. His hire was quite the catch for the org when we took him private.

In addition, @datarink (Kevin Kan) was pretty much my favorite Twitter person for data visualization techniques. So was really impressed when he also came to work under Eric a while back to not only assist with data analysis, but also developing the visual means of conveying that analysis to the front office.

And now, per Eric's Twitter account, looks like the org is also hiring a data engineer and a data scientist to prepare for the onslaught of data that will come from player tracking technology, the very thing he was talking about towards the end of his presentation during the Q&A session.

We're certainly not being left behind at the station on this front.

 

[Bunch of Jurcos]

Thanks for sharing that link. Video is about 20 minutes long for those that have it on a tab.

Eric Tulsky was/is widely respected as an originator, not just a copycat, in the advanced stats universe. His hire was quite the catch for the org when we took him private.

In addition, @datarink (Kevin Kan) was pretty much my favorite Twitter person for data visualization techniques. So was really impressed when he also came to work under Eric a while back to not only assist with data analysis, but also developing the visual means of conveying that analysis to the front office.

And now, per Eric's Twitter account, looks like the org is also hiring a data engineer and a data scientist to prepare for the onslaught of data that will come from player tracking technology, the very thing he was talking about towards the end of his presentation during the Q&A session.

We're certainly not being left behind at the station on this front.

I wish I had gone to grad school now to be a data scientist. I believe this area of analytics is going to be a huge part of the future even though it has grown quite a bit in the past few years. It would be awesome to be at the forefront of big data.

 

[BrindAmours Army]

I’ve done that before and now have over twenty tabs waiting for me to get back to later.

Yeahhh I have a slight problem with tabs. I don't remember the last time I had under 100 open.

 

[Ahoy there]

Thanks for sharing that link. Video is about 20 minutes long for those that have it on a tab.

Eric Tulsky was/is widely respected as an originator, not just a copycat, in the advanced stats universe. His hire was quite the catch for the org when we took him private.

In addition, @datarink (Kevin Kan) was pretty much my favorite Twitter person for data visualization techniques. So was really impressed when he also came to work under Eric a while back to not only assist with data analysis, but also developing the visual means of conveying that analysis to the front office.

And now, per Eric's Twitter account, looks like the org is also hiring a data engineer and a data scientist to prepare for the onslaught of data that will come from player tracking technology, the very thing he was talking about towards the end of his presentation during the Q&A session.

We're certainly not being left behind at the station on this front.

Player tracking tech was interesting, and I'm seeing it pop up more with the Canes. Wondering out loud if part of the impetus behind this innovation is consistently fielding a competitive squad while not spending to the cap. Advance analytics will identify who to take a risk on (Nino, Mrz) and when to swing for a star.

 

[Canes]

I wish I had gone to grad school now to be a data scientist. I believe this area of analytics is going to be a huge part of the future even though it has grown quite a bit in the past few years. It would be awesome to be at the forefront of big data.

There are always fads in sports but I definitely think data/advanced statistics are here to stay.

Yeahhh I have a slight problem with tabs. I don't remember the last time I had under 100 open.

I can't stand the tabs being so small so I'll have 4 or 5 windows open with 10-15 tabs each.

 

[GIN ANTONIC]

What does Tlusty know about analytics? He should stick to what he knows which is leeching points off of E. Staal and taking selfies.