Newly hired analytics specialists: Dubas vs Dellow vs Mehta?

[greger16]

Hey HF world, its my first post here so I'm still getting my feet wet for sure (go easy on me!).

With all of the recent hiring of analytic advocates, (Kyle Dubas for Toronto, Tyler Dellow for Edmonton, Sunny Mehta for New Jersey) my curiosity wandered to the potential impact each of these individuals bring to their respective clubs.

In saying that, from what I've found on each of them through simple google searches, I am still foggy as to what each can bring to their respective clubs with their own expertise.

Kyle Dubas - CHL GM who seemingly turned a bottom feeding SOO team to a contending/desirable player location. Also a pioneer of advanced stats at the OHL level. As an organization they recorded their own stats deemed important to them - stats not otherwise tracked in game logs.

Tyler Dellow - Outspoken blogger who is(was) a very active advocate for advanced stats, and vocal critic when it comes to management and gameplay - particularly against the Edmonton Oilers

Sunny Mehta - Poker star and author of two books on beating the odds of poker. One would assume he knows his way around numbers and a spreadsheet?

Like I said, my research on these fellows have been rather superficial (what the Google machine provides me) so if anyone has more comprehensive information about the potential impact these hires will have on their respective clubs in regards to contracts, asset management, or even style of play, please share! I would love to have a solid mindset going into this season and future seasons about the impact we will see from analytics at the ice level as well as on an organizational level. Also figured it would be fun to compare how these new minds stack up against each other (if at all).

Cheers!

Click to read more...

 

[Cloned]

Dellow and Mehta are friends IIRC.

Dellow is the more vocal of the two but they are both extremely intelligent.

The real question is who will hire Vic Ferrari?

 

[Hivemind]

These are the three very public hires, but plenty of other teams already have analytics departments or use outside consultants. As none of us are actually in these front offices, it's hard to say just how they will be utilized, how much management will listen to them, and what type of input they will have to which processes. It is worth noting that Dubas is an assistant general manager, while the other two are not (suggesting that he'll have more input on the actual decision making, as opposed to just supplying analysis and information).

 

[NHLPaul]

Both Dellow and Mehta are also friends with Vic Ferrari (now known to be one Timothy Barnes). Both responded often to his posts on "Irreverent Oiler Fans" too and had great discussions. Worthy of checking out if you haven't.

 

[Kevin Forbes]

Probably worthwhile to add Brian MacDonald to this list as well.

It might be just me, but I'm intrigued by the different roles/job titles being thrown around:
Obviously Dubas is an Assistant GM.
In cases like MacDonald and Mehta, they're Director of Hockey Analytics.
Dellow has been called a consultant by the Oilers, but his exact title isn't listed. Eric Tulsky is also consulting with an undisclosed NHL club.

 

[hatterson]

Impact on the individual clubs will obviously depend on how much weight their opinions are given. You can have the worlds greatest stats guy in your office, but if you constantly ignore his recommendations there's no difference between him and a monkey with a dartboard on your club.

That being said, Dubas' title would lend one to believe he would have a decent amount of input on decisions followed by MacDonald and Mehta as directors and Dellow as a consultant. However, there's many organizations where the consultant ends up having the most influence in practice and the "assistant GM" ends up being no more than the "assistant to the GM" and without being inside any of these organizations, I can't really say how they will operate.

As a Leafs fan I really hope that Dubas is given a significant amount of say as I believe analytics is something we're sorely missing and have screwed up a lot in the past, but that's just hope, not actual insider insight.

 

[BeardyCanuck03]

It will be interesting to follow these teams and if we can see any improvements that can be attributed to them. The thing with advanced stats is that they are results and the good teams can take those results and figure out why these results are happening. Does a player have bad corsi because of his play, his linemates, the system...

 

[Bardown warrior]

Dubas wasn't hired as a pure analytic's guy, the fact that he uses them just means it's a weapon in his seemingly vast arsenal; the others are consultants for teams, while Dubas is an assistant general manager, who will have to use much more then analytic's to run the leafs.

 

[Hivemind]

Another thread mentioned that the Oilers already have an analytics department to work with the GMs, and that Dellow was hired by Eakins to work with the coaching staff.

 

[Colin226]

Impact on the individual clubs will obviously depend on how much weight their opinions are given. You can have the worlds greatest stats guy in your office, but if you constantly ignore his recommendations there's no difference between him and a monkey with a dartboard on your club.

That being said, Dubas' title would lend one to believe he would have a decent amount of input on decisions followed by MacDonald and Mehta as directors and Dellow as a consultant. However, there's many organizations where the consultant ends up having the most influence in practice and the "assistant GM" ends up being no more than the "assistant to the GM" and without being inside any of these organizations, I can't really say how they will operate.

As a Leafs fan I really hope that Dubas is given a significant amount of say as I believe analytics is something we're sorely missing and have screwed up a lot in the past, but that's just hope, not actual insider insight.

The bolded above will be key with Mehta and something Devils fans are very curious about.. Lou came out and said that he's going to be the liason between Mehta's work and its impact on team decisions, so I'm interested to see how much Lou buys in.. On the one hand, Lou just seems to be very old school while on the other hand he was a math teacher in the 80's and said he's looking forward to having an Analytics Dept..

I've seen Devils fans statistically show how awful Stephen Gionta can be, yet our coach and GM seem to love him and refer to his non-statistical impacts such as "bringing energy" so it'll be fun to see statistics versus intuition this season in NJ

 

[Jumptheshark]

as someone who works numbers year round, the problem I have with how everyone has gone goo goo gaga over advance or fancy stats--there are too many ways to blow holes into them. Because, unlike baseball, hockey is a very physical and demanding game, while you are able to work out some stats, because hockey is more of a team sport then the individual stats that pushed money ball, the hockey stats are can be filled with more conjecture and open for debate.

 

[getyourselfsomerest]

Well Dubas is an assistant GM with tangible managerial experience who is actually going to be making decisions. The other two guys are bloggers who will be nothing more than advisers to their respective teams.(no offense)

 

[Sam Spade]

as someone who works numbers year round, the problem I have with how everyone has gone goo goo gaga over advance or fancy stats--there are too many ways to blow holes into them. Because, unlike baseball, hockey is a very physical and demanding game, while you are able to work out some stats, because hockey is more of a team sport then the individual stats that pushed money ball, the hockey stats are can be filled with more conjecture and open for debate.

I'm with you.

Personally I think the stats should have SOME weight but the human element should be the ultimate decider (Ken Holland called them "gut analytics").

This may be harsh but I really hope all these big hires fail miserably.

 

[KaseMeOutside]

it doesn't matter unless the organizations as a whole fully support them, if the people making player decisions ignore them it doesn't matter

 

[Kevin Forbes]

Well Dubas is an assistant GM with tangible managerial experience who is actually going to be making decisions. The other two guys are bloggers who will be nothing more than advisers to their respective teams.(no offense)

That's completely correct, Dubas brings different qualities to the table compared to Dellow and Mehta. When it comes to the use of analytics, Dubas is probably closer to Dallas Eakins (using analytics to support and supplement his work) rather than Tyler Dellow (analytics specialist). Including him in the conversation doesn't fit, which is kind of what I was getting at with my first post about roles.

I'm with you.
Personally I think the stats should have SOME weight but the human element should be the ultimate decider (Ken Holland called them "gut analytics").
This may be harsh but I really hope all these big hires fail miserably.

There's kind of two different strings here.
I don't think anyone is saying that the Oilers/Devils/Panthers/Leafs/whoever are going to just blindly follow the recommendations of their new hires when it comes to hockey decisions. I view this as recognition that there is value in the work that they do and that they want to have an additional voice and perspective around the table when it comes to making decisions. Something to add to their current methods of evaluation and decision making.
The hope that these analysts fail is a bizarre thing to say unless you have a particular dislike for the teams involved. Teams have used different methods to evaluate their players and their opposition over the years, these hirings are not too much different from a team hiring a new pro scout.

 

[hatterson]

as someone who works numbers year round, the problem I have with how everyone has gone goo goo gaga over advance or fancy stats--there are too many ways to blow holes into them. Because, unlike baseball, hockey is a very physical and demanding game, while you are able to work out some stats, because hockey is more of a team sport then the individual stats that pushed money ball, the hockey stats are can be filled with more conjecture and open for debate.

I'm with you.
Personally I think the stats should have SOME weight but the human element should be the ultimate decider (Ken Holland called them "gut analytics").
This may be harsh but I really hope all these big hires fail miserably.

I think you'll be hard pressed to find a stats guy who says "Well player X has a higher <statname> than player Y so he's better" Basically everyone who is interested in analytics beyond a desire to sound smart understands that hockey is a tough sport to accurately measure (we have an entire thread about that debate) because of the team aspect and the physical nature of the game (eg, a player getting rocked by a check and being afraid to go into the corner for the rest of the game is a very real thing).

Stats cannot be properly used in a vacuum. You can't just grab an excel sheet, throw in a couple functions, hit sort, come out with a hockey team and expect it to win a championship. However, we must also recognize that our perception of events as we watch a game are flawed (sometimes deeply so) by biases and the limitations of our brain. Stats are very valuable in combating those biases, in grounding our observations with cold hard facts.

 

[Halibut]

Another thread mentioned that the Oilers already have an analytics department to work with the GMs, and that Dellow was hired by Eakins to work with the coaching staff.

That's the story, personally I think they hired him to shut him up.

 

[Pyrophorus]

That's the story, personally I think they hired him to shut him up.

...or found value in his criticisms.

 

[Mr Sakich]

Another thread mentioned that the Oilers already have an analytics department to work with the GMs, and that Dellow was hired by Eakins to work with the coaching staff.

that was my post. The oilers have gone heavily into analytics. Dellow was hired to work with the coaching staff. They have 3 guys who work with the GM.

They also hire Darkhorse Analytics to compile stats for the draft elligable players. Two recent draft picks who were apparently selected because of Darkhorse findings are Marco Roy and Greg Chase.

It will be interesting to watch Oiler drafting over the next few years to see if any trends occur in their drafting and if it is successful.

 

[TLeafsFan]

Dubas will have the most influence on his respective team. Plus assistant GM of the Leafs, should mean GM of the Marlies. He's already called a couple guys up from the SOO.