Devils Analytics Thread

[MadCaptain]

I'm with MCDevils with Camel's scoring. I'm a bit curious to see how he did the past 3 years in that stat though. 1 year is nice but its not a lot of data to base decisions off of.

 

[MCDevils]

I'm with MCDevils with Camel's scoring. I'm a bit curious to see how he did the past 3 years in that stat though. 1 year is nice but its not a lot of data to base decisions off of.

Over the past three seasons, he's 50th in actual goals scored, 4th in Goals/game, since he's had some injury problems, which is the biggest red flag to worry about.

Also, I'm not going to totally discredit a guy for scoring 20 goals.

There were only 21 guys in the league last year that scored 30+. Only 101 scored 20+. That's only 3 per team, so adding another guy who can get that many is still a good pickup. Besides we need a goal scorer on that line with Zajac and Jagr, because Jagr has said time and time again that he sees himself as a passer, not a scorer (which I find ridiculous btw, but whatever, haha).

And he's a natural LW, and by having him play with Jagr and Zajac, you can let Adam and Elias play together, which say AH play his best hockey to date last season.

 

[MCDevils]

Guys around Camm's totals over the past three seasosns are Hartnel, Giroux, JVR, Franzen.

What if I told you over the past three years, Evander Kane has only scored 7 more total goals than Camm.

Like I said he has injury problems, so they need him to stay healthy. Odds are he won't play 82 games, but as long as he's healthy enough to get us to April/May/June hockey, then fine.

 

[MCDevils]

You can check the following link to see Cammalleri's stats over time.

Mike Cammalleri (born on June 8, 1982) - Hockey Stats - QuantHockey.com

Highly recommend checking out the stat for Goals/game. You can click on the 90th percentile to see how he ranks in the league in terms of goal scoring. He's been at or above the 90th percentile in the NHL for the last half decade. He's a good pickup.

 

[nj_cup]

Yeah but you can't just add 15 goals and say they will give us 3 more wins. It doesn't work that way. Jägr breaking down this year rookies on d playing more mins catching hot goalies at the bad times Cory getting hurt all that will come into play

Mc how many goals you think he will put up this year

 

[Grue]

I'm also not afraid to get into nerd-level discussions about data, where it comes from, and how its used. I have enough experience with modelling to know that you can manipulate the data to have it show whatever you want.

Cool thread. I did the same for the Army (the modeling, not the data manipulation - well, not much).

I've always wondered the best way to apply it to hockey. Football and especially baseball are saturated markets for analysts.

 

[MadCaptain]

Cool thread. I did the same for the Army (the modeling, not the data manipulation - well, not much).

I've always wondered the best way to apply it to hockey. Football and especially baseball are saturated markets for analysts.

Baseball is the easiest, and likely the furthest along. Even baseball has its screw ups though. Its really bad with wild pitchers who are just as likely to throw a batting practice pitch as they are a completely unhittable one.

I get the feeling that most of us don't know a whole lot about the modeling.

Whats the link to your thread? Maybe I can cut/paste some stuff to here to further the discussion.

 

[MCDevils]

Jagr will prob score 20-25 again. If we can add 15 more goals to our differential, get to somewhere around +10 we'll be in the playoffs.

I wish I could find the article again, but something like adding or subtracting 5 goals to your differential equals a 1.5 win swing.

 

[Indian Chief]

Yeah but you can't just add 15 goals and say they will give us 3 more wins. It doesn't work that way. Jägr breaking down this year rookies on d playing more mins catching hot goalies at the bad times Cory getting hurt all that will come into play

Mc how many goals you think he will put up this year

If Cory goes down the first week of the season, the Devils are screwed. But the truth is, those things are always in play and you can never predict them. So you look at what the most likely scenarios are and go from there. Maybe Jagr drops off a bit, but having Schneider playing more and no Brodeur easily balances that out. Aside from Jagr, all of last seasons additions were garbage. If they get anything from Ryder, Clowe, Brunner, et cetera, then they are a better team. If someone accidentally scores a shootout goal, they are a better team.

Cammalleri being on the team pushes other guys down on the depth chart, which absolutely makes this team better. We complained all year about guys being on lines they were not qualified to man. So honestly, I think it all balances out to the point that you can say, 'yes the Devils are adding 15 more goals and 3 more wins'. Does that depend on a lot of factors? Yes, it always does. If Quick tears his ACL in the pre-season do I think the Kings are winning another cup? Hell no. But you can't predict that. So you take the most likely scenarios (Jagr, Cory, young D, additions, subtractions) and use the analytics to point you in a direction.

 

[MCDevils]

If Cory goes down the first week of the season, the Devils are screwed. But the truth is, those things are always in play and you can never predict them. So you look at what the most likely scenarios are and go from there. Maybe Jagr drops off a bit, but having Schneider playing more and no Brodeur easily balances that out. Aside from Jagr, all of last seasons additions were garbage. If they get anything from Ryder, Clowe, Brunner, et cetera, then they are a better team. If someone accidentally scores a shootout goal, they are a better team.

Cammalleri being on the team pushes other guys down on the depth chart, which absolutely makes this team better. We complained all year about guys being on lines they were not qualified to man. So honestly, I think it all balances out to the point that you can say, 'yes the Devils are adding 15 more goals and 3 more wins'. Does that depend on a lot of factors? Yes, it always does. If Quick tears his ACL in the pre-season do I think the Kings are winning another cup? Hell no. But you can't predict that. So you take the most likely scenarios (Jagr, Cory, young D, additions, subtractions) and use the analytics to point you in a direction.

Just to reiterate, when I was saying that 15G will get us three wins, there's actually a very accurate way to measure additional wins. Obviously it's not perfect but if the Devils can stay at a similar GA mark and get 15 more GF we're going to get those three extra wins, and that doesn't even count SO points.

 

[MCDevils]

Deny all you want, but the stat-masters are being brought in everywhere.

http://sports.yahoo.com/blogs/nhl-p...ing-of-fancy-stats-is-complete-153753176.html

 

[MCDevils]

Also, I finally found one the articles above. On average every six goal change of goal differential give you an additional win.

So if the Devils get ~15 more goals and keep their defense about the same, they'll get 6 more points in the standings, and potentially be in a playoff spot, or even if they just win some shootouts....

 

[MadCaptain]

Deny all you want, but the stat-masters are being brought in everywhere.

http://sports.yahoo.com/blogs/nhl-p...ing-of-fancy-stats-is-complete-153753176.html

That is why this discussion is needed. These analysts are here right now. They may not be here to stay, but they have an influence in front offices today.

Same with the shootout. No matter how much I hate it, they teams still go to shootouts and seasons are still decided by awarding bonus points for a skills competition.


Indian: It sounds like you're against all types of stats, including GAA, goals, assists, etc. It doesn't really make any sense. GMs were predicting injuries and player rise/decline way before computers even allowed for advanced stats.

 

[Grue]

Baseball is the easiest, and likely the furthest along. Even baseball has its screw ups though. Its really bad with wild pitchers who are just as likely to throw a batting practice pitch as they are a completely unhittable one.

I get the feeling that most of us don't know a whole lot about the modeling.

Whats the link to your thread? Maybe I can cut/paste some stuff to here to further the discussion.

Sorry I wasn't clear. I did modeling for the army. I didn't make a thread about it (sorry, no link)

We did brigade level and below combat simulations to estimate the impact on engineering or weapon system changes. For example, before spending 100 billion on a new armored vehicle, you model the capability in a larger battlefield environment. You compare results against existing capabilities, and then recommend if the system was an improvement. If significant, then some higher up General takes the results to the Pentagon who works with congressman and they lobby for funding.

 

[MadCaptain]

Sorry I wasn't clear. I did modeling for the army. I didn't make a thread about it (sorry, no link)

We did brigade level and below combat simulations to estimate the impact on engineering or weapon system changes. For example, before spending 100 billion on a new armored vehicle, you model the capability in a larger battlefield environment. You compare results against existing capabilities, and then recommend if the system was an improvement. If significant, then some higher up General takes the results to the Pentagon who works with congressman and they lobby for funding.

I had a completely different thought. The Devils' fan group is sometimes referred to as the Devils Army and I figured they had a forum somewhere. I was stupid in hindsight because you're not a Devils fan to begin with and would have no idea what I was talking about.

You're experience just proves a point that Modelling is widespread throughout many different fields already, not just sports. I've done work with modelling programs for watershed analysis, including experiments comparing different programs against each other. One of the biggest problems I ran into was with models getting too complex. They got to a point where you can make the output be whatever you want with minor changes to assumed variables.

That is why I'm interested in the calculations behind these numbers. Certain stats can be of great value while others can have serious flaws that stay persistent through groupthink. My experience with hockey stats is very limited and I don't really know what is out there right now.

 

[Indian Chief]

Indian: It sounds like you're against all types of stats, including GAA, goals, assists, etc. It doesn't really make any sense. GMs were predicting injuries and player rise/decline way before computers even allowed for advanced stats.

Nope, wrong, but maybe I wasn't clear. I quoted Cup because I was refuting what he said. His point was that MC can't say that you add 15 goals and get 3 more wins because of a confluence of other factors. I was basically arguing that in this case you can. Everything about the Devils from last year to this year is balanced in my mind, to the point that you CAN say 'add 15 goals and get 3 more wins'. I ended by saying that you can never predict a freak injury or a free agent that suddenly drops off after you give him 5 years/$25M. (You can try, but if it was really effective, you wouldn't have given a guy that contract.) BUT you take what is in front of you and then use analytics like the one MC brought up about the +/- effect that goals have on wins and predict from there.

I'm not arguing against analytics, in fact I find them interesting. I see Cup's point that you can't use them in a vacuum and think you'll predict next year with perfect accuracy. However, when used responsibly you can find fascinating patterns.

Maybe my previous post makes more sense in that context. I was just addressing Cup, not MC, that's why I quoted what Cup had said.

 

[nj_cup]

guess i'm just old school.

give me a guy that will put up 40 goals (a star player) then you dont have to worry about x+y=z amount of wins.

the devils without camalari this year would've fought for the same playoff spot they are going to fight for with him in the lineup this year. 7th/8th. i'm still not sold on them getting in but i can see it happening. so i guess he is worth 25 million?

 

[MCDevils]

guess i'm just old school.

give me a guy that will put up 40 goals (a star player) then you dont have to worry about x+y=z amount of wins.

the devils without camalari this year would've fought for the same playoff spot they are going to fight for with him in the lineup this year. 7th/8th. i'm still not sold on them getting in but i can see it happening. so i guess he is worth 25 million?

#TorontoMapleLeafs

 

[Indian Chief]

guess i'm just old school.

give me a guy that will put up 40 goals (a star player) then you dont have to worry about x+y=z amount of wins.

the devils without camalari this year would've fought for the same playoff spot they are going to fight for with him in the lineup this year. 7th/8th. i'm still not sold on them getting in but i can see it happening. so i guess he is worth 25 million?

I see what you are saying, except there really are not many 40 goal guys any more (maybe 5?). Hell, how many guys scored even 30 goals last year, 20-25? It also doesn't seem to work out for too many of those guys. Washington has a guy that scores 50 and I don't see them winning anything in my lifetime.

You're also not wrong about what this team is with or without Camm, but without him they miss the playoffs. With him they make it. At least, that's what I believe. Maybe it is still that 7/8 spot but I'd rather be in than out. Can't make a run if you're at home watching.

 

[nj_cup]

#TorontoMapleLeafs

when this team gets shut out by the flames at home 1-0 with the playoffs on the line against kari romo again...tell how important those 15 goals over 82 games really are.