What's new
By:
USCHO Fan Forum

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

  • The USCHO Fan Forum has migrated to a new plaform, xenForo. Most of the function of the forum should work in familiar ways. Please note that you can switch between light and dark modes by clicking on the gear icon in the upper right of the main menu bar. We are hoping that this new platform will prove to be faster and more reliable. Please feel free to explore its features.

Statistics Models

Re: Statistics Models

I just got caught up on the NHL 2012-13 realignment. It seems that the biggest complaint of the public is inter-division imbalance w.r.t. playoff structure. I'm inclined to think that having the first 2 rounds take place within your division is a good counter-perk to this grievance.

I'm thinking the imbalance is more an issue when it comes to draft seeding. With a more isolated schedule, W/L/O records are likely going to bring injustice to seeding. the emphasis of the SOS component of KRACH makes me think that it might be a suitable tool for more accurate seeding. But, having not played with the model firsthand yet, I wonder if the team schedules will be varied enough, given that your schedule will be just about identical to everyone else in your division and that every team matches up with every other team at least twice.

What do you guys think? Is there a tool that can do the job better? Will the NHL ever adopt something so "hard to understand?"
 
Re: Statistics Models

Numbers said:
And, generally, I have another question that seems to belong here.

If college hockey wishes to choose its' NCAA field by game results only, and KRACH (can we please find a better name? And, how would a statistician really refer to this method?) does so as well as any, how do we deal with the following problem?

Currently, the top of the list is filled with CCHA teams. I mean filled. I don't really have a problem with that, but I have a feeling that the math works out that way because the number of non-conf games is small, so a couple of handfuls of good results elevate the entire league.

Again, I don't really have a problem with that - if you want to use results, then use results. What I wish was that there were a way to smear the benefit a little. Does anyone understand what I mean?

Maybe in short it would be like this: KRACH makes the non-conference results of all the teams in one's league to be very important, because of the high number of insulated games within conferences. How can we tone that down a little?

Thanks,
Numbers
Click to expand...

Intra-conference effects are mostly a result of what is termed in statistics as "leverage"... this is a concept of linear modeling that says that certain observations, through their predictors (independent variables, etc... depends on which book you're using) will frequently serve to be over-influential on the model. This often comes when there's only certain situations which serve to suitably inform on the overall picture. Think of it sort of like the matrix concept of linear dependence. In this case, if there are results that are far from what is most likely its because of a few observations that have high influence... non-conference games are that type of thing because there's only so much information relating to the general position of the conference in relation to each other. This is also the classic "SEC" effect... it really depends on what's true... the opposite is the "Mount Union effect" or we could dub it the "Boise effect".

Anyhoo... how do you deal with the over-influence of a successful conference? Usually you can't. There are certain procedures (hierarchical models as discussed before) which could dampen the values and bring them closer to center. Bayes techniques may impose a different form of dampening but they're incredibly complicated for the sports ranking problem. The real question is "what do you assume" and can you express it in a tractable mathematical form.

The real answer to this... its a problem (from a math perspective) until the schedules become more open
 
Last edited:
Re: Statistics Models

Umileated said:
I just got caught up on the NHL 2012-13 realignment. It seems that the biggest complaint of the public is inter-division imbalance w.r.t. playoff structure. I'm inclined to think that having the first 2 rounds take place within your division is a good counter-perk to this grievance.

I'm thinking the imbalance is more an issue when it comes to draft seeding. With a more isolated schedule, W/L/O records are likely going to bring injustice to seeding. the emphasis of the SOS component of KRACH makes me think that it might be a suitable tool for more accurate seeding. But, having not played with the model firsthand yet, I wonder if the team schedules will be varied enough, given that your schedule will be just about identical to everyone else in your division and that every team matches up with every other team at least twice.

What do you guys think? Is there a tool that can do the job better? Will the NHL ever adopt something so "hard to understand?"
Click to expand...

Point blank... hockey is the most resistive to quantitative analysis of all the major sports. (Football > Soccer > Baseball > Basketball > ... > *tumble weed* > ... > Hockey, in that order) I would like that to change... for all I know they had people at the last NESSIS at Harvard. In spirit, you are correct, but you're working against the good ol' boys and waving this weird math stuff in their face.

That being said, my understanding and observations of computer rankings in sports with more regular schedules is that things follow closely but not identically to the standings (neglecting the loss point of the NHL)... you'll tend to get some flips and flops. As long as its not like an AHL schedule it'll be similar enough. Last I heard, is that the schedules aren't that isolated. Further, if you threw something like KRACH out there you may see people trying to game the system through the schedules or forsake any sane notion of scheduling.

Its a cute thing to think about... maybe slightly interesting to analyze... but establishing utility isn't clear. Remember, these are the same people who gave us the loss point because they didn't want anybody feeling bad.
 
Last edited:
Re: Statistics Models

Patman said:
Point blank... hockey is the most resistive to quantitative analysis of all the major sports. (Football > Soccer > Baseball > Basketball > ... > *tumble weed* > ... > Hockey, in that order) I would like that to change... for all I know they had people at the last NESSIS at Harvard. In spirit, you are correct, but you're working against the good ol' boys and waving this weird math stuff in their face.

That being said, my understanding and observations of computer rankings in sports with more regular schedules is that things follow closely but not identically to the standings (neglecting the loss point of the NHL)... you'll tend to get some flips and flops. As long as its not like an AHL schedule it'll be similar enough. Last I heard, is that the schedules aren't that isolated. Further, if you threw something like KRACH out there you may see people trying to game the system through the schedules or forsake any sane notion of scheduling.

Its a cute thing to think about... maybe slightly interesting to analyze... but establishing utility isn't clear. Remember, these are the same people who gave us the loss point because they didn't want anybody feeling bad.
Click to expand...

My understanding from what I've read on the re-alignment is this:
4 divisions of either 7 or 8 teams.
A given team will play every non-division opponent exactly twice - once at home, and one away.
A given team will play every division opponent 6 times (with 7 teams in division), or an annual alternating 5-6 times (8 team divisions).

So the division/non-division schedule weighting will be either 36/46 or 38/44.
 
Re: Statistics Models

Patman said:
Point blank... hockey is the most resistive to quantitative analysis of all the major sports. (Football > Soccer > Baseball > Basketball > ... > *tumble weed* > ... > Hockey, in that order) I would like that to change... for all I know they had people at the last NESSIS at Harvard. In spirit, you are correct, but you're working against the good ol' boys and waving this weird math stuff in their face.

That being said, my understanding and observations of computer rankings in sports with more regular schedules is that things follow closely but not identically to the standings (neglecting the loss point of the NHL)... you'll tend to get some flips and flops. As long as its not like an AHL schedule it'll be similar enough. Last I heard, is that the schedules aren't that isolated. Further, if you threw something like KRACH out there you may see people trying to game the system through the schedules or forsake any sane notion of scheduling.

Its a cute thing to think about... maybe slightly interesting to analyze... but establishing utility isn't clear. Remember, these are the same people who gave us the loss point because they didn't want anybody feeling bad.
Click to expand...

Hockey people complain about math... yet our determination of who makes the college national tournament is the only one that is completely objective...
 
Re: Statistics Models

FlagDUDE08 said:
Hockey people complain about math... yet our determination of who makes the college national tournament is the only one that is completely objective...
Click to expand...

yeah... curious that it evolved like that... part of me wonders if a lot of that has to do with the old east vs. west haziness back pre-widespread internet.
 
You must log in or register to reply here.
Back
Top