Thread: GBU: Chokejob
View Single Post
Old
03-16-2011, 12:36 PM
  #99
thefifagod
I'm The Survivor
 
thefifagod's Avatar
 
Join Date: Jul 2008
Posts: 4,042
vCash: 500
Quote:
Originally Posted by jlr View Post
Is it perfect? No. Obviously, it doesn't predict the future - no model can. But what it does do is give you a pretty good idea of what's likely to happen - based on historical data. I think that's relavent, and useful information.



Actually, I think it wrote us off because we flat out sucked, and teams that suck as bad as we did at that point in the season don't usually turn things around and make the playoffs. I could be wrong, but I think they use an extremely simple model that either says we win 50% of our remaining games, or a certain percentage of them based on previous games. I don't think it even attempts to factor in things like injuries, fatigue, etc.

And we don't know that things like that even affect the numbers predictably. We've played better without Roy, the Pens were fine without Crosby until Malkin went down, etc.. So adding in more variables may actually make the percentages less accurate.




I don't want a firesale, and never did. But I think it's downright delusional to think this team is anywhere near the contender they need to be, and to hold onto expiring contracts for players who aren't in our future plans, instead of moving them for things that will actually have value after July 1. I don't think moving out dead weight like Connolly, Rivet (now gone), Lalime (now superglued to the bench), Grier, etc. would have substantially changed our playoff chances. If we manage to get in, then great, it will be good for a lot of the young guys to get the experience. But I'm not willing to sacrifice much, if anything, in the way of future assets to get there this year.

I don't think you can ever predict something as complicated as hockey using any statistical model. But when you have to make decisions - like what to do with expiring UFAs - I think it's useful to know what the past results say about your teams chances. Obviously, knowing your roster, how they play and practice, etc ,is going to be more valuable than raw numbers, but I don't think the historical data is worthless.
Agree with this down the line. This is as accurate as a simulation can be and completely disregarding it is pretty egregious. Sure we could start to struggle but that is accounted for in some of the simulations too. Plus they do adjust the weighted probabilities based on games that have happened and current teams' runs. Perhaps it's because I'm a double mathematics and statistics major who is focused in statistics that I can see the validity of such things, but from my experience this is an extremely well done simulation that certainly has value.

thefifagod is offline   Reply With Quote