User Pass
Home Sign Up Contact Log In
Forum > Oceania AAA Leagues > Oceania AAA #2 > Week 2 Computer Rankings (Alpha only atm)
rutgar
offline
Link
 
ALPHA CONFERENCE

1. Outback 0.8183
2. Grinnell 0.73665
3. Brisbane 0.71839
4. Auckland 0.69382
5. Pittsburgh 0.6718
6. Aberystwyth 0.57969
7. Samoa 0.5685
8. Western 0.50136
9. Launceston 0.48149
10. Villanova 0.46048
11. Los Angeles 0.4257
12. Suva 0.32012
13. Tahiti 0.30515
14. Oregon 0.26704
15. Oberon 0.25722
16. Hollywood 0.1943


ZETA CONFERENCE

1. Fiji Stingers 0.754
2. Cook Island Convicts 0.734
3. Tasmania Dragons 0.691
4. OffTopica Forumites 0.688
5. Honolulu Skulls 0.685
6. Sydney Pelicans 0.672
7. Palau Princes 0.668
8. Collector Planet Galaxy 0.505
9. G-Town Connection 0.502
10. Dampier Devils 0.33
11. Charm City Headset Jockie 0.328
12. Delran Mankats 0.323
13. Cairns Fire 0.307
14. Alaska Wild 0.305
15. USAORG Travelers 0.266
16. Buen Retiro Longhorns 0.243
Last edited Aug 6, 2008 11:31:38
 
rutgar
offline
Link
 
1. Why is Outback #1 and Grinnell #2 when most rankings have Grinnell #1 and Outback around #4?

Outback is 2-0 with both wins coming on the road. Additionally, both teams that Outback beat won the other game they played so both have records of 1-1 with their only loss coming to Outback. Meanwhile, Grinnell is 2-0, but with one win at home and the other on the road. Additionally, Grinnell's second opponent, Suva, lost their only other game. An extra road win and competition with better records h

I believe all other cases where there is a distinction between teams of the same level can be explained in much the same fashion.

I expect that over the course of the season, these rankings will look a lot like the standings. This is because every team plays every other team, except for the last game of the seaon. Additionally, I don't consider these rankings to be a "Power Rankings." To me, power rankings put more weight in how a team has done recently than games that occured months ago while these rankings count every game equally. I'm working on a computation power ranking and will post it when finished.

Additionally, I will get the Zeta conference rankings posted later today. At this point, the only part of the process I've coded is the part that solves the matrix, I'm still entering results manually. I'm hoping to be able to parse scores from the various team pages and completely automate the process, I just need to find the time.
 
rutgar
offline
Link
 
The problem with any computer ranking is that the computer cannot see what happens on the field and cannot take into account things such as sloppy play, ill-timed turnovers, etc. Some of the things that can go into a computer ranking are wins, losses, location, and margin of victory. These are the items that factor into my rankings.

I use a matrix to compute rankings. I cannot begin to take full credit for the idea, that goes to Dr. Wesley Colley who developed Colley's Bias Free College Football Ranking Method, one of the rankings used by the BCS ( http://www.colleyrankings.com ). There are some difference to my rankings and his. I include bonuses and penalties for winning on the road versus losing at home, and have a scaled bonus/penalty for margin of victory/defeat.

There are some basic assumptions used in my computer rankings. First of all, what happened last year means nothing, the current system is viewed as a closed system. Everyone starts off equal. There is an assumed homefield advantage, which is why there is a bonus for winning on the road and a penalty for losing at home. There is no bonus for winning at home or losing on the road. Bonuses/penalties for winning/losing cap at >30 points.

The reason the BCS doesn't release the computer portion of it's rankings until the middle of the season is that a certain sample size is required for computer rankings to really make sense. I'm posting mine now because I want to get them out there and maybe they will generate some discussion, but I've also explained why certain things are what they are.
Last edited Aug 6, 2008 11:00:24
 
rutgar
offline
Link
 
For those of you that read the forums more than I do...

Has it ever been stated how home field works in this game? Is it an advantage? Is it something that factors into outcomes?
 
Intricate08
offline
Link
 
Originally posted by rutgar
For those of you that read the forums more than I do...

Has it ever been stated how home field works in this game? Is it an advantage? Is it something that factors into outcomes?


No.
 
Zurai
offline
Link
 
Home field means nothing in the game, other than deciding who gets the revenue from the game.
 
rutgar
offline
Link
 
Thanks for the info. I'll make adjustments tonight.
 
trojans4me
offline
Link
 
I would like it so that home field would cause visiting teams to commit more false starts or waste TO's due to crowd noise. It would add some realism to the game.
 
Ruger7mmmag
offline
Link
 
Nice, we're not last!
 
mzsleeper02
offline
Link
 
Nice work. I do wish home field advantage counted for something.
 
Phantazm
offline
Link
 
This is a cool thread, thanks for the effort!
 


You are not logged in. Please log in if you want to post a reply.