New AllegSkill ranks

Allegiance discussion not belonging in another forum.
Correct
Posts: 1046
Joined: Thu May 22, 2008 6:31 pm

Post by Correct »

Sounds like a challenge.
TakingArms wrote:QUOTE (TakingArms @ Aug 9 2009, 07:15 AM) it's interesting how politics turns ordinarily funny, kind-hearted people into vicious, hateful attack mongers. Except IB, he's just always that way.

People just take stuff too seriously I think. Except IB, of course.
zombywoof
Posts: 6523
Joined: Mon Jul 14, 2008 4:59 am
Location: Over the Rainbow

Post by zombywoof »

MrChaos wrote:QUOTE (MrChaos @ Jul 17 2009, 12:28 AM) You actually can't but you think you can, read the book Eyeball. I mean no offense... you know whenever we are on you have to predict the winner at launch every single game your on no cherry picking, I'll keep track and after say 300 games we'll post it, maybe a 100. It will be an interesting experiment
Also highly inaccurate and completely irrelevant unless the teams don't change over the course of the game.
Image
Don't find fault, find a remedy; anybody can complain.
Cookie Monster wrote:QUOTE (Cookie Monster @ Apr 1 2009, 09:35 PM) But I don't read the forums I only post.
MrChaos
Posts: 8352
Joined: Tue Mar 21, 2006 8:00 am

Post by MrChaos »

phoenix1 wrote:QUOTE (phoenix1 @ Jul 17 2009, 02:51 AM) Also highly inaccurate and completely irrelevant unless the teams don't change over the course of the game.
:unsure:

Read the words Im typing: AllegSkills ACCURATELY predicts the winner 80% of the time at game LAUNCH. Much of the other 20% is the ebb and flow post launch IMHO. The system doesn't get to "re-pick"

This is the claim Sir Baker's work is making. WE did throw out a goodish portion of games for being to small, 2 plus teams (the math would make you sob like a hysterical little bitch), those with data collection hiccups, etc etc Im guessish but 8-12%
Ssssh
zombywoof
Posts: 6523
Joined: Mon Jul 14, 2008 4:59 am
Location: Over the Rainbow

Post by zombywoof »

Really. I thought it was predicting the results by seeing the team across the whole game.

That's very interesting.
Image
Don't find fault, find a remedy; anybody can complain.
Cookie Monster wrote:QUOTE (Cookie Monster @ Apr 1 2009, 09:35 PM) But I don't read the forums I only post.
MrChaos
Posts: 8352
Joined: Tue Mar 21, 2006 8:00 am

Post by MrChaos »

phoenix1 wrote:QUOTE (phoenix1 @ Jul 17 2009, 03:41 AM) Really. I thought it was predicting the results by seeing the team across the whole game.

That's very interesting.

Hmmm I meant to edit it, but didnt last night since the read my lips type statement is misleading :blush: and well just never mind :blush:
Ssssh
Adept
Posts: 8660
Joined: Thu Sep 27, 2007 12:53 pm
Location: Turku, Finland

Post by Adept »

Allegskill is measures the skill to win at Allegiance. Much as we'd probably like it to be all about personal skill at playing the game, it's to a large extent also knowing which commanders to back, and whether you're trying to maximise your chance to win when you pick who to fly for.

It's still pretty sweet, and miles better than anything we've had before.
ImageImageImageImageImage
<bp|> Maybe when I grow up I can be a troll like PsycH
<bp|> or an obsessive compulsive paladin of law like Adept
badpazzword
Posts: 3627
Joined: Thu Jan 12, 2006 8:00 am
Contact:

Post by badpazzword »

cashto wrote:QUOTE (cashto @ Jul 17 2009, 02:17 AM) A more serious problem is whether a (21) is worth the same as, more than, or less than three (7)s.
Allegskill does not work like that. AllegR4 uses sum of ranks to get a 'team rank', while AllegSkill uses mu's and sigma's to produce much better matchups (R4 came out earlier). A problem in the not-yet-finished implementation of AllegSkill is getting those numbers to the autobalance portion of Allegiance so that R5 (or R6 or whatever) can bring the full AllegSkill awesomesauce to our games.
Last edited by badpazzword on Fri Jul 17, 2009 11:57 am, edited 1 time in total.
Have gaming questions? Get expert answers! Image Image
finnbryant
Posts: 360
Joined: Sat Feb 17, 2007 11:18 am
Location: England

Post by finnbryant »

so the only way to give comms anything close to allegskills true knowledge would be to replace the team rank with a probability of winning?

eg
yellow team (45%)
blue team (55%)

(ignoring draw probability here)


if im understanding right, that is the only point when allegskills potential can be condensed into a number without becoming blurry, semi-meaningless and difficult to truely understand.

as to the original discussion, why cant people just trust baker? hes proven himself to be far more capable than you at understanding this type of problem, so shutup and let hims do his thing.
Last edited by finnbryant on Fri Jul 17, 2009 2:20 pm, edited 1 time in total.
Raveen
Posts: 9104
Joined: Wed Mar 16, 2005 8:00 am
Location: Birmingham, UK
Contact:

Post by Raveen »

Finn Bryant wrote:QUOTE (Finn Bryant @ Jul 17 2009, 03:16 PM) why cant people just trust baker? hes proven himself to be far more capable than you at understanding this type of problem, so shutup and let hims do his thing.
QFT
ImageImage
Spidey: Can't think of a reason I'd need to know anything
Imago
Posts: 1440
Joined: Tue Sep 23, 2003 7:00 am
Location: Minneapolis, MN
Contact:

Post by Imago »

hey brood you crazy animal.... maybe it has something to do /w the stats not making it to the database.....AFAIK the server triggers events (large, clunky, xml ones) only once and does not retry if it fails.....the way the server "pushes" the stats is retarded. just wanted to say the word retarded, cause when you said it sounded so cool!

baker don't let any of these people slow you down...... :iluv:
Image

These bugs haven't been fixed yet because don't have any developers interested in fixing them up. --Tigereye
Imago's stupid-sensor is supersensitive. --RealPandemonium
The art is managing the flow of the drama to achieve the desired results. --Big_Beta_Tester
joeld wrote:But we’ve been amazed at the level to which some of the Allegiance fans have remained hard-core.
Post Reply