AllegSkill Online

Plaything · Post by **Plaything** » Tue Jan 20, 2009 5:40 am

QUOTE However, you can 'force' it up, in essence, if you do things like fly around in circles one game and then go out and kill everything you see the next.[/quote]

THATS why I went UP!

Zruty · Post by **Zruty** » Tue Jan 20, 2009 8:40 am

Hehehe, had a laugh this morning.
HotNoob, maybe you are missing a tiny bit in your calculations? Easily can happen with such an enlightened person

POINT 1.
Well, let's take X = (X1, X2, ..., XN) - a series of measures, like in your example.
E(X) is defined as (X1+X2+...+XN)/N. That's what is present in the example, and it is called expectation.
Sigma (standard deviation) is Sqrt[ E( (X - E(X))^2 ) ]
You suggest to calculate Sigma as SigmaNoob = E(X - E(X)). But you can easily see that SigmaNoob = 0 ALWAYS! Check it yourself if you want

POINT 2.
You are assuming that each player gets some amount of points after each game, and Mu and Sigma are based on the distribution of these points. You are wrong. AllegSkill system doesn't involve any points in any way: the only object it operates with is whether a player's team WON the game or LOST it.
This way a player can't SUDDENLY improve a TON. He can start winning all of his games, that'll be 'perfect play'. If, as an addition, he always antistacks, i.e. causes the 'weaker' team to win all the time by playing for it, his rating will go up rather fast.

Spinoza · Post by **Spinoza** » Tue Jan 20, 2009 3:34 pm

Oh, I so hate to complain about what in general a huge improvement and massive coding project but I must.

My numbers are mu=17.43 sigma=2.23 rank=6.4
This means that though I'm an experienced, squadded player and an anti stacker (-8) I'm now basically stuck with a number after my name.

I'm not arguing with my rank ... 6 is a great number, but must I have it follow my callsign ?
Could we not make an exception for players with a low sigma ?

Bathawk · Post by **Bathawk** » Tue Jan 20, 2009 3:56 pm

Clay_Pigeon wrote:QUOTE (Clay_Pigeon @ Jan 20 2009, 03:23 AM) *watches hotnoob calculate standard deviation*

*washes hands*

Zruty wrote:QUOTE (Zruty @ Jan 20 2009, 09:40 AM) [...]SigmaNoob[...]

Maybe someone could change hotnoobs callsign for a day or two?

Post by **pkk** » Tue Jan 20, 2009 4:47 pm

Tigereye wrote:QUOTE (Tigereye @ Jan 20 2009, 02:21 AM) Okay I think I've fixed the Opera bug. Someone with opera, sort the columns and let me know if the "order" value is properly appearing in the "order" column instead of the "ranking" column as before

It's not fixed, but I don't care...

sgt_baker · Post by **sgt_baker** » Tue Jan 20, 2009 4:47 pm

Spinoza wrote:QUOTE (Spinoza @ Jan 20 2009, 03:34 PM) Oh, I so hate to complain about what in general a huge improvement and massive coding project but I must.

My numbers are mu=17.43 sigma=2.23 rank=6.4
This means that though I'm an experienced, squadded player and an anti stacker (-8) I'm now basically stuck with a number after my name.

I'm not arguing with my rank ... 6 is a great number, but must I have it follow my callsign ?
Could we not make an exception for players with a low sigma ?

I think there might be a problem with the newbie number thing. Newbness is defined by sigma, not absolute rank. I'll look into it.

spliffyhat · Post by **spliffyhat** » Tue Jan 20, 2009 4:58 pm

juckto wrote:QUOTE (juckto @ Jan 19 2009, 06:47 PM) Do they speak English in What?

Badp wrote:QUOTE (Badp @ Jan 19 2009, 07:08 PM) Going to university isn't enough to understand those mere ten pages, trust me.

don't know what's in that link of yours Badp , three times I clicked it and 3 times it crashed FireFox.

MrChaos · Post by **MrChaos** » Tue Jan 20, 2009 6:01 pm

Badp wrote:QUOTE (Badp @ Jan 19 2009, 07:32 PM) * lot's of caramel goodness inside *

So your "rank" increases all the while your Mu can be going to it's proper spot in the nether regions. As must have been said 120,000 times: Sigma is the uncertainty of your placement inside the population of those playing Allegiance, the Mu is where you place. The underlying assumption of the PDF, and CDF is it normal distributed and the the fact that making changes to the little bits of data placement inside the population that are each of us which doesn't destroy the assumption is the magic behind it all.

Baker Needs a Life Fun Fact: FIVE, yes five seperate and distinct systems were looked into before settling on TrueSkills as the basis for AllegSkills.

Clay_Pigeon wrote:QUOTE (Clay_Pigeon @ Jan 19 2009, 07:42 PM) hotnoob,

frankly, your suggestions make no sense, and imply that you really haven't paid too much attention to the articles. You're also talking about things like replacing variance with absolute value of sigma and implying that sigma can be negative (it can't). This suggests that you haven't done a lot of work with statistics either. On top of everything, you're being condescending towards people who have spent a lot of time with this material, and frankly making an ass of yourself in the process. I recommend stopping while you're ahead.

-T

(who has taken plenty of math courses thankyouverymuch)

edit

never mind.... sigma can be subtracted subtracted not negative you chowderhead MrChaos. $#@! me I was thinking plus/minus about a two tailed test example and not the "negative" bit.

Correct wrote:QUOTE (Correct @ Jan 19 2009, 09:48 PM) I think I understand Hotnoob, I'll try and translate for him.

Okay. Example.

A bad player, lets call him SirChaos, sucks badly at the game. He changes to a mouse and gets much better. As he continues to win games AS thinks he should lose his sigma goes up along with his mu, causing no conservative rank change until the sigma starts going back down as AS expects him to win.

Hotnoob just suggested fixing it so the conservative rank will go up straight away instead of delayed.

Hotnoob also seems to think people are leaving because they are ranked lowly [and probably correctly]

He is of course wrong in that point, only he cares enough about his rank to leave.

SirChaos sucks!!!!! The bit about 25/300 keeps your sigma, while not daisy fresh, at least as not as withered as 90 year old nun's.... where was I.... and allows the system to recognize the fact quicker that SirChaos is now almost to voob level seemingly over night rather then his usual suckage of epic proportions. In the end of you get better the system reacts to it. The bit was injected to make it snap to quicker but at the expense of some very minor artifical uncertainity. Horse trade if you follow. Eyeball

on me for the above post made me really laugh out loud

badpazzword · Post by **badpazzword** » Tue Jan 20, 2009 6:55 pm

spideycw wrote:QUOTE (spideycw @ Jan 19 2009, 07:36 PM) I notice I have 0 Command Draws listed but I can think of at least 4 off the top of my head. What gives?

Draws have been discarded because they really were #abandon's. When AllegSkill sees a draw it understands the sides are "even", which isn't really true in Allegiance: most of the draws are such ludicrously stacked, complex numbers would have been required to fully compute them.

AllegSkill 2 - The Complex Plan. Coming to a ranking system rotator near you

Dogbones · Post by **Dogbones** » Tue Jan 20, 2009 7:25 pm

hotnoob wrote:QUOTE (hotnoob @ Jan 19 2009, 11:35 PM) i have a nice IQ of 157
just to let ya no.

Oooh, oooh, I like these sorts of comments. My IQ is in the 172 to 176 range and I have an MD/PhD. I also have a very large shoe size...

Last I checked none of these has anything to do with saying something or appearing stupid. It does mean I am still paying off student loans even with 75K worth of grants/scholarships.