ELO Discussion

[MrChaos]

This may explain the recent spate of low numbers and such:

http://www.freeallegiance.org/forums/index...ost&p=49373

MrChaos <--- appears to need to play a bunch more games /unsure.gif" style="vertical-align:middle" emoid=":unsure:" border="0" alt="unsure.gif" />

[cuculet]

Definetly something wrong with elo.

In 2 days i was 7 --> 4 --> 5 --> 5 --> 11

Lol. Im not a 11. /blush.gif" style="vertical-align:middle" emoid="" border="0" alt="blush.gif" />


Edit (15 min later)

Just entered alleg and i'm (12) inter 5.

So by tomorow ill be vet 3 ?????? Somebody adjust the elo.

[X_Avenger_X]

Some people care about seeing a number after their names ... others dont.

Personally, I pay little attention to the latest set of (##) after peoples names and still go by who the pilot is regardless of (##).

That is normally a good reliable indication of skill. My advice ? Quit worrying about the silly numbers /wink.gif" style="vertical-align:middle" emoid=";)" border="0" alt="wink.gif" />

[Pook]

QUOTE (bastard @ Nov 21 2006, 10:17 PM) Personally, I think ELO is flawed, well, because I am a (5).

There was a bug in a recent update to the way the new player rank modifier was calculated which has been fixed. Chances are you're no longer a 5.

[SNAFU]

Say (theoretically) there were 2 people with exactly the same level of skill and weren't improving.

One played 14 times per week, the other 7.

Even tho they were as good as one another, the one who played twice as much would have a higher, or lower, elo than the other?

How is this a true representation of skill?

[Pook]

QUOTE (HotBird6 @ Nov 22 2006, 10:04 AM) Say (theoretically) there were 2 people with exactly the same level of skill and weren't improving. One played 14 times per week, the other 7.

Most likely impossible to play twice as much as someone else yet not have more experience than they do, but for the sake of argument, sure.

QUOTE (HotBird6 @ Nov 22 2006, 10:04 AM) Even tho they were as good as one another, the one who played twice as much would have a higher, or lower, elo than the other?

You're confusing ELO and RANK. ELO has absolutely nothing to do with how often you play other than how quickly your rating converges upon it's true value.

I believe that you want to be posting in the thread about the newbie rank modifier.

[SNAFU]

But, but....

I think it's best I just give up trying to understand and crawl back under my stone.

[Raveen]

QUOTE (HotBird6 @ Nov 22 2006, 04:04 PM) Say (theoretically) there were 2 people with exactly the same level of skill and weren't improving.

One played 14 times per week, the other 7.

Even tho they were as good as one another, the one who played twice as much would have a higher, or lower, elo than the other?

How is this a true representation of skill?

Assuming that both players are not newbies then Player 1 will play 7 games, gain some ELO by winning some games, lose some ELO by losing games. Player 2 will play 14 games, gain some ELO by winning some games, lose some ELO by losing games. Assuming both players have hit their final converged ELO rating then they should both fluctuate around the same basic value no matter how many games they play.

[mcwarren4]

Just curious, but could we use all of the old games from the first round of ELO in calculating player skill? Using the new modifiers, throwing out games that are outside of a 70/30 stack limit, we should be able to 'simulate' close to 10,000 games and calculate ranks should we not? I haven't thought it through but on the surface it seems like an idea that might work.

[jgbaxter]

Well, it's possible. The issue with that and why there have been resets is, that the old data is based on players ranks using the elo system at the time, commanders and players were looking at those ranks to be able to balance games. Now since elo has gone through changes, and at times large changes, it becomes more accurate, and the commanders and players make more informed decisions as time goes on and elo gets to a better standard, which therefore means that the decisions previous were slightly less informed occasionally. /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />

So, we could, I suppose, though we'd be using somewhat faulty data in that situation. Mind you, with certain qualifiers as to what games could be kept in the calculations then it's 'possible' with strict restrictions like lowering the threshold for the older games, maybe as low as 60/40. /unsure.gif" style="vertical-align:middle" emoid=":unsure:" border="0" alt="unsure.gif" />

It's a tough call. /blush.gif" style="vertical-align:middle" emoid="" border="0" alt="blush.gif" />