That is quite intersting about the colors. I didn't expect it hit home so much. Lol
Another one I like to ask people is why the crave a pizza at dinne time, instead of a cheeseburger, on any given day?
While one part of me questions why I like what I like, another part of me thinks to myself, why waste time and just get busy enjoying it.
>>I don't get why you seem to have such a strong belief that you can create a better system than already exist.
A quick note on the issue of faith 
No, actually, on the issue of 'already exists'. Obviously the system doesn't *already exist* or the discussion would be very different. What is going on is that you/we are trying to adapt a system that already exists to a very different playing field, vs creating a new one.
One problem with the 'consider this six games' theory, altho I do think it is best, is that you aren't playing six games against the three other players. You are, in fact, playing one game against three other players. Thus what is really going on is a bizarre combination of 'how good you are versus how good the others are on average' AND 'how good you are versus how good the others are *in total*'.
As we all know each player needs to keep an eye on all other players, and when one good player goes after you even a very weak other player can hop in and get some points against you. Points that, in one sense, they didn't earn, since it wasn't their great play that opened you up, but the great play of another player.
However, the shoe on the other foot, a weak player will not tend to see these opportunities and take them... thus justifying his weak rating.
All in all I think a good discussion is in order while the current system continues. However would it be possible to code in another system as well and compare how well it predicts the games? Not use it for matching, but just for seeing if it is a better predictor. Because, in the long run, that is what we are looking for, no? A system that can fairly accurately predict how player one will do against player two.
Using elo this way with multiplayer games has been done before. It is not a new concept. I do think this game allows more for cooperation between players and it is easier to play against a specific player than most multiplayer games though.
I think you make some valid points @VAOhlman and it could be worth discussing. One thing that can be done is to make ratings change less when there is a big rating gap between the players. The thing is though that what Christopher_Parsons was suggesting was not about issues with the multiplayer format, but just being against elo in general. I am also really not against taking an existing system and find some improvements to it or modification to fit what you want with it better. But making a better one from scratch is a different story.
I think the approach of taking the game as 4 individual games is flawed because it doesn't reflect the nature of the 4 player game. Some examples:
- A game has 3 high rated players and 1 much lower rated. The high rated who has the low rated in front gets destroyed by the 2 on his sides. He is considered to have lost directly to the lower rated and loses a lot of points, but in fact the game might have even been worse for him than if the one in front had been higher rated.
- Again a game with 3 high rated, one much lower. One of the high rated plays a great game and manages to kill the other 2 high rated, but this costs him most his pieces and he gets second (and lets say the way the game went, this was his best try at getting first). He still loses points because the individual loss to a much lower rated opponent counts more than the 2 wins.
- A high rated plays against 3 much lower rated. Here any position other than 1st loses him a lot of points, yet in this game getting first doesn't depend fully on you even if you are much better.
My first idea to improve it would be these 2 changes:
1. You consider each player to have played 3 games against the average of the other 3 opponents, instead of their 3 individual ratings.
2. On top of that, you assign a small amount of points to the finishing positions in the game, independent of the ratings of the players. For example +10 +5 -5 -10
I think the approach of taking the game as 4 individual games is flawed because it doesn't reflect the nature of the 4 player game. Some examples:
- A game has 3 high rated players and 1 much lower rated. The high rated who has the low rated in front gets destroyed by the 2 on his sides. He is considered to have lost directly to the lower rated and loses a lot of points, but in fact the game might have even been worse for him than if the one in front had been higher rated.
- Again a game with 3 high rated, one much lower. One of the high rated plays a great game and manages to kill the other 2 high rated, but this costs him most his pieces and he gets second (and lets say the way the game went, this was his best try at getting first). He still loses points because the individual loss to a much lower rated opponent counts more than the 2 wins.
- A high rated plays against 3 much lower rated. Here any position other than 1st loses him a lot of points, yet in this game getting first doesn't depend fully on you even if you are much better.
My first idea to improve it would be these 2 changes:
1. You consider each player to have played 3 games against the average of the other 3 opponents, instead of their 3 individual ratings.
2. On top of that, you assign a small amount of points to the finishing positions in the game, independent of the ratings of the players. For example +10 +5 -5 -10
+1
I also think this concept can be carried over into a hypothetical pointless version, created to subvert the problems with the current system. Even where only mates and resignations or draws affect ratings, counting this as three individual games played at one time, don't truly do justice to performance or result. You can add collusion to the list of problems with the current system. If a well known titled player isn't dared to be ganged up on by others, but they quickly attack any nobody, it creates an anomaly that the current system has no way of addressing either. This is a special type of game and needs a very specific and special type of rating system, that is custom crafted to account for it's particular nuances.
I think the rating system should be restructured to not reward you for coming 2nd or 3rd, but only grant points for wins.
There are games where a player plays well but get checkmated early. Anyone who just survives should not get free points just because they managed to somehow wait long enough for the eventual winner to claim win.
Another example is where one person offers to team up completely with another player. (Has happened in one game with me yet). And then the two of them kill the other two players for free rating points. (They are guaranteed to be 1st and 2nd, which almost always increases ratings).
This should not be possible under any circumstance. Even /stop-chat cannot stop it because you would have to disable chat "before" the first person offers the alliance.
I think all of this is avoidable if the rating system is simplified. If you come 1st, you win the rating point equivalent that Elo should give. If you don't, then you lose a fraction (1/3 say?) of whatever you would otherwise lose.
Anyone else notice their rating change twice after a game? It does not seem possible to start a new game while the people that beat you are still going (any longer!), so that can't be it. Also, any chance of a points refund for losing to He Who Plays One Game Using Two Accounts?
As an example I finished 2nd just now and my rating climbed to 1399 [I had a bad day yesterday]...and a few seconds later I got a message saying 'your rating is now 1419.'
how would you calculate the rating if each player was 1200?
teams: why does the higher rated winner not gain fewer points than his lower rated partner?
if opponents avg rating is X, shouldn't rating gain depend on your rating vs X? and be different for each player?
FFA: consider 6 games
Teams: consider 4 games
players 1&2 beat 3&4:
- 1 v 3 1:0
- 1 v 4 1:0
- 2 v 3 1:0
- 2 v 4 1:0
no need for 1 v 2 and 3 v 4
and sum up rating gains/losses as in FFA
Both teams are calculated based on average rating, so therefor both teammates will gain/lose the same amount of rating points. If your team is expected to win 60% it does not matter if you are the higher rated or lower rated player.
If the average rating is the best measure to use is a different question.
is nathan on to something here or not? he says
it's interesting that it takes average rating .. and gives both same amount ..
..
if .. played .. down to 500 rating, then could boost partners rating.
..
you dont gain points based on your own rating.
think about what happens when say you have a team A, a 2600 GM with a 1400 amateur, vs team B, two 2000 rated players.
If they were to play a regular game of chess, i would guess team A's strength around 2600, and team B's maybe 2100?
in 4p teams, i think team A should be a clear favourite most of the time. (depending on the extent of the amateur's ability/willingness to cooperate etc, team A's strength could be anywhere from 2605 down to 1400, while team B is likely to perform around 2000-2100 almost always)
say team A wins, who deserves how much?
treating it as 4 games is i suggested above fails because imo the 1400 guy does not deserve a huge rating boost to the tune of beating two 2000s.
using average, it's 2000 beat 2000, so i'm guessing +9 for A, -9 for B
hard to say what's fair. ideally somehing like
2600 gets a win over 2100, say +4
both 2000s get a loss to 2600, say -3 each
the 1400 should get +4 also?
for what it's worth, if i were to bet money (along with the players involved, to ensure the patzer's interest in cooperating with his GM partner 
), i would base my bets on (gut math)
2000+2000 = 2100,
1400+2600 = 2605,
1600+1800 = 1850.
i'm finding it hard to come up with something reasonable when the rating gap within a team is large.
I think the most extreme solution is to use the highest rated players rating in each team for the calculation (instead of average). This encourages teams of players of similar strength. Of course it would make random partners more painful and more rating points at risk for higher rated players.
One thing worth noting is that it is a lot easier to cooperate in a variant like this compared to bughouse, so the strength of a team can more easily move towards the strength of the highest rated player.
Matethief is right. I suggested the same idea to ignoble
The rating system has been updated.
( Top rated player in team * (2/3) + lower rated player * (1/3) ) / 2
2. The corresponding team ratings are used to determine the difference given to the winners and losers.
My favorite colour is yellow. It has the wavelength of light that pleases my eyes the most. ok, not really, my parents choose a colour for each kid to make it easy to differentiate clothes, toothbrush, towels and whatnot and the colour for me was yellow. As a teenager I got really addicted by the colour and now basically all my T-shirts are yellow and if I get to choose a colour I almost always choose yellow. Since I am the one most addicted to my colour I always get that colour when playing board games. All my friends and relatives knows this, so while typically the youngest kids get to choose their colour in board games first that is not the case when I participate since I need the colour yellow. It is part of my identity at this point. It is also very fitting for my profile picture to be a pokemon that is yellow.
Did I get sidetracked? Sorry, I am very passionate about my favorite colour. I will not try to convince others that it is the best colour since I do not want to fight for getting the right colour in games. It is great that there finally is a chess variant where I can play with yellow pieces. I actually made a chess set with pieces in wood when I was about 11 years old that I painted yellow. My friend made his pieces green and then we could play each other between classes.
As a side note I have two younger brothers that are twins. They got the colours blue and red. If they switch shirt with each other it is very easy to mix them up and think that one of them is the other one. While they are not as addicted to their colour as I am it is still sort of part of their identity. This was especially true when they were younger, now they look more different to each other.