I'm currently aware 1v1 elo increases your team elo, but not vice versa, I'm aware 1v1 skill strongly reflects your ability at team games, but I think it should have much less effect on team elo... especially if you never play them. I suggest, if team elo cannot be independant from 1v1 elo, 1v1 should weigh less and to grade anywhere as top team player, you should, at the very least, participate in them... the team ladder would shake up alot i believe.
+3 / -0
|
quote:
if team elo cannot be independant from 1v1 elo |
Technically there is no obstruction for team elo being fully distinct.
+4 / -0
|
Think of it as "1v1" Elo vs "Everything" Elo. I think a good system would be to have distinct Elo rankings for 1v1, Team, and FFA. Where: 1v1 - 2 teams with only 1 player per team Team - 2 teams with at least >1 player on one team FFA - >2 teams with any number of players each I'm being overly specific just to show that those categories can cover any combination of team counts and player counts.
+9 / -0
|
there better team players with lower elo then most high elo 1vs1 players. just saying it a different ball game.
+1 / -0
|
quote:
Technically there is no obstruction for team elo being fully distinct.
|
From this I conclude that "team elo" is not the average of 1v1 elo and a hidden team only elo, but a regular elo that changes with every game you play (what you play more often is "weighted more"). Is it true that the average elo of a team in a team game behaves like 1v1 elo or "team elo" in a 1v1 and that the elo loss or win is distributed equally among the players of a team? This would mean that 1v1s adapt 1v1 elo and "team elo" faster to skill changes than team games adapt "team elo". On the other hand playing team games leads to less fluctuation of elo at constant skill. This kind of stronger weighting 1v1s and small teams is justified because there is more self-importance and changes in the elo distribution among ALL players depend on the number of games. On the whole the current system is not bad (except too high elo fluctuation in relation to skill change rate when playing 1v1). The actual problem is that a player can decide what influence 1v1s and team games have on his "team elo", but skill depends on the game type (at least a bit). Furthermore elo fluctuation varies among players depending on the relations of played game types. Also the players' spread of elo mean values varies depending on game types. Actually you would need an own elo for any specific game type, where every elo also influences elo of similar types, but this would be too complicated, so I suggest tha same as   luckywaldo7.
+1 / -0
|
I think it would be enough to just exclude 1v1 elo from the pool, so 1v1 and 'everything else' rather than 'everything'. But one of you guys is probably going to have to have the will ability and motivation to actually do this, so...
+0 / -0
|
"there better team players with lower elo then most high elo 1vs1 players. just saying it a different ball game." Please send them our way, PRO would love to have some high level team matches!
+3 / -0
|
Personally I think it's genius that 1v1 elo influences teams elo, and teamgames don't influence 1v1 elo, but I agree with 4hundred that it shouldn't be influencing it that much, because i've gotten myself into #44 in team games being really horrible as for top 50. This can be really annoying, because when I go to teams my elo is WAY higher than it should be, resulting in placing me against teamgame veterans who can defeat me easily. I have won 6 games today, in which 2 of them I was performing good, keeping up with my team, while in the rest of them I was kinda carried by 1500-1700 elo players which is definitely not right(I am 1850). Oh, and receiving additional comms and units of missing players while not being the best on the team makes it even worse.
+3 / -0
|
wat spam said i dont like team games, i dont want to play them properly but my elo will never reflect this thanks to 1v1 does more harm than good linking 1v1 to team elo if you ask me
+0 / -0
|
If for nothing else, separate the two (or three if you count FFA as well which is a good idea too) just to make it easier for people to see how well they are doing in each game type. By showing team elo as a combination of 1v1 + team, it's hard to get a sense of how well you play in team games. Not to mention people play very differently in the different types of games so their ranks should reflect that and the balancer should work accordingly.
+4 / -0
|
Yay lets add chicken elo too!
+2 / -0
|
I agree that they should be separate. I am in the teams top20 from playing 1v1. This puts my team in a really bad position because I play like 1650 in teams. Team games are very different to 1v1 in ZK. 1v1 elo shows real game skill, but teams has lots more complications than that such as coordinating a team that won't listen.
+3 / -0
|
|
|
quote:
But one of you guys is probably going to have to have the will ability and motivation to actually do this, so... |
Well, it's not worth even trying until it has an official Licho seal of approval, or overwhelming community support such that he cannot ignore.
+0 / -0
|
quote:
But then you can't balance teams by the "Everything" elo. Thats the wrong way.
Like    TheSponge said. |
I think Sponge actually said the contrary: that one shouldn't include 1v1 data in teams elo calculations. quote:
Well, it's not worth even trying until it has an official Licho seal of approval |
Teams elo vs 1v1 elo has already been split, previously there was just one ladder. I would consider inclusion of 1v1 data in everything-elo a bug, especially since it currently lies* about being "teamgame elo". (* See "Top 50 team players" - that's a lie because it includes 1v1 data). Bugs don't require any explicit authorization to be fixed.
+4 / -0
|
I think changing it to the way Sak suggested would be best now. 2 elos, one for 1v1 and one for everything else. Would be a small step for big gain.
+4 / -0
|
|
|
As far as I am aware the current system is here because it is one of the simplest to implement. Apparently many bits of infrastructure depend on the existence of a single elo number and cannot deal with using different elo types in different situations. For example the Elo displayed in the playerlist in 1v1s is actually everything elo. When 1v1 elo was added (due to overwhelming support) it was patched over the top in a way which works. So I think anyone is free to improve it by splitting the elos completely and adding new categories.
+0 / -0
|
for what its worth, im not a fan of the 'everything else' thing, FFA is very different to teams and 1v1 aswell though a 3rd table is probably too much... but if not, a rolling table of past month would be cool too, it would motivate and enable newer and\or more active players to find themselves in a table.
+0 / -0
|
FFA is not so common that it seriously impacts elo so much, in fact it could be excluded totally since it's mostly for bragging rights (no balance considerations) and bragging rights in FFA is kind of silly. Which, again, I'm totally in support of you going and doing that if you like.
+1 / -0
|
