New ELO RatingsTogether with the introduction of the Matchmaking system, the old Teams and 1v1 ELO ratings have been replaced. Instead, we now have a "casual ELO" as well as a "competitive ELO". The casual one replaces the old Teams ELO and will be hidden after the next update. It is used to balance all non Matchmaking games. The competitive ELO takes the place of the previous 1v1 ELO and will remain visible on ladders, in the lobbies and in the game. It is only used and affected by Matchmaking games, even though it is displayed in non Matchmaking games. Rank icons will depict this rating instead of the Teams rating from now on. If you have been using widgets such as the Deluxe Playerlist, be aware that those will now always display your competitive rating, even if that's not the rating used for the game you're currently playing. For Lobby DevsThe only transmitted ELO value is now EffectiveMmElo: 42
+4 / -4
|
quote:
only used and affected by 1v1 Matchmaking games |
Are you sure? I kind of expect it to be affected by all matchmaking games.
+0 / -0
|
or maybe not.. Also I don't really agree with using one value to balance both 4v4 and 1v1, but I'm just delivering the news.
+2 / -0
|
Why stop displaying casual elo in deluxe player list? Nobody complained !! And why make casual elo hidden from own profile??
+5 / -0
|
quote:
It is only used and affected by Matchmaking games, even though it is displayed in non Matchmaking games. |
I hope it's clear now.. quote:
   Connetable
And why make casual elo hidden from own profile?? |
The idea is more like: There is no casual ELO.If you want to increase your skill rating, you need to play "fair" games. Custom rooms allow you to play all kinds of unbalanced games which can screw with elo(1v2, funny maps, boosting/downranking). Casual ELO is just supposed to help balance those games better than using random balance.
+2 / -0
|
Connetable I'm sort of split on this one too.. On one hand newbies can be feeling pressure if they know all games are rated, and allies might be agressive towards them if they know "it matters". On the other hand I also like having some rating in all games I play. In the end I asked people in dev channels and they voted for hiding casual elo.. It can be changed if we decide otherwise.
+0 / -0
|
Maybe this is the time to test whether a game size dependent weighted average of (1v1Elo or smallerTeamsElo) and BiggerTeamsElo yields a significant advantage over just one value for matchmaking..
+1 / -0
|
For the record, Hearthstone does things very similar: There's a casual mode and a ranked one, and while there's no rank (or "elo") indicator in casual mode like there is in ranked, opponents do get notably better after you win a couple of games. There's no mechanical difference in both game types (whereas matchmaking here won't put you into a Trololol game afaik), but I think the solution is good.
+0 / -0
|
Mixing teams and 1v1 elo makes literally no sense. We are regressing. Case study A: Me My teams elo is garbage, because I'm bad at teamgames, but if it used 1v1 elo it would think I'm really good at teamgames.
+6 / -0
|
There were further changes. Site (on http://test.zero-k.info/ not yet deployed) now calculates rank (numerical order) of each active player. All competitive ranks (and elo) are visible. Each player also sees his own casual rank. Top50 shows both top 50 competitive and 50 best casuals. This should be a decent compromise without being too harsh. As for 1v1 and "teams" being mixed together: in MM elo is as much about "bragging" as about balance. Also 2v2 and 1v1 should definitely require very similar skills and player has full control over whether he gets 1v1 or teams.
+0 / -0
|
I am against this change. Please make teams/casual elo visible ingame again. It is not about bragging, it is about knowing which allies you can rely on and which enemies are the most dangerous and which are less. And please keep separate elo for 1v1 and teams if you want balanced games. Otherwise it is too easy to drop your teams elo by playing 1v1s. Which leads to less balanced games later. It is much more important than some special elo for a handful of competetive players who care about ranking. Why change the system that worked well in the first place?
+6 / -0
|
1v1-3v3 games require largely the same skillsets so the same elo ranking for them isn't unreasonable. 4v4 is a bit on the edge, but still fine if balanced by the same Elo as 1v1 imo. To back up my claims I merely have anecdotal evidence. The large majority of my games have been duels (maybe 90%), but I never found it hard to match my teams Elo with my 1v1 Elo (it currently actually surpasses that), since the only occasional team games I played were small teams (2v2-4v4). Clusterfucks are much different imo.
+0 / -0
|
|
|
You mostly play large games though. My claim is that small teams and 1v1 skillsets are largely similar, although I do agree large teams and 1v1 skillsets differ quite a bit.
+0 / -0
|
|
|
quote:
It is not about bragging, it is about knowing which allies you can rely on and which enemies are the most dangerous and which are less. |
Huge Elo gap only exists because of 10v10. In a proper game you can count on everyone because everybody is at ~same elo. quote:
If my 1 vs 1 elo will go 100 lower I will brake the team games balance... |
Clusterfuck still uses casual elo for balancing
+1 / -0
|
I think that as in many changes, often not critical, such as changing the name of Wyvern, would have been the case to ask the player, through a poll, if they preferred to appear on their Elo 1v1 and not as previously. There are players like my case that does not play hardly ever in 1v1 and prefers games in small room. How can I grow my Elo ?. Or frequently organized 1v1 tournaments?
+0 / -0
|
I think that the assumption that 1v1 skill and small teams skill are similar enough to warrant having the same rating is faulty. Speaking for myself, I'm pretty sure I'm a measurably better small teams player than I am a 1v1 player (and measurably better at both than at large teams, for what that's worth) - possibly because my frequently flaky eco skills are challenged less with a team to cover for me.
+2 / -0
|
quote:
Also 2v2 and 1v1 should definitely require very similar skills and player has full control over whether he gets 1v1 or teams. |
quote:
1v1-3v3 games require largely the same skillsets so the same elo ranking for them isn't unreasonable. 4v4 is a bit on the edge, but still fine if balanced by the same Elo as 1v1 imo. |
That's exactly what my idea of game size dependent weighted average of 1v1Elo and teamElo tries to achieve. Here are some example weightings for the linked formulas: game size: 1v1EloWeighting, teamEloWeighting 1v1: 1, 0 1v2: 0.67, 0.33 2v2: 0.5, 0.5 3v3: 0.33, 0.67 4v4: 0.25, 0.75 5v5: 0.2, 0.8 ... quote:
I think that the assumption that 1v1 skill and small teams skill are similar enough to warrant having the same rating is faulty. |
With my 2nd formula, 1v1elo can be separated from smallTeamElo, bigTeamElo: game size: 1v1EloWeighting, smallTeamEloWeighting, bigTeamEloWeighting 1v1: 1, 0, 0 1v2: 0, 1, 0 2v2: 0, 0.75, 0.25 3v3: 0, 0.5, 0.5 4v4: 0, 0.375, 0.625 5v5: 0, 0.3, 0.7 [Spoiler]without 1v2 consideration:
game size: 1v1EloWeighting, smallTeamEloWeighting, bigTeamEloWeighting
1v1: 1, 0, 0
1v2: 0, 1.333, -0.333
2v2: 0, 1, 0
3v3: 0, 0.667, 0.333
4v4: 0, 0.5, 0.5
5v5: 0, 0.4, 0.6 But it's just an idea. Some testing would be good.
+2 / -0
|
  manero if you like small games, you can use matchmaker to get into team games. Also planetwars will work in competitive mode too. Other than that, you can still try to get into top 50 list of casual players, this still exists.
+0 / -0
|
