OR  Zero-K Name:    Password:   

Help! I'm overrated!

66 posts, 1878 views
Post comment
Filter:    Player:  
Page of 4 (66 records)

15 months ago
Until today, I had only played teamgames and chickens and stuff since I started playing Zero-K again.
My casual rating were about 2200-2250 yesterday putting me in the top 10.

But today I decided to play some 1v1 in the matchmaking and this made my casual rating skyrocket to a whooping 2517(!!!) making me the highest rated player in Casual.

Just because I play a few 1v1, that doesn't make me the best team player in the game.
Sigero is currently a much better player in teamgames and I suffer from lag both due to network issues and hardware bottlenecks lategame in large team games.

How can my team game rating go so high when my 1v1 rating is even lower?

It is really backward that I should become the highest rated team player from playing a few 1v1 games when I'm not even the highest rated 1v1 player.

How will this be fixed?

Do I have to use 1 account for 1v1 and another account for team games?
+2 / -0
15 months ago
just keep playing and you will loose because your not the best
+0 / -0

15 months ago
IIRC matchmaker and 1v1 influence casual and teams because, perhaps unintuitively, that seemed to increase the rating system's predictive power.
+0 / -0

15 months ago
Losing isn't fun. Losing several hundred rating from 1v1 games in team games would take like forever.
Also, what will be the actual balance point for the teamgame rating?
Will it actually stabilize once my 1v1 rating is set and after I play ten, twenty, 30, 40 or however many teamgames I need to lower my casual rating?
+0 / -0
"Losing isn´t fun..."

I honestly feel with you.
+0 / -0
Casual rating is general rating, not team game rating. Rating is a lot about statistics and unfortunately there is a very strong correlation between good 1v1 play and good team play, uncoupling team games from 1v1 would make balance worse for most players.

Thanks to your sacrifice, Zero-K lobsters are now winning more than ever!

PS: Don't smurf if you don't want your rating jumping around like a lobster in a hot bowl. There's no way to escape the elo hell that is team games anyway.

Edit: Just to clarify why rating team games independently is pointless. I've compared ratings taken purely from 1v1 games vs ratings taken purely from team games and the 1v1 ratings were better at predicting both the 1v1 and the team games!

When it comes to accurately rating players, there's no point having more than a single, combined rating. Even including games on trollmaps improves the rating accuracy. The "1v1" (Matchmaker) rating is only there as a bragging rating really and doesn't include casual games so you don't have to worry about losing brag points. If MM popularity increases, the rating should become more meaningful. Why do you think there should be a team game rating?
+4 / -1
15 months ago
i jsut want to point out that there is no 1v1 Ranking.
There is only a Casual and a MM-Ranking, both have team and 1v1 games.

What i dont understand however is that those two Ratings are connected somehow.
The ratings seem to miss their real purpose in that way. I mean, when both nearly point out the same, whats the idea of having two then?
+0 / -0

15 months ago
I completely disagree with CHrankAdminDeinFreund. The end goal is not maximum accuracy. As far as I can tell the MM rating uses all MM games for its rating while casual rating uses all games. Casual should just use casual.
+3 / -0

15 months ago
"Why do you think there should be a team game rating?"
Because that team games and 1v1 games requires different skills to some degree.
Optimal play in team games require coordination, team work and communication between players while 1v1 requires none of that. Some skills that are more important in 1v1 is raiding, expanding and trading your units off for momentum and map control, while unit preservation, ecoing, reclaiming and knowing lategame units is more important in team games.

"When it comes to accurately rating players, there's no point having more than a single, combined rating. Even including games on trollmaps improves the rating accuracy."
I don't see how troll maps could increase rating accuracy. If you take maps like Duke Nukem where you can easily lose because of RNG, how will that make rating more accurate? Imagine a map where the only that thing decides the victor is a literal throw of the dice, how can that make rating more accurate?
If you take maps like Speed Metal, Cow or other strange maps, they might still require skill to win but this skillset is different from 1v1.

Now, 1v1 is surely the fastest way to get accurate information about the general skill of the individual player.
After all, if it is a 10vs10 team game, on average, individual players only have about 10% impact on the outcome of the game. It takes many more games to average the skill of the player in those situations.
However, if I have played many teamgames and suddenly jumps far above everyone else just from playing a few 1v1. Then something is wrong.
+0 / -0

15 months ago
Also I have read up a little bit on the WHR and nothing says it is good for rating individual players in teamgames.
If you could elaborate further on that, it could be interesting.
+0 / -0
lel. I predicted this exact scenario.

+0 / -0
The thing is Godde is actually this good, he will have so disproportionate a win rate that he will sit several hundred rank points above anyone else. But he cannot carry a whole team to this level when he has only a tiny fraction of the teams income, his effect over the outcome is limited.
+1 / -0
15 months ago
that was always the destiny of the stronger players
+0 / -0
15 months ago
Godde is doomed. Now his ranking is 2586 :D

You need lessons on being crap, pal.

I give 1 on 1 coaching, and I guarantee after a few modestly priced lessons you won't be nearly so highly ranked ;)
+4 / -0
If competitive and casual WHR were separate, Godde could still use only one account, play somewhat bad in teamgames and the best he can in 1v1, and not have skyrocketing casual whr.
+1 / -0
Different size game leads to different skills.
1v1 is much more a story of micro than cluster game. I do not see much nukes in 1v1.
In fact, I think there are three skill classes:
- 1v1
- 2v2 and 3v3
- 4v4 and more
Therefore, we could have three separate ladders with 1v1 the king of ladders.
But it sounds too messy, so 1v1 and casual ladder is ok for me (as long they are distinct).
+0 / -0

15 months ago
Yeah, just separating the matchmaker rating and the casual rating would make me happy.
+0 / -0
15 months ago
Aren't you talking about multiple issues at the same time?

1) How is the rating computed
Q: is the top 50 casual custom ladder, impacted by "all games that you play" ? (including matchmaker)
Q: is the change Godde has seen due to his 1v1 games or is it due to the daily recomputation (I remember somebody (DeinFreund?) mentioning that once per day WHR recomputes)?

2) How is the balance made in a team game based on rankings
Q: does team size has an effect on balance or just average team rating is considered?

+0 / -0

14 months ago
Slaab imho in 1v1 micro is a skill that is mostly important during the raider phase of the game. After that it´s importance gradually declines while eco, expansion and macromanagment of multiple unit-groups become important.
+1 / -0
NLrankmalric those are good questions. It seems casual rating is impacted by all games (not sure about noelo and silly maps, though). SErankGodde's change seems to be due to the hourly recomputation based on his 1v1 games. AFAIK team balance only uses rating average currently. See this thread about team skill.

SErankGodde your arguments about separating 1v1 and team ratings are valid. And there are also arguments against it. For example, it can be good to separate ratings for players who play 10 games a day, but there are also players who play 2 games per month. If they are played on troll maps, it can be better to include them to everything.

How all those arguments weigh for prediction/balance quality on the whole can only be shown mathematically. I suppose this is what CHrankAdminDeinFreund did, so I trust what he says.

What we can discuss about is the convenience of a system even if that worsens the prediction/balance quality or new systems that could improve quality but are not tested yet.

Furthermore, I think that subtracting one time stdev from mean to get shown rating is enough, rather than two times.
+2 / -0
Page of 4 (66 records)