COH2.org ladders: Something fishy here?
Posts: 688
That said, I have been wondering if the ladder position is actually working right.
What got me started was that I was like position 1.600 one day and the next day - without playing at all - I was at 6.000 and something.
I thought that might just be because I had been surpassed by other players who had scored wins in the meanwhile. After all, the player base must be growing (albeit slowly?).
However, yesterday something odd happened. In the - for competitiveness totally unusable - 3v3 category, I was at a non-aweinspiring 98.000-something place(yes, lots of losses due to ppl not ocooperating in automatch 3v3+). My win ration was 42%
Then, after 1 win (yes, it DOES happen), my win ratio was up by about 1%, BUT MY POSITION HAD JUMPED 90.000??
Now, I understand that these ladders use ELO? And that they probably include a lot of players who may not have played more than a few games. Or maybe it even includes players without any multiplayer games in 3v3+ categories at all. So maybe these players are lumped together in the middle of the ladder and not moving ever.
In that case, however, I would expect them all to be standing with either a 50% ratio or with a zero ratio/zero rating. In either case, my single win should hardly let me overtake all these idle players?
Because I assume, that is why I jump 90.000 positions. It can hardly be that as many as 90.000 players have ratings that close that one single win promotes me that much?
A final explanation is ofc, that there is a technical problem with the ladder. I have no idea, I just wonder what is actually going on?
Posts: 1620 | Subs: 2
Posts: 688
After what I read there must be something like just 100k to 150k players, so this is a big deal.
Posts: 16697 | Subs: 12
The big problem that I see now is that some inactive players have extremely high Elo's that are untouchable. That makes the ladder stagnant and no fun. Relic needs to reset and introduce seasons, or introduce some form of ladder decay. Right now the ladders count games from the beta, and show players at the top that haven't played in months... that's no good.
Posts: 688
I've passed 50 games in my most played category. I assume the placement should be narrowed down now?
I think the last 3 wins 'only' moved me about 10.000 positions.
As discussed elsewhere, there seems to be a problem with the way automatcher uses the rating, though. Even in my 'stabilized' category, I still draw teammates with next to no xp. I suppose they just got the game recently and that you don't start over on xp if you rename your account?
Or is it possible to wipe your COH2 user in Steam and start at zero xp, but keep the rating?
In that case, low xp players might make sense in automatching with high xp, but in general I find that the have no clue and it really does take the last bit of balance out of team games to get someone with <20 xp.
Posts: 1838 | Subs: 17
There will be massive swings during your placement matches. That's part of how the Elo system works. You go through a "probation" period (in chess it's usually around 20 games) where it uses a different formula with bigger swings to take you more quickly to your appropriate position. After your "probation" is over, and you are placed generally where you belong, the Elo points gained or lost will be less volatile after each match.
The big problem that I see now is that some inactive players have extremely high Elo's that are untouchable. That makes the ladder stagnant and no fun. Relic needs to reset and introduce seasons, or introduce some form of ladder decay. Right now the ladders count games from the beta, and show players at the top that haven't played in months... that's no good.
Rank 2 as Ostheer 4 life.
Posts: 523
There will be massive swings during your placement matches. That's part of how the Elo system works. You go through a "probation" period (in chess it's usually around 20 games) where it uses a different formula with bigger swings to take you more quickly to your appropriate position. After your "probation" is over, and you are placed generally where you belong, the Elo points gained or lost will be less volatile after each match.
The big problem that I see now is that some inactive players have extremely high Elo's that are untouchable. That makes the ladder stagnant and no fun. Relic needs to reset and introduce seasons, or introduce some form of ladder decay. Right now the ladders count games from the beta, and show players at the top that haven't played in months... that's no good.
This. It also makes us active players less incentive to play. Game is currently unbalanced, people who dont play stay at really high rank..
Posts: 1820 | Subs: 2
Posts: 1708 | Subs: 2
Posts: 33
Posts: 934
For me personally I still frequently go up and down the top 50-200 scale as I play and try out different stuff. I think there is much room to move below the top 50 imo, so the majoirty of people won't really be discouraged. If there was a stat reset I couldn't care less.
I actually thought the majority of people are against that though, I remember Basilone being burn't down about it...
Posts: 1006
Posts: 786
I think the last 3 wins 'only' moved me about 10.000 positions.
As discussed elsewhere, there seems to be a problem with the way automatcher uses the rating, though. Even in my 'stabilized' category, I still draw teammates with next to no xp. I suppose they just got the game recently and that you don't start over on xp if you rename your account?
the upper you go on the ranking and the less positions you will climb per win obviously (think of ELO distribution between players like a gaussian distribution (bell curve)).
In team games your teammates aren't necessarily of your same ELO rating, they could have any. The team gets an overall ELO that's the average of the players', then the automatch looks for a team with a similar average ELO
Posts: 1944 | Subs: 2
can't be bothered to spam games to pass people with lower win % just because they played a lot more than I did.
Posts: 813 | Subs: 1
Posts: 786
can't be bothered to spam games to pass people with lower win % just because they played a lot more than I did.
yeah, that's exactly how it works
Posts: 3293
Posts: 688
In team games your teammates aren't necessarily of your same ELO rating, they could have any. The team gets an overall ELO that's the average of the players', then the automatch looks for a team with a similar average ELO
Evidently what it might be trying to do, but I don't think it makes sense.
Average rating for a team isn't enough. Players also need to be within some range of eachother to make for an effective. 1 player at level 100 (made up numbers) plus two at level 10 isn't as good a team as three players each around level 40.
As it is now, it feels like the automatching for team games has a slipery slope of its own: I can't get decent 2v2 partners because I got into a losing streak initially. I'll be lucky to draw somebody who is active AND knows the basics.
So I keep losing 2 of 3 games and keep dropping in level. I could care less about 3v3+ because that mode is mostly for fun, but it bothers me that I'm in a pit with 2v2 that I can't seem to escape.
Posts: 17
To see advanced, intermediate and beginner ladders and seasons. Once you become a good enough player you get a chance to go to a higher league or season with better players and such i think this would be awesome. But that's just my opinion. If anyone can pull it off its coh2.org
Livestreams
10 | |||||
1 | |||||
1 | |||||
23 | |||||
9 | |||||
9 | |||||
8 | |||||
4 | |||||
2 | |||||
1 |
Ladders Top 10
-
#Steam AliasWL%Streak
- 1.655231.739+15
- 2.842223.791+5
- 3.35458.859+3
- 4.939410.696+5
- 5.599234.719+7
- 6.278108.720+29
- 7.307114.729+3
- 8.645.928+5
- 9.10629.785+7
- 10.527.881+18
Replay highlight
- cblanco ★
- 보드카 중대
- VonManteuffel
- Heartless Jäger
Board Info
12 posts in the last week
26 posts in the last month
Welcome our newest member, LegalMetrologyConsul
Most online: 2043 users on 29 Oct 2023, 01:04 AM