| 1 |
We have the tools, so there's no point guessing about whether it's better to have different ELO values for 1v1/2v2-4v4 or to merge them into a single value. I have evaluated multiple possible systems using [url=http://zero-k.info/Forum/Thread/22898]this scoring[/url].
|
1 |
We have the tools, so there's no point guessing about whether it's better to have different ELO values for 1v1/2v2-4v4 or to merge them into a single value. I have evaluated multiple possible systems using [url=http://zero-k.info/Forum/Thread/22898]this scoring[/url].
|
| 2 |
\n
|
2 |
\n
|
| 3 |
== ZK ELO ==
|
3 |
== ZK ELO ==
|
| 4 |
Elo as it is currently implemented in Zero-K except for leaving out newbie malus(which only matters little in the long run). The old prefix refers to using K=32 instead of K=64.
|
4 |
Elo as it is currently implemented in Zero-K except for leaving out newbie malus(which only matters little in the long run). The old prefix refers to using K=32 instead of K=64.
|
| 5 |
\n
|
5 |
\n
|
| 6 |
== Split ELO ==
|
6 |
== Split ELO ==
|
| 7 |
Using two different Elo values, one for games with more than S players and one for the others. Otherwise the same as ZK ELO.
|
7 |
Using two different Elo values, one for games with more than S players and one for the others. Otherwise the same as ZK ELO.
|
| 8 |
\n
|
8 |
\n
|
| 9 |
== Interpolated ELO ==
|
9 |
== Interpolated ELO ==
|
| 10 |
Using two different Elo values and weighting them with S/Players(Max. weight = 1). Suggested by @Brackman
|
10 |
Using two different Elo values and weighting them with S/Players(Max. weight = 1). Suggested by @Brackman
|
| 11 |
\n
|
11 |
\n
|
| 12 |
== Linear Mixed ELO ==
|
12 |
== Linear Mixed ELO ==
|
| 13 |
Using N different Elo values, each one for a specific player count, linearly interpolating them in between.
|
13 |
Using N different Elo values, each one for a specific player count, linearly interpolating them in between.
|
| 14 |
\n
|
14 |
\n
|
| 15 |
== Always 0.5 ==
|
15 |
== Always 0.5 ==
|
| 16 |
Dummy rating system that will always return equal chances. This would lead to random balance.
|
16 |
Dummy rating system that will always return equal chances. This would lead to random balance.
|
| 17 |
\n
|
17 |
\n
|
| 18 |
== Results ==
|
18 |
== Results ==
|
| 19 |
These systems were evaluated over all even 1v1-4v4 battles not played on a funny map or against bots.
|
19 |
These systems were evaluated over all even 1v1-4v4 battles not played on a funny map or against bots.
|
| 20 |
\n
|
20 |
\n
|
| 21 |
Scores:
|
21 |
Scores:
|
| 22 |
{
{
{
Always
0.
5:
0.
0
|
22 |
{
{
{
Interp
ELO
(
S=5,
K=64)
:
0.
1789421646708225
|
| 23 |
Interp
ELO
(
S=5,
K=64)
:
0.
1789421646708225
|
23 |
ZK
Elo:
0.
17810675231268006
|
| 24 |
Lin Mix ELO (2, 8): 0.17601081142739458
|
24 |
Lin Mix ELO (2, 8): 0.17601081142739458
|
| 25 |
Split ELO (S=4, K=64): 0.1711465984015903
|
25 |
Split ELO (S=4, K=64): 0.1711465984015903
|
| 26 |
ZK
Elo:
0.
17810675231268006
|
26 |
Old
ZK
Elo:
0.
16988054006151806
|
| 27 |
Old
ZK
Elo:
0.
16988054006151806}
}
}
|
27 |
Always
0.
5:
0.
0}
}
}
|
| 28 |
\n
|
28 |
\n
|
| 29 |
As you see, splitting the games is actually a bad idea, it reduces the available data to calculate ratings from. The single ELO value is only beaten with little margin by the interpolated one.
|
29 |
As you see, splitting the games is actually a bad idea, it reduces the available data to calculate ratings from. The single ELO value is only beaten with little margin by the interpolated one.
|
| 30 |
\n
|
30 |
\n
|
| 31 |
I have also tried different constants for the systems, but limited it to the best result for each system.
|
31 |
I have also tried different constants for the systems, but limited it to the best result for each system.
|
Predictiveness of the new ELO Split in 1v1-4v4