Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-Agent Reinforcement Learning on Google Research Football (Scenario Win Rates)
Loading...
97
Win Rate: Pass and Shoot
Expert
41.88
56.19
70.5
84.81
Apr 7, 2026
Win Rate: Pass and Shoot
Win Rate: Corner
Win Rate: Counterattack (Easy)
Win Rate: 11 vs 11 (Easy)
Win Rate: 11 vs 11 (Medium)
Win Rate: 11 vs 11 (Hard)
Updated 11d ago
Evaluation Results
Method
Method
Links
Win Rate: Pass and Shoot
Win Rate: Corner
Win Rate: Counterattack (Easy)
Win Rate: 11 vs 11 (Easy)
Win Rate: 11 vs 11 (Medium)
Win Rate: 11 vs 11 (Hard)
Expert
Task Specificity=Singl...
2026.04
97
40
89
99
100
94
MARL-GPT
Training Scope=Multi-E...
2026.04
96
43
89
98
98
68
BC-LSTM
Training Scope=Single-...
2026.04
90
22
88
43
30
24
DT
Training Scope=Single-...
2026.04
80
60
88
0
0
0
RATE
Training Scope=Single-...
2026.04
78
58
85
4
1
1
CQL
Training Scope=Single-...
2026.04
60
30
87
38
35
34
BC
Training Scope=Single-...
2026.04
44
37
86
40
41
40
Feedback
Search any
task
Search any
task
