Multi-Agent Reinforcement Learning on Google Research Football 3vs1
100Goal RateVanilla MAPPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Vanilla MAPPOPolicy Status=Converged2026.03 | 100 | 0 | |
| ETD-MAPPO (Twin-Critics)Policy Status=Converged & Optimized2026.03 | 100 | 5.2 | |
| Static Entropy (No Critics)Policy Status=Total Collapse2026.03 | 0 | 78 |