[2205.06968] No-regret learning for repeated non-cooperative games with lossy bandits