Researchers crack multi-player bidding game complexity
New paper solves decades-old problem in game theory with practical AI implications.
Computer scientists Guy Avni and Fatima Murra have published a landmark paper introducing the first comprehensive study of multi-player discrete-bidding games, a class of graph-based strategic interactions where players compete for token movement rights through structured bidding.
The research demonstrates that under standard tie-breaking assumptions, these games achieve determinacy—meaning from any starting configuration, one coalition of players can always force a win. This breakthrough extends to mean-payoff games and guarantees the existence of pure Nash equilibria in qualitative variants. Critically, the team established that determining winning coalitions in even simple reachability games is PSPACE-hard, contrasting sharply with two-player variants that remain in NP/coNP despite binary budget representations.
- First analysis of multi-player discrete-bidding games by Guy Avni and Fatima Murra (arXiv:2607.27456)
- Proves determinacy and existence of pure Nash equilibria in coalition-based play
- Complexity jumps to PSPACE-hard for reachability games (vs NP/coNP in two-player variants)
Why It Matters
Enables more accurate modeling of multi-agent AI systems where discrete resource allocation determines outcomes.