- Content map: SMU H3 Game Theory Map
Chapter 8: Strategic Moves
Strategic Moves
Core Idea
Definition:
A strategic move is an observable and irreversible action that changes the game in order to manipulate later play to one’s advantage.
Requirements
- The move must be observable by other players.
- The move must be irreversible or costly to reverse.
- The move must be part of the game description.
- The move must change incentives, not merely announce intentions.
Insight:
Strategic moves are useful because they change what is optimal later.
Credibility
Threats and Promises
- A threat punishes an unwanted action.
- A promise rewards a desired action.
- Both matter only when the other player believes they will be carried out.
Definition:
Credibility means the threatened or promised action is sequentially rational in the relevant subgame.
Connection to SGPE
- A Nash equilibrium may rely on a non-credible threat.
- SGPE removes threats that are not optimal when the subgame is reached.
- A strategic move changes the game so that the threat or promise becomes part of an SGPE.
Result:
A credible strategic move makes follow-through optimal, not merely announced.
Reputation and Commitment
Reputation
- Repetition allows players to build reputations.
- Reputation may require costly short-run actions.
- The cost is paid now to affect future beliefs and future behaviour.
Commitment
Definition:
Commitment is a strategic move that removes future choices or changes future payoffs.
- Removing choices is a way to bind future behaviour.
- Changing payoffs can make enforcement optimal.
- Commitment may require adopting an action that was dominated in the original game.
Insight:
Commitment can be valuable because it restricts the player who commits.
Credible Enforcement
Implementation
- Set clear rules before the relevant subgame.
- Make rules known to all players.
- Enforce the rules when the triggering event occurs.
- External rules or institutions can make enforcement easier.
Result:
Credibility comes from the ability to stick with the rule when backing down would be tempting.
Doubt About Rationality
Strategic Irrationality
- A player can benefit when others assign positive probability to a stubborn or irrational type.
- If the probability is high enough, opponents may concede.
- Behaviour that creates sufficient doubt about rationality can be a strategic move.
Insight:
Reputation can work through beliefs about both preferences and rationality.
Repeated Punishment and Threats
Punishment Strategies
- Repeated games can sustain cooperation by threatening future punishment.
- A punishment phase may last for periods after a deviation.
- The punishment counter can reset if another deviation occurs during the punishment phase.
Threat Design
- In SGPE, credible threats are usually not implemented because the target avoids triggering them.
- If implemented, the threat can be costly to the threatening player.
- Excessively harsh punishment can be less credible because it is too costly to carry out.
- Example of an excessive threat: refusing future defence if the other country does not open its markets.
Result:
Threats must balance deterrence against the cost of actually enforcing them.
Brinkmanship and Salami Tactics
Brinkmanship
Definition:
Brinkmanship makes a very large threat look possible rather than certain.
- The threat is not a certain decision to punish.
- Instead, the player creates a risk of escalation.
- This can make a large threat more believable.
- Example: relations deteriorate until political constraints may block future assistance.
Salami Tactics
Definition:
Salami tactics use small steps toward non-compliance while delaying or blaming external constraints.
- Each step is small enough to look tolerable.
- The cumulative effect can be substantial.
- Vague boundaries make enforcement harder.
- Example: accept market opening in principle, then delay implementation by citing political constraints.
Insight:
Brinkmanship works by increasing perceived risk; salami tactics work by shrinking each apparent violation.