Every part until now assumed you were the only one deciding. You never are. Clients, contractors, regulators and competitors all make their own choices — and the best move for you depends on the moves you expect from them.
When multiple decision-makers each pursue their own aims, and each one's outcome depends on what the others do, you've left the world of single-agent optimisation and entered game theory. This final part is about that strategic uncertainty, and it starts with the simplest case: everyone decides once, simultaneously, knowing everyone's payoffs.
1The building blocks
A normal-form game lists each agent's possible actions and the payoff each agent receives for every combination of choices. From this a few ideas do most of the work. A dominant strategy is one that's best for you no matter what others do — when it exists, choose it. More often none does, and the central concept is the Nash equilibrium: a combination of strategies where no agent can do better by changing their own choice alone. It's the natural notion of a stable outcome — the resting point once everyone is best-responding to everyone else. Agents may also use mixed strategies, deliberately randomising to stay unpredictable.
2The dilemma at the heart of it
The famous cautionary tale is the prisoner's dilemma, and it's painfully familiar to anyone who has watched a project relationship sour. Two parties would both do best by cooperating — yet each, reasoning individually, is drawn to compete, and they end up in a mutually worse outcome that neither wanted. The tragedy is that the bad outcome is the stable one: it's the Nash equilibrium.
3Reasoning about who you're up against
Equilibrium assumes everyone is a flawless strategist, which real counterparties aren't. So a complementary approach models the other agent's likely behaviour — their goals, their sophistication, how many steps ahead they actually reason — and computes your best response to that model rather than to an idealised opponent. In practice, understanding how a specific counterparty tends to decide is often more useful than assuming perfect play.
When outcomes depend on others' choices, "optimal for me" is meaningless without a model of them. A Nash equilibrium predicts where mutual best-responding settles — and the prisoner's dilemma warns that the stable outcome can be worse for everyone than an unreachable cooperative one.
Treat counterparties as strategic actors, not fixed obstacles: anticipate the client's, contractor's, or regulator's best response to your move before you make it. When you see a relationship drifting toward mutual defection — adversarial claims, defensive posturing, information hoarding — recognise the prisoner's dilemma at work, and remember its escape route: change the payoffs. Contracts, incentives, aligned milestones and repeated dealing all exist to make cooperation the rational choice rather than the risky one.