Part 6 — Coalition mechanics

How to build, maintain, and exit coalitions

Coalition is the primary mechanism for exceeding the individual energy cap. But coalitions are not free — they carry alignment costs, defection risk, and exit complexity. Understanding coalition mechanics is the difference between leverage and dependency.

Theory — Axelrod, The Evolution of Cooperation

In iterated games, cooperation emerges and sustains when: interactions repeat, agents can identify each other, and defection is visible and punishable. Tit-for-tat outperforms pure cooperation and pure defection over time. Coalition stability depends not on trust alone, but on the structure of the game — repeated, legible, and with consequences.

Who to ally with

Ally with agents whose energy is complementary, not redundant. The surplus in a coalition comes from combining different capabilities. Two agents with the same strengths produce less than the sum of their parts — no superadditivity.

Check mode alignment. Coalition between agents in different modes is fragile. A survival-mode agent and a control-mode agent have different risk tolerances, time horizons, and definitions of success. Mismatched modes produce coalition stress even when explicit goals are aligned.

Assess their selection pressure. Allying with an agent under expulsion risk imports their problems. Before entering coalition, understand what the system is already doing to your potential ally.

What makes a coalition stable

Visible contribution tracking

Each agent's contribution is legible to all members. Opacity about who is doing what is the earliest signal of coming defection.

Repeated interaction structure

Coalitions that interact once are high-defection risk. Repeated contact — with memory of prior behavior — is the structural basis for sustained cooperation.

Defined exit conditions

Coalitions without clear exit conditions become traps. Agents who cannot leave cleanly will defect uncleanly. Agreed exit terms, set at formation, are a stability mechanism — not a pessimistic contingency.

Structure at the right threshold

Below ~5 agents, self-coordination works. Above that, informal coordination degrades. Introduce structure — roles, decision rights, escalation paths — before the coalition is large enough to need it urgently.

Detecting capture

Coalition capture is when a coalition you joined begins operating primarily in another agent's interest — not yours. It is gradual and often invisible until significant energy has been transferred.

Are coalition decisions consistently benefiting the same agent?Occasional asymmetry is normal. Systematic asymmetry over time is capture.

Has your exit cost increased since joining?Capture works by raising exit costs gradually — social, financial, reputational. If leaving feels significantly harder than it did at entry, examine why.

Are you being kept in survival mode within the coalition?A coalition that keeps one member permanently resource-constrained is extracting from them. Genuine coalition enables all members to progress through modes.

Example — clean exit

A founding team approaches a natural divergence point — one founder wants to scale, the other wants to stay small. Instead of forcing alignment or waiting for a crisis, they activate the exit terms they defined at formation: clear equity treatment, defined handoff period, agreed communication to stakeholders. The coalition dissolves without destruction because the exit structure was built before it was needed.