Initial seed from Week 1 backtest. Player table reflects post-Operation-Epic-Fury reality: US exec + Pentagon + MAGA base anti-ground (low x, high salience+capability) drag equilibrium below midpoint despite Israeli/neocon/Russian pro-ground advocacy.
weekly cyclep = 0.285-0.010
W21 weekly cycle rescore: 4 capability updates applied (GCC +7 restraint leverage, Iran-theocratic +3 cost-imposition, Israel -3, Iran-nationalist -2); dropped the air-strike-tied hawks proposal. Ground position held -- escalation this week was air/missile, excluded by resolution criteria.
gcc.c: 40 → 47Trump publicly stated he called off a planned Tuesday strike on Iran 'at the request of Gulf states' -- a discrete, demonstrated instance of GCC leverage directly altering US military action, exceeding the prior estimate of their clout on this issue.[1]
israel-government.c: 45 → 42Three-quarters of Democrats now oppose aid to Israel and the major European powers are jointly pressuring Jerusalem over settlements/settler violence; eroding US-domestic and allied support marginally weakens Israel's ability to lobby the US toward a ground commitment.[1][2]
iran-theocratic.c: 65 → 68Tehran asserted expanded Hormuz control (a chokehold lever) and explicitly threatened to strike beyond the Middle East if the US resumes attacks -- fresh demonstrations of the asymmetric cost-imposition capability that raises the price of any US ground intervention.[1][2]
iran-nationalist.c: 20 → 18Reporting that a small IRGC-linked fraternity runs decision-making plus a surge of political executions (32+ verified) shows the regime consolidating rather than fracturing, further suppressing the secular faction whose leverage was already contingent on regime fracture.[1][2]
weekly cyclep = 0.295+0.004
Weekly cycle 2026-W20: re-scored after applying 6 cited updates from digest.
us-executive.x: 25 → 28Trump publicly rejected Iran's peace terms as 'totally unacceptable,' said Iran 'will fold,' is weighing restarting Project Freedom, and US forces continued exchanging fire with Iranian units this week. Stance hardened versus the post-Epic-Fury baseline, modestly raising the tail risk of ground escalation if airpower/naval pressure fails -- but the operational mode remains naval/air, so the move is small.[1][2][3]
iran-theocratic.x: 5 → 4Tehran is doubling down on diplomatic demands (sovereignty over Hormuz, sanctions relief, reparations) and refusing to dismantle nuclear facilities, while domestic economic pressure (mass layoffs) intensifies. Their behavior continues to signal maximum avoidance of any trigger that would invite US ground forces. Tiny downward nudge.[1][2][3]
gcc.s: 0.9 → 0.95Former Qatari PM publicly calling for a 'Gulf NATO' and warning the Hormuz crisis is the 'most dangerous fallout' of the war, while BBC reports Gulf economies face decade-long damage. GCC engagement on the issue is intensifying, raising salience without yet shifting their stance against ground escalation.[1][2][3]
china.c: 65 → 67Wang Yi hosting Iran's FM Araghchi in Beijing and explicitly calling for the Strait to reopen, plus a Trump-Xi summit with the Iran war on the agenda, reinforces Beijing's active mediator role. Capability for de-escalation (and thus avoidance of US ground commitment) ticks up slightly.[1][2]
russia.s: 0.7 → 0.75Russia is more visibly engaged this week -- pushing a proposal to remove Iranian enriched uranium, amplifying Iran's negotiating talking points via TASS, and using the Caspian trade route to resupply Tehran. Salience up modestly; position unchanged because the same evidence pulls both ways (wants US drained, but also active de-escalation moves).[1][2][3]
israel-government.x: 80 → 82Netanyahu explicitly refused to halt the Hezbollah campaign for any Iran peace deal and convened a security cabinet during Trump's hardening rhetoric, while Ben Gurion is functioning as a US military hub. Israeli posture is doubling down on a regional reshape that prefers deeper US commitment; small upward move.[1][2][3]
Polymarket resolves YES on a military offensive intended to establish control -- narrower than our 'any attributable ground-combat engagement,' so on definition alone ours should sit higher. On the May strike-escalation news PM jumped from 19.5% (2026-05-11) to ~28-30% ($31M volume), converging on ours (28.5%) and closing the prior ~10pp gap where we sat above it. That convergence is itself the signal: PM's narrower 'invade' question repriced almost to our broader one, suggesting the market is pricing air-strike escalation as ground-invasion risk that our EUM holds out of the ground number (air/missile strikes are excluded by our resolution criteria). PM trades intraday in a ~28-30% band, so treat sub-2pp deltas as noise. Auto-refreshed by engine/benchmark.py.
Resolution criteria
At least 3 of {AP, Reuters, NYT, WSJ, BBC, AFP} report an attributable US ground-combat engagement on Iranian soil by 2026-12-31.
combat
A non-special-forces, attributable engagement with US uniformed personnel either firing or receiving fire on Iranian soil.
ground
On the surface of Iranian sovereign territory (including disputed islands claimed by Iran).
attributable
Either officially acknowledged by the US government or independently confirmed by at least 3 of the source-of-truth outlets.
Excluded:
Drone or missile strikes from offshore or outside Iranian airspace
Naval action in the Strait of Hormuz or Persian Gulf
Special Operations Forces operations not officially acknowledged
Pilot-rescue or other one-off CSAR missions
Cyber operations
Air-to-ground strikes without ground forces inserted
Methodology
Simplified one-issue Expected Utility Model (Bueno de Mesquita). Each stakeholder has a position, salience, and capability; the equilibrium is computed via pairwise bargaining and squashed to a probability with a logistic centered at the issue midpoint.
See docs/theory.md for the derivation and worked example. Score engine: engine/score.py.
Every change to a stakeholder number is tied to one or more event URLs from a multi-bias news digest. Nothing changes silently.