# Truthtrance: compound shipment promises 14 September 2026. This extends the [reserve-shipment prototype](TRUTHTRANCE_SHIPMENT_PROMISES.md) with shipment-only AND/OR questions. Public mode and publication gates remain unchanged. ## Meaning and authority The [GF9 November 2020 FAQ, printed page 8](https://www.gf9games.com/dune/wp-content/uploads/2020/11/Dune-FAQ-Nov-2020.pdf) permits AND/OR questions and binds definite answers while compliance remains possible. Each condition here describes the same eventual reserve shipment in the active opportunity, using a destination and minimum physical force count. This does not promise two shipments or combine present facts with future actions. AND is true only if every condition matches. OR is true if any condition matches. A No answer negates the whole expression: No to AND requires at least one false condition; No to OR requires every condition to be false. Skipping shipment makes all these positive shipment conditions false. Separately answered questions must all be honored together. For example, Yes to “at least six to Carthag OR at least four to Arrakeen” permits either branch. No to that same question permits fewer than six to Carthag, fewer than four to Arrakeen, another destination or no shipment. AND with two different destinations cannot be fulfilled by a single shipment, so Yes is unavailable. ## Connected prototype The question form adds a second destination/count condition and an AND/OR choice. The shared engine expression supports grouping within software limits of sixteen leaves and four nested groups; the first visual builder exposes two conditions. Existing flat questions and saved promises retain their meaning. Public question/history text preserves grouping, while answer feasibility and suggested preparations stay private to the respondent. The completion search considers every relevant destination, physical count and legal sector, with the same controlled preparations and funding as single promises. It evaluates the entire expression and all earlier promises together. Actual shipment, skipping, voluntary spending and opposing-effect release use the same evaluator. A fulfilled promise records the shipment event and survives later movement or casualties. All four AI profiles rank the projected feasible answers using only public destination value and their own reserves. Strategic preference never determines whether an answer is possible. The existing complete continuation remains the fallback when a strategic shipment shortlist misses a legal route. ## Evidence and remaining scope Focused engine, controls, AI and authenticated SQLite cases cover whole-expression truth, alternative destinations, old saves, joint commitments, custody, private restoration and duplicate actions. Final results and source fingerprints belong to the private checkpoint report and Git message. The isolated browser used genuine base setup before conserved turn-two staging. The human asked “twenty to Carthag OR four to Arrakeen.” The respondent's legal AI policy answered Yes, and the public grouped question/answer survived refresh. Its next AI action shipped four to Arrakeen, spent four spice and fulfilled the aggregate promise once. All six accepted actions replayed exactly; all private views, card/force custody and the original seat row were preserved. This is a targeted browser scenario, not a complete human game. The existing active-unused-shipment scope remains: base Basic, and supported base Advanced games without Guild or optional modules. Earlier timing, mixed fact/future predicates, expansion shipment combinations and arbitrary prose remain unfinished. This prototype does not settle pending material rulings or certify complete Truthtrance. The opening backup preserves 1,004 local rooms and their seat/recovery records. Reuse the healthy server and keep all subsequent games.