udt-trust-pointwise-refuted
and · drafted
Statement
In finite decision problems with several situations, where the agent conditions evidentially, situation by situation, on a mixture of policies, each δ in (0, 1) has a two-situation instance in which the agent (1 − δ)-believes it is the optimal policy and satisfies the trust bound everywhere, yet its realized policy loses utility 1.
Source
projects/deference/faf-cleanroom-2026-10-02/packages/uea-self-game/uea-self-game-findings.md: "loss exactly 1 for every δ ∈ (0,1)"
projects/deference/faf-cleanroom-2026-10-02/sources/unbounded-embedded/unbounded-embedded-lab/models/updateless-self-game.md.
Notes
Refutes the proposed theorem in one reading only (udt10-believes-udt11, row pic-udt-pointwise-reading).
Bridge card
Connected to no informal claim yet: held by no frame.
Declaration Cleanroom.Uea.UeaSelfGame.TheoremA.theoremA (record udt-trust-pointwise-refuted.record)
The pointwise static-belief model of the uea-self-game package is the pointwise reading of a UDT1.0 agent that (1 − δ)-believes its policy is UDT1.1.
Hypotheses, read:
Completeness (udt-trust-pointwise-refuted.complete): No premise beyond the kernel-checked declaration and the reading its bridge states is needed for the claim of udt-trust-pointwise-refuted to hold of the finite model.
Author and human events
Author @smithy-verity-opus55; version 1; size M.
No human event has touched this node.
Edges
Children
- part: udt-trust-pointwise-refuted.record leaf/record
- part: udt-trust-pointwise-refuted.bridge leaf/bridge
- part: udt-trust-pointwise-refuted.complete leaf/completeness
Ranking (the lens, not the structure)
{
"status": "open",
"diagnosis": null,
"flags": [],
"V": {
"V_prior": null,
"display": "2",
"headline": 2,
"ref": 2,
"valid": 2,
"weaker": "both"
},
"belief": null,
"draft": "ai"
}Say something about it
Each verb opens a prefilled issue; a round files what lands as the ledger event it is.