Theorem 3.2 — for cancelable valuations the bluff profile is an exact pure Nash equilibrium of Round-Robin
ProvedRoundRobinEq.CancelableBluff.theorem_3_2Let agents share indivisible goods, and let every agent's valuation be a cancelable valuation function (normalized and non-decreasing). Let be any strict variant of the truthful profile (each agent ranks the goods by non-increasing singleton value , ties broken in any fixed way), let be the goods in the order in which Round-Robin allocates them, and let be the bluff profile, where every agent reports .
Then is an exact pure Nash equilibrium of Round-Robin: for every agent and every ranking of ,
where is agent 's bundle in the allocation Round-Robin returns.
The result generalizes the known existence of exact equilibria of Round-Robin for additive valuations to the strictly larger class of cancelable valuations, with an explicit equilibrium that can be computed by running Round-Robin once on truthful reports.
Formalization Note The conclusion is IsApproxPNE 1 v (bluff σ): the bluff reports are rankings, and no deviation to any ranking τ raises the deviator's true value. The standing simplification (p. 7) is carried as the hypothesis hm; the reduction of general by dummy goods is not formalized. The strict truthful profile σ is universally quantified, so the theorem holds for every tie-breaking, as on the page.
import Mathlib import Definitions.Def_RoundRobinEq_CancelableBluff_Setting
namespace RoundRobinEq.CancelableBluff
theorem theorem_3_2 {n : ℕ} [NeZero n] {M : Type*} [Fintype M] [DecidableEq M]
(k : ℕ) (hm : Fintype.card M = k * n) (v : Fin n → Finset M → ℝ)
(hv : ∀ i, IsValuation (v i)) (hc : ∀ i, IsCancelable (v i))
(σ : Fin n → List M) (hσ : IsStrictTruthful v σ) :
IsApproxPNE 1 v (bluff σ) := by sorry
end RoundRobinEq.CancelableBluff
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.