Deterministic competitiveness carries over to the randomised notion
ProvedAJOnline.ofStrategy_randCompetitiveLet be a costed game, and let be a strategy on that is -competitive:
Then , read as a behaviour strategy, is -randomised-competitive against oblivious adversaries.
The randomised notion therefore extends the deterministic one.
import Mathlib import Definitions.Def_AJOnline_CostedGame
namespace AJOnline
open AJGames
open scoped ENNReal NNReal
universe u
/-- **Deterministic competitiveness carries over.** If `σ` is `(ρ, b)`-competitive, the behaviour
strategy `ofStrategy σ` is `(ρ, b)`-competitive against oblivious adversaries: against any script, its
play-out is the single play of `σ`. AJ axioms used: strategy axioms (nil, receptive; first-O and
determinism through `ofStrategy`) and G3 through `ofStrategy`'s legality. No tensor axiom. -/
theorem ofStrategy_randCompetitive (C : CostedGame.{u}) (σ : Strategy C.G) (ρ b : ℝ≥0)
(h : Competitive C σ ρ b) : RandCompetitive C (ofStrategy σ) ρ b := by sorry
end AJOnlineRead-back
What the Lean code literally says, in plain math · claude-opus-5-5
Games and strategies. Players are and . A game has a move type , a labelling , and a set of finite move sequences that contains the empty sequence , is prefix-closed, and whose members alternate in label. A (deterministic) strategy contains , is prefix-closed, has every nonempty member beginning with an -move, is deterministic ( with both -labelled implies ) and receptive (, , implies ).
Costed games. A costed game is a game together with two arbitrary functions from finite move sequences (legal or not) to . For :
- is -competitive if for every ,
Behaviour strategies and play-out. A behaviour strategy gives for every finite sequence a distribution on , with the constraint that forces nonempty with last move labelled , , and . For a script of moves, is the distribution on sequences defined by: empty script gives the point mass at ; for , if and , draw and return the point mass at if , or continue with if ; otherwise return the point mass at . The expected costs over a run from the empty history are
sums over all finite sequences, in (with the convention ). is randomised -competitive if for every finite script of moves (any moves at all, including -labelled or illegal ones),
Induced behaviour strategy. puts, at history , the point mass on one -labelled with (chosen by the axiom of choice) if such exists, and the point mass on otherwise.
Statement. For every costed game (any universe), every strategy on its game, and all :
All arithmetic is in : and may take the value , and with the term equals . The hypothesis only constrains on members of ; values on other sequences are unconstrained.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.