symmetric_game_regret
For a positive natural number n, an n-by-n matrix A over ℝ, and a strategy p in stdSimplex ℝ (Fin n), the symmetric game regret symmetric_game_regret A p is Finset.univ.sup' Finset.univ_nonempty (fun i => Matrix.mulVec A p.val i) - dotProduct p.val (Matrix.mulVec A p.val).
Statement
def symmetric_game_regret {n : ℕ} [NeZero n] (A : Matrix (Fin n) (Fin n) ℝ) (p : stdSimplex ℝ (Fin n)) : ℝ- Defined in
- arXiv:2610.25524 Section 1.1
- Verified
- 20 Oct 2026
- Axioms
- none listed
- Used by 3
- saddle_gap_low_value_symmetric_approx_nashLet n be a positive natural number with
[NeZero n]. Let A : Matrix (Fin n) (Fin n) ℝ be an n-by-n real matrix with entries in [0, 1]… - symmetric_game_regret_le_of_sup_sub_inf_leFor a positive natural number n, an n-by-n real matrix A, a strategy x in
stdSimplex ℝ (Fin n), and a real number ε, if… - symmetric_game_regret_le_sup_sub_infFor any strategy x in stdSimplex ℝ (Fin n), the regret symmetric_game_regret A x is at most the maximum over i of (A *ᵥ x) i minus the…
- saddle_gap_low_value_symmetric_approx_nashLet n be a positive natural number with
Read back from the Lean
Given a positive natural number $n$ (via [NeZero n]), an $n \times n$ real matrix $A$, and a probability vector $p$ in the standard simplex on Fin n over $\mathbb{R}$ (i.e., $p \in \Delta^{n-1}$), symmetric_game_regret A p is defined as the real number:
$$\left(\max_{i \in \{0, \dots, n-1\}} (A p)_i\right) - p \cdot (A p)$$
that is, the maximum entry of the matrix-vector product $A p$ minus the dot product of $p$ with $A p$.
Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.
Lean source view module on GitHub
/-- Symmetric game regret of a strategy p: max_i e_iᵀ A p - pᵀ A p. -/ def symmetric_game_regret {n : ℕ} [NeZero n] (A : Matrix (Fin n) (Fin n) ℝ) (p : stdSimplex ℝ (Fin n)) : ℝ := (Finset.univ.sup' Finset.univ_nonempty (fun i => Matrix.mulVec A p.val i)) - dotProduct p.val (Matrix.mulVec A p.val)