Lemma F.2: derivative of scaled own-price variance
ProvedBFWFluidLimit.scaled_variance_hasDerivAtprobabilitystochastic-approximation
Suppose t≠0 and u,w have derivatives (p−u(t))/t and (p²−w(t))/t at t. Then the derivative of s↦s(w(s)−u(s)²) at t is (p−u(t))². This is the scalar differential identity in Lemma F.2 of the paper, from which the own-price variance floor follows.
Preamble
import Mathlib.Analysis.Calculus.Deriv.Pow import Mathlib.Analysis.Calculus.Deriv.Mul import Mathlib.Analysis.Calculus.Deriv.Add import Mathlib.Analysis.Calculus.Deriv.MeanValue import Mathlib.Topology.MetricSpace.Lipschitz import Mathlib.Tactic.FieldSimp import Mathlib.Tactic.Ring import Mathlib.Tactic.Linarith import Mathlib.Tactic.SplitIfs
Formal statement
namespace BFWFluidLimit
theorem scaled_variance_hasDerivAt (u w : ℝ → ℝ) (t p : ℝ) (ht : t ≠ 0)
(hu : HasDerivAt u ((p - u t) / t) t)
(hw : HasDerivAt w ((p ^ 2 - w t) / t) t) :
HasDerivAt (fun s => s * (w s - u s ^ 2)) ((p - u t) ^ 2) t := by sorry
end BFWFluidLimit
Source
J. Baek, V. F. Farias, F. Wu, Misspecified Estimate-then-Optimize Leads to Supra-Competitive Prices, arXiv:2605.16064v2, https://arxiv.org/pdf/2605.16064; Lemma F.2 p. 49.
Read-back
What the Lean code literally says, in plain math · gpt-6
At a nonzero time, assume the own-price mean and second-moment coordinates have the stated mean-field derivatives. The derivative of time times centered own-price variance is the square of the posted-price deviation from the mean. This is an exact differential identity, not an assumed variance-floor inequality.