Proof of Theorem 4 — linear growth of Ĵ_t on the event A_t
ProvedKeskinZeevi.IncumbentSufficient.Jhat_linear_growth_on_goodEventFix a pricing policy, a slope , a constant , a period and a realization of the shocks. Suppose that condition (ii) of Theorem 4 holds at time along this realization:
If the realization lies in the event (the greedy deviations for are all within of ), then
So the information metric grows linearly in on when : a policy that tracks the greedy price, which is bounded away from the uninformative incumbent price, accumulates information at a constant rate.
Formalization Note. The statement is deterministic (pathwise). The paper applies it for ; it is stated here for , where so that the sum in (ii) starts at and covers every index of . The sum in condition (ii) starts at rather than the paper's , because is defined only for greedy ILS. is the natural logarithm, applied to the real number .
import Mathlib import Definitions.Def_KeskinZeevi_IncumbentSufficient_LeastSquares
namespace KeskinZeevi.IncumbentSufficient
/-- Keskin–Zeevi 2014, p. 1159, proof of Theorem 4, display after (A16): pathwise, for `t ≥ 3`,
if condition (ii) holds at time `t - 1`, i.e. `Σ_{s=1}^{t-1} (ψ(b_s) - x_{s+1})² ≤ κ₁ log(t - 1)`,
then on the event `A_t` of (A15)
`Ĵ_t ≥ (ψ(β)²/8 - κ₁ log(t - 1)/(t - 1)) (t - 1)`. -/
theorem Jhat_linear_growth_on_goodEvent (M : Model) {Ω : Type*} (π : Policy M) (β : ℝ)
(ε : ℕ → Ω → ℝ) (κ₁ : ℝ) (t : ℕ) (ht : 3 ≤ t) (ω : Ω)
(hii : ∑ s ∈ Finset.Icc 1 (t - 1),
(M.psi (truncEst π β ε s ω) - dev π β ε (s + 1) ω) ^ 2 ≤ κ₁ * Real.log ((t : ℝ) - 1))
(hA : ω ∈ goodEvent π β ε t) :
(M.psi β ^ 2 / 8 - κ₁ * Real.log ((t : ℝ) - 1) / ((t : ℝ) - 1)) * ((t : ℝ) - 1) ≤
Jhat π β ε t ω := by sorry
end KeskinZeevi.IncumbentSufficient
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Setting. Take any model , any type , any policy , any real (not required to lie in ), any noise functions , any real (no sign assumed), any natural and any .
Notation. Everything is evaluated along for the process that generates under and :
- are the price deviations;
- are the truncated least-squares estimates, where is set to when ;
- .
Hypotheses.
- The cumulative pricing error up to period is small:
- . That is, for every natural with :
Conclusion.
This is a pathwise statement. No measure is involved.
Degenerate cases.
- Well-definedness: since , the real number . So there is no division by zero, and .
- : hypothesis 1 cannot hold, because its left side is and its right side is . The statement is then vacuous.
- : hypothesis 1 forces for all .
- : this is allowed, including the case with . Hypothesis 2 then forces in the range. The conclusion becomes .
- Index range for : the range in is , and it never includes here.
- Empty : the statement is vacuous.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.