Lemma 1, p. 9 — reduction to zero cost and unit mean: the ratio is unchanged and µc = 1, Sc = (S+M−1)/M, Mc = 1, Dc = D/M
ProvedValuePersPricing.Deviation.lemma_1p2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1personalized-pricingprice-discrimination
Let be a market (a valuation with almost surely and finite mean ) with , i.e. margin . Let be the law of the standardized valuation
Then is a market with mean , and
so that . Moreover , , if has scale then has scale , and if has unbounded support so does .
This reduction lets every bound be proved for , and transported back to a general cost and mean.
Formalization Note is the pushforward standardize ν c of ν under .
Differences from the source:
- generalization: the ratio identity R_PP(F,c)/R_SP(F,c) = R_PP(F_c,0)/R_SP(F_c,0) → the two scaling identities R_SP(F,c) = (µ−c)R_SP(F_c,0) and R_PP(F,c) = (µ−c)R_PP(F_c,0); reason: the proof on p. ec1 shows exactly these, and they imply the ratio identity whenever it is defined.
- generalization: the statistics of F_c → also that (F_c, 0) satisfies the standing assumptions and that unbounded support is preserved (S = ∞ gives S_c = ∞); reason: needed to apply the c = 0 results to F_c, including in the unbounded case; the extra conjuncts only add to the lemma.
- encoding: "let V_c := (V − c)/(µ − c)" → hypothesis c < µ; reason: V_c is undefined when µ = c, and µ ≥ c always holds.
Preamble
import Mathlib import Definitions.Def_ValuePersPricing_Deviation_Model open MeasureTheory Set Filter Topology
Formal statement
namespace ValuePersPricing.Deviation
/-- Lemma 1 (p. 9): reduction to zero cost and unit mean. With `F_c` the law of
`V_c = (V − c)/(µ − c)`, the standardized market has cost `0` and mean `1`; both profits scale by
`µ − c` (so the ratio `R_PP/R_SP` is unchanged); `M_c = 1`, `D_c = D/M`,
`S_c = (S + M − 1)/M`, and unbounded support is preserved. -/
theorem lemma_1 (ν : Measure ℝ) (c : ℝ) (hm : Market ν c) (hc : c < mean ν) :
Market (standardize ν c) 0 ∧
mean (standardize ν c) = 1 ∧
spProfit ν c = (mean ν - c) * spProfit (standardize ν c) 0 ∧
ppProfit ν c = (mean ν - c) * ppProfit (standardize ν c) 0 ∧
margin (standardize ν c) 0 = 1 ∧
coefDev (standardize ν c) = coefDev ν / margin ν c ∧
(∀ S : ℝ, IsScale ν S → IsScale (standardize ν c) (sc (margin ν c) S)) ∧
(Unbounded ν → Unbounded (standardize ν c)) := by sorry
end ValuePersPricing.Deviation
Source
Elmachtoub, Gupta & Hamilton, The Value of Personalized Pricing, SSRN 3127719 (review version, PDF sha256 635e22eb…), p. 9 (PDF 9), Lemma 1; proof p. ec1–ec2 (PDF 33–34)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.