Theorem 1, p. 11 — every pacing game admits a second price pacing equilibrium
ProvedMultPacing.Existence.theorem_1budgetsequilibrium-existencep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1pacingsecond-price-auction
Let be a finite set of buyers and a finite set of divisible goods, each sold in a second price auction. Each buyer has valuations and a budget . Then the pacing game has a second price pacing equilibrium: there are pacing multipliers , allocations and prices such that
- for every good , with equality when some ;
- implies for every ;
- implies (and if is the only buyer);
- for every buyer ;
- implies .
In short,
The result holds for all finite numbers of buyers and goods and all valuations, with no genericity or tie assumption. It shows that the budget-management rule of multiplicative pacing always has a stable outcome in which every buyer best responds (Proposition 2 of the paper) and the market clears.
Differences from the source:
- encoding: buyers and goods → arbitrary finite types
N,G; reason: the same instances up to relabelling. - encoding: "the per-unit price " → a price vector quantified with and and pinned to the highest other bid wherever , with ; reason: Definition 2 names without quantifying it, and an unpinned only multiplies allocations equal to .
- encoding: "the highest bid other than 's bid" → the maximum over , when there is no other buyer; reason: bids are nonnegative, and the paper gives a lone bidder price (p. 16).
- encoding: budgets → real numbers ; reason: §3's model has finite budgets; appears only in examples outside the model.
Preamble
import Mathlib import Definitions.Def_MultPacing_Existence_Model
Formal statement
namespace MultPacing.Existence
theorem theorem_1 {N G : Type*} [Fintype N] [DecidableEq N] [Fintype G] (P : PacingGame N G) :
∃ (α : N → ℝ) (x : N → G → ℝ) (p : G → ℝ), IsSPPE P α x p := by sorry
end MultPacing.Existence
Source
Conitzer, Kroer, Sodomka & Stier-Moses, Multiplicative Pacing Equilibria in Auction Markets, arXiv:1706.07151v5, p. 11 (PDF 11), Theorem 1 (with Definition 2, p. 9)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.