Adopt the setting, notation, and hypotheses of Level-Revealed Conditioning for Jointly Driven Solutions of the Controlled N-Agent Dynamics: the probability space (Ω,F,P) with transition and observation clocks and the σ-algebra S0; the clock labels a with clocks Ya and rate bounds Ba; and, for j∈{1,…,J}, the S0-measurable initial states ςj, the policies hj, and the given solutions of the controlled N-agent dynamics on [0,T], with consumed clock times Aja and system filtrations (Ftsys,j)t∈[0,T]. For t∈[0,T] let Ft=Ftsys,1∨⋯∨Ftsys,J be the σ-algebra generated by the union, and for a clock label a define the frontier consumed time At∨,a=max(A1a(t),…,AJa(t)). Let μp and exp be as in Predictable-Window Moment Identities for the Homogeneous Poisson Process, and write E for the expectation, expectations of [0,∞]-valued random variables being their integrals with respect to P.
1. (Frontier windows) Let t∈[0,T], let a be a clock label, let Vˉ and Λˉ be nonnegative reals, and let v and λ be Ft-measurable random variables with At∨,a(ω)≤v(ω)≤Vˉ and 0≤λ(ω)≤Λˉ for every ω. Then for every natural number p≥1 and every Ft-measurable Z:Ω→[0,∞], in [0,∞]:
E[Z(Yv+λa−Yva)p]=E[Zμp(λ)]andE[Z1{Yv+λa−Yva≥1}]=E[Z(1−exp(−λ))]≤E[Zλ];
in particular, taking p=1, E[Z(Yv+λa−Yva)]=E[Zλ].
2. (Crossing compensation) For all real 0≤s≤w≤T, every clock label a, and every Fs-measurable Z:Ω→[0,∞], in [0,∞]:
E[Z(YAw∨,aa−YAs∨,aa)]=E[Z(Aw∨,a−As∨,a)].
3. (Product windows) Let t∈[0,T], let a1,…,ak be distinct clock labels with k≥1 a natural number, let Vˉ and Λˉ be nonnegative reals, and for each n∈{1,…,k} let vn and λn be Ft-measurable random variables with At∨,an(ω)≤vn(ω)≤Vˉ and 0≤λn(ω)≤Λˉ for every ω. Then for all natural numbers p1,…,pk≥1 and every Ft-measurable Z:Ω→[0,∞], in [0,∞] and with the finite product notation:
E[Z∏n=1k(Yvn+λnan−Yvnan)pn]=E[Z∏n=1kμpn(λn)];
moreover the identity continues to hold when, for the indices n in an arbitrary subset of {1,…,k}, the factor (Yvn+λnan−Yvnan)pn is replaced by 1{Yvn+λnan−Yvnan≥1} and correspondingly μpn(λn) by 1−exp(−λn).