Reason: New pure tool: cell-count form of a compensated counting-path functional along a time change, with window-discrepancy error bounds (P7.3).
Statement
Let R>0 and s>0 be real numbers, let J≥1 and l≥1 be natural numbers, and let 0=b0<b1<⋯<bJ=R be real numbers (the cell boundaries), with cell lengthsμj=bj−bj−1 (1≤j≤J) and μmax=max1≤j≤Jμj. Let p be a counting path. Define its cell countsKj=p(bj)−p(bj−1) (1≤j≤J), its compensated pathMˉ(x)=p(x)−x (x∈[0,R]), and, for a real number w≥0, its window discrepancy
Let w1≥0 be a real number and let C,Cˉ:[0,s]→[0,R] be measurable maps with ∣Cu−Cˉu∣≤w1 for every u∈[0,s]. Let v∈Rl and let H:[0,s]→Rl be bounded measurable; put ∥H∥1=∫[0,s]∣Hu∣du, the integrand being bounded measurable by Norm Bound for a Vector-Valued Lebesgue Integral over a Compact Interval. Products of a real number with a point of Rl are scalar multiples. For a measurable C′:[0,s]→[0,R] define the compensated functional
Λ(C′)=vMˉ(Cs′)+∫[0,s]HuMˉ(Cu′)du∈Rl
(defined by claim 1), and define the cell coefficients
1. (Measurability and the partial-cell decomposition.) For every measurable C′:[0,s]→[0,R] the map u↦p(Cu′) is bounded measurable, so that u↦HuMˉ(Cu′) is bounded measurable and Λ(C′) is defined; the maps u↦1{bj≤Cˉu}Hu are bounded measurable, so that the αj are defined. Moreover, for every x∈[0,R],
Mˉ(x)−j=1∑J(Kj−μj)1{bj≤x}≤Discμmax(p).
2. (Time-change error.) For every u∈[0,s], ∣Mˉ(Cu)−Mˉ(Cˉu)∣≤Discw1(p), and
Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.