lemmaroutine

Direct collision charge bound

Direct collision charge is ≤ 2−p2Ch∗B−1/(2s)2^{-p}2^{Ch_*}B^{-1/(2s)}; rest is child discrepancy
collision charge in the coupling recurrence
Lemma·lem:fp-direct·§9.4 Bucket algebra and the collision charge

Statement (paper)

The direct collision contribution to any final pre-incoming output has norm at most 2−p 2Ch∗B−1/(2s), 2^{-p}\,2^{C h_*}B^{-1/(2s)}, for a fixed CC independent of D1,D2,ϵD_1,D_2,\epsilon introduced below. All remaining contributions on Fxc\mathcal F_x^c are bounded by actual--clean child discrepancy norms with an amplification at most 2Cd2^{C d} at slack dd.

In plain English

On the complement of the key-success event, the direct collision contribution to any final pre-incoming output has norm at most 2−p 2Ch∗B−1/(2s)2^{-p}\,2^{C h_*}B^{-1/(2s)}, with CC independent of D1,D2,ϵD_1,D_2,\epsilon. Every other contribution there is bounded by actual–clean child discrepancy norms, with amplification at most 2Cd2^{Cd} at slack dd.

Why it's here

This lemma splits the bucket error from the bucket discrepancy bound into two parts: a collision term whose size is controlled by the collision probability together with the small clean-height row norm, and a remainder controlled by child discrepancies at cost 2O(d)2^{O(d)} via the raw difference estimate. The uniform block coupling proposition uses this split to bound the recurrence by an escape/collision charge plus child discrepancies.
Classified routine:
This splits the earlier bucket discrepancy bound into a collision term and child-discrepancy remainders, using the raw difference estimate and buffer Lipschitz/clamping lemmas. It is bookkeeping on top of earlier nodes, not a new idea.

Inside the proof

Sources

PaperOpen
Section
§9.4 Bucket algebra and the collision charge
Page
—
Label
lem:fp-direct
Line
sections/fingerprints.tex:523
Leannot aligned

Phase 2 aligns this node to a Lean declaration.