Context
The potential charges the smaller of arithmetic guard capacity and actual fixed-weight word supply, correcting capacity-only overcounting.
Definitions
- counts full guard slots allowed by the best completion budget and the least-representative-greater-than-two convention.
- With and ,
Hypotheses and scope
- The target layer and fixed-weight feasibility convention are fixed.
- The child guard cylinders partition the parent extensions, and infeasible children contribute zero.
Proof or evidence
Child cap inclusion gives ; Pascal gives ; and proves monotonicity. At depth , and the cap is exactly the acceptance indicator. The loss law follows from .
Verification notes
This run checked the child-cap inclusion, the min inequality, the full-depth equality, and the bottleneck algebra. The repository explicitly records that the post-freeze transcript is preserved evidence rather than an independently rerun all-length checker.
Limitations
- Monotonicity does not imply a uniform factor loss.
- This capped count is distinct from the abstract lift-slot count in FPRD-T141.
- No standalone executable checker for the all-length T143 theorem was identified.
Notes
Monotonicity combines the child guard-cylinder partition, monotonicity of the maximal-completion cap, and Pascal's identity; the loss formula is the identity 2 min(x,y)=x+y-|x-y|. It caps arithmetic capacity by combinatorial supply but gives no uniform factor loss. The separate post-freeze transcript is preserved evidence, not an independently rerun checker for this all-length theorem. Ncap is not the abstract lift-slot count Nslot of FPRD-T141; the two potentials need not share a root value.