[pinterbanget]

Summary

This adds research/pr4359-ledger-rebalance/. It answers the open question of my #4410. In zxv's #4359 ledger (controller/outer_parameters.py), the complex uniform saving appears only in 0.002304 − b. The BIT saving appears only in bit_saving − a, with a = 0.0023025, after the cover fallback and atom conversion. The nine outer cost rows depend on neither. #4359's choices of 0.002304 and 0.00276 sit far above what the ledger needs, and since H1 ∝ 1/δ², they are what makes the ranks astronomical.

Claim type: an exact parameter rebalancing within #4359's ledger. No new κ: κ stays 0.0023, and #4359 is not accepted.

Result

The rebalanced values are:

• b = 0.00230255;
• complex uniform saving = 0.0023026;
• BIT C0 = 0.002311 and C1 = 0.00231, giving an exact ordinary saving of 0.0023031853797, with 0.002303 admitted.

All 54 ledger predicates, the BIT conversion and the fallback stay strict. The nine outer margins are unchanged. The superpotentials are certified at #4359's unchanged tested savings, and the ranks follow #4405's refined induction:

Family	#4359	#4405	rebalanced
Complex	2^276540184361	2^1546084442	2^268300549
BIT	2^6552928412	2^42843607	2^14664

Control. The transcribed ledger reproduces #4359's nine published margins, from 0.00230197698 to 77.00004, along with all 54 predicates and the exact BIT conversion 27418653797/10¹³. Both earlier rank columns are recomputed too.

Evidence

python3.12 -B scripts/research_package.py --package research/pr4359-ledger-rebalance --check-only
  PASS_PACKAGE_METADATA_AND_PINS (not a mathematical acceptance check)
python3.12 -B scripts/research_package.py --package research/pr4359-ledger-rebalance
  PASS_REGISTERED_RESEARCH_PACKAGE research/pr4359-ledger-rebalance outcome=completed   (~1 s, stdlib only)

Head: e32b03203d3805b9a0c839d1a00f01bffce99dcc. Base: canonical main 12fa8c5c, priority revision 7.

Not covered

• All of Conditional κ = 0.0023 (+5.690e-4, +32.87% over #4181): productive helper jobs compiled into a uniform supplier #4359's hypotheses. The precision and storage constants must be recomputed for these different templates, as Conditional κ = 0.0023 (+5.690e-4, +32.87% over #4181): productive helper jobs compiled into a uniform supplier #4359 already requires.
• The inherited 63-predicate ledger named in Refine PR4359: conditional κ = 0.00230509 and smaller explicit template ranks #4405.
• Materialized templates.

Task and overlap

• Card: independent-moment, refined.
• Claim: Handoffs and requests for independent checks #1166 (comment)
• Overlap: no rebalancing had been proposed.

Credit

@zxv (#4359), @1337isnot1337 (#4405), and Jacob Alan Sussman (#4268).

Prepared by Ryanson Jonathan with substantial Claude Code (Claude Opus 5.5) assistance; no independent human review. Apache-2.0. This PR does not merge anything or change the frontier.

🤖 Generated with Claude Code

[pinterbanget]

Update (head 7d6a7df): the verifier now also transcribes the 19 functional checks of #4359's transfer-audit/outer_transfer_audit.py; all 19 stay positive at the rebalanced values. This closes the package's former open question 1, except for the 13 historical-endpoint predicates of the inherited ledger, which compare only κ.

[pinterbanget]

Note: zxv's parallel #4411 (posted two minutes before this PR) retunes the same consumer more aggressively and replaces the power majorant with a logarithmic one, giving ranks 2^465000 / 2^7100 at κ = 0.0023. Those supersede this PR's 2^268300549 / 2^14664. #4425 independently reviews #4411's majorant. This PR's ledger observation (the savings enter only through uniform − b and bit_saving − a) is consistent with #4411's retuning.