EEML v0.25 · changelog-en

v0.25 · 27 September 2026 Current

EEML as a unifying principle and four core rules, separating starting points, model inputs, consequences and protocol.
Possibilities, weight and consistency separated; ε bridge and tied minima made explicit.
Marginalisation, four-way couplings, memory, correlated noise, fluctuation-response and relative entropy developed.
Reverse engineering and a retrodiction protocol added, with reference models and failure criteria.
K3 need not be unique or most efficient; χ=24 remains geometric. Kummer construction and duality context added.
ε≈1 for discharges, N_gen=χ/8 and direct 24→α claims explicitly historical-revised; no new measurements claimed.
NL and EN updated together. Earlier pages, proof details and error reports preserved.

This version reorganises the research; it does not establish EEML as a physical law or reproduce all earlier calculations. The correction register →

Complete snapshot before this revision (NL/EN)

Earlier versions

v0.24 · 16 September 2026

Addition to the α route: a scaling lemma, an exact pre-test and four sharpened test criteria, worked out after a proposal by Robin Genis.

Scaling lemma: bare heat-trace and determinant costs on a fixed K3 select no finite size; the determinant scales as C + (22/15) log L.
Counterexample: a stabilised size with a quantised contribution does not yet fix α.
Exact criterion for positive posynomials: existence, uniqueness, flat directions and constancy of a quantity, with proof.
Three-term model with a unique minimum, e*³ = n/c, and a targeted test b > 0 and a < 2b for a third term.
16 symbolic check groups and 12 test cases published, without the measured α as input.

No derivation of α: the third contribution, its coefficient and the electromagnetic identification remain open. v0.23 is preserved.

Download the α route bundle · Reproduction and scope

The route to α →

Earlier version: v0.23

v0.23 · 16 September 2026

The original EEML work received a mathematical and computational check. Valid results return with proofs, conditions and examples.

Two original Lean files rebuilt: one core theorem and four symmetry results pass. The original formal core already used the correct half-line.
Kraft, Bayes, Pareto, Laplace weights and a conditional Born derivation developed.
K3 derivations χ=24, σ=−16, scalar a₄=4/15 and ζ(0)=−11/15 restored.
Modular calculation and coupling running recalculated.
An energy-matrix error and a factor two in coupling normalisation identified. Dependent numerical claims remain unconfirmed.
Check code, selected original sources, inputs, results and Lean logs published.

The aim remains deriving α. This is not a complete reproduction of all historical research projects. A full v0.22 backup was made before this update. The original site backup and public text archive remain preserved.

Download the verification bundle · Reproduction and scope

The complete verification →

Previous version: v0.22

v0.22 · 16 September 2026

The philosophical core has been rewritten around the Effortless Effect, the undifferentiated zero point and deliberate self-application. The research aim remains a derivation of α, with K3 and χ=24 as a candidate route.

Corrections recorded

The exponential theorem now has a non-negative domain, explicit dependence of weight on resistance, positivity and non-triviality conditions.
Weight, probability measure, a minimum and stationarity are distinguished.
The fixed-volume Einstein-Hilbert claim is corrected to Einstein metrics rather than necessarily Ricci-flat metrics.
Describing nothing is no longer treated as proof of its impossibility. No resistance is no longer treated as proof of instability.
Self-application is a consistency requirement, not an established unique law.
The conclusion that α must be an address without a derivation is withdrawn as a general conclusion. Failure of particular methods does not establish impossibility.
K3 and particle-selection results are stated within their assumptions, rather than as necessary consequences of the meta-law.
General confidence scores and the claim of a fully falsifiable meta-law are removed.

What this revision does not settle

The earlier code and formal proof projects have not been rerun. Their results are not collectively declared false. Their correspondence with broader published claims requires review. No new experimental verdicts are reported.

Preservation

A complete backup of the live site was made before editing. The public snapshot below preserves the earlier content, including failed predictions and the historical article. Original media and the preprint remain available as historical material. Historical annotations are not current endorsements.

Earlier versions · snapshot 2026-09-16

Earlier media

The following videos illustrate earlier versions and may contain claims revised above.

Effortless Effect Meta-Law v2 (video) · The Effortless Effect (video) · Preprint (PDF)

v0.25 · 27 September 2026