MEFI Model Comparison Laboratory v0.3

MEFI Theory · reproducible comparison bench

Model Comparison Laboratory v0.3

Run a conventional baseline beside the unchanged MEFI core, preserve both raw outputs, import real observations, and calculate comparison metrics only after the simulations finish. The page does not declare a winner without comparable evidence.

ONE UNCHANGED MEFI COREΔQ ALWAYS NONZERORAW OUTPUT FIRSTANALYSIS AFTER SIMULATIONNO AUTO-WINNER

Canonical MEFI core used in every moduleF_MEFI(r,t) = k_c/r² − k_r/[r²(1+r)] + ΔQ · f_UFR(t)

Runnable conventional solvers9

MEFI governing formulas1

Observed rows0

Benchmark statusNO DATA

Current winner—

Framework contrast · what this page is actually demonstrating

The conventional side is allowed to use the equations appropriate to its domain. The MEFI side is not. MEFI must retain the same core in every case. This section describes structural differences only; benchmark performance is reported separately and only when comparable observations are loaded.

Conventional governing structureChanges with selected domain

—

MEFI governing structureOne unchanged core

k_r compression · k_c expansion · r relationship · nonzero signed ΔQ · f_UFR(t). F_MEFI is the current snapshot.

Current conventional configuration—

Displayed factors and domain assumptions are explicit in the conventional ledger.

MEFI domain switchingNone

The substrate and active relationship change; the governing core does not.

Conventional model

equations shown—

configured factors—

finite outputs—

raw conventional output

Observed / reference

benchmark evidence

Load actual observations or a completed benchmark bundle. Accepted bundles can contain x, observed, conventional, mefi, allowing historical comparisons to be replayed without altering either model.

CSV: x, observed[, conventional, mefi]
JSON: {"x":[],"observed":[],"conventional":[],"mefi":[],"meta":{...}}

No observed dataset loaded. Simulations can run, but no superiority result will be calculated.

rows0

source—

units—

imported observation only

MEFI

same core every relationship

The domain can change completely while the governing MEFI core remains unchanged. The upper plot is one instantaneous cross-r snapshot. The continuous engine below repeatedly recalculates the same core as its visible input relationships change.

F_MEFI(r,t) = k_c/r² − k_r/[r²(1+r)] + ΔQ·f_UFR(t) Π(t) = ΔQ·C(t) is retained as a diagnostic readout; it does not replace or drive the core. No domain-specific equation, propagation rule, attenuation rule, fitting law or substitute solver generates the MEFI state.

k_r · compression

k_c · expansion

r minimum

r maximum

ΔQ · nonzero signed difference

f_UFR(t) · active UFR occupation

C(t) · coherence diagnostic

governing equations1

core factors5

current Π = ΔQ·C—

F snapshot span—

instantaneous F_MEFI snapshots across r · one time slice

MEFI continuous reevaluation · snapshot history & stability ranges

Every row is another evaluation of the exact same core. The page never invents a second dynamical law. You may import a real MEFI relationship history, or generate a clearly labeled transparent test trajectory by interpolating between visible anchor states.

External MEFI context: Earth UFR, solar/CME, geomagnetic, planetary, laboratory, or other enclosing-node influences can be retained with every row. They are provenance/context fields. This page does not secretly convert those observations into f_UFR(t); the active MEFI relationship values must be supplied or resolved by the MEFI source process.

Anchor A · starting relationship

Anchor B · changed relationship

Anchor C · changed relationship

Samples per A→B→C→A cycle

External context label

Earth UFR context · optional

Coherent tag for generated test rows

Snapshots0

Current F_MEFI—

Occupation envelope—

Mean |ΔF|—

Current Π—

Coherent-tagged range—

No continuous trajectory loaded yet. The instantaneous MEFI snapshot above remains valid as a single evaluation.

F_MEFI · expansion term · compression term · ΔQ·f_UFR

nonzero ΔQ · f_UFR(t) · C(t) · r · all visible input history

t	k_r	k_c	r	ΔQ	f_UFR	C	Π	Expansion	Compression	ΔQ·f_UFR	F_MEFI	Earth UFR	Context

Side-by-side result ledger

Metrics summarize completed outputs; they never generate MEFI state. Raw mode requires comparable target units. Shape mode uses post-simulation min–max normalization and is labeled as shape-only.

overlay activates after import or UI demo

x	Observed	Conventional	MEFI	\|Conv−Obs\|	\|MEFI−Obs\|

Conventional vs observation

MAE —
RMSE —
Correlation —

MEFI vs observation

MAE —
RMSE —
Correlation —

No result is declared until observed/reference data and comparable model outputs are present.

Conventional configuration ledger

MEFI configuration ledger

One governing core formula across every module.
k_r remains compression; k_c remains expansion.
r is active relationship / separation.
ΔQ is always nonzero signed difference.
f_UFR(t) is current UFR occupation.
F_MEFI is a snapshot, not a permanent state.

Built-in raw plots intentionally keep their native coordinates. A direct winner requires a bundle whose conventional and MEFI outputs are already expressed against the same observed target. Shape mode is pattern analysis, not equal physical units.

Comparison-engine coverage

✓ Newton / Kepler

Runnable inverse-square and two-body orbital baselines.

✓ Schwarzschild

Runnable gravitational redshift and null-geodesic light-bending solvers.

✓ Stellar structure

Numerically integrated Lane–Emden polytrope.

✓ Neutron-star structure

Numerically integrated TOV system with an explicit polytropic EOS.

✓ Cosmology / thermal

ΛCDM background H(z) and Planck thermal-spectrum baselines.

✓ N-body gravity

Deterministic leapfrog integration with energy-drift validation.

Next: full CMB Cℓ

Import CAMB/CLASS or published model vectors; do not substitute a toy anisotropy model.

Next: production codes

Attach external research-grade outputs for numerical relativity, stellar evolution, plasma/MHD, and large-N cosmology.