Evidence

Every claim tested against published data. Zero free parameters in the physics. The Lorentz invariant derived, the cosmological constant resolved, five Standard Model constants matched. Drug response predictions at 3.6% MAE. Patient-level AUC 0.808 from two measurements. The geometry either predicts the data or it does not.

384/384

Positions matched

10⁻¹²²

Cosmological constant (0 params)

0.808

Patient AUC (NSCLC)

Free parameters

Universal Sequence — 24 Systems × 16 Positions

Twenty-four dissipative processes across eight scientific domains, each mapped to sixteen positions. Every position verified against domain-specific literature. 384 out of 384 positional matches.

Domain	Systems	Positions	Match	Source
Physics	Stellar evolution, Nucleosynthesis, Baryon asymmetry	48/48	100%	Paper 2
Chemistry	Crystallisation, Nanocrystals, Protein folding	48/48	100%	Paper 2
Biology	Cell cycle, Immune response, Embryogenesis	48/48	100%	Paper 2
Microbiology	Biofilm formation, Sporulation, Quorum sensing	48/48	100%	Paper 2
Ecology	Forest succession, River systems, Perturbation cycles	48/48	100%	Paper 2
Geophysics	Volcanism, Water cycle, Soil formation	48/48	100%	Paper 2
Medicine	Cancer treatment, Drug combinations, Wound healing	48/48	100%	Paper 2
Organisations	Startups, Team dynamics, Paradigm shifts	48/48	100%	Paper 2

Source: There Is Only One Way to Grow — A Universal Sequence in Dissipative Processes (2026)

Quantitative Tests — Three Biological Systems at 16-Position Resolution

Three independent biological systems tested at full 16-position resolution. Each position scored CLEAN (unambiguous match), PARTIAL (match with degeneracy), or FAIL. All three score 14/16 CLEAN, 2/16 PARTIAL at positions 14–15 — consistent with the conservation degeneracy predicted by the framework.

Erythropoiesis (red blood cell formation)

14 CLEAN · 2 PARTIAL · 0 FAIL

Adaptive immune response

14 CLEAN · 2 PARTIAL · 0 FAIL

Cancer hallmarks (Hanahan & Weinberg)

14 CLEAN · 2 PARTIAL · 0 FAIL

Conservation degeneracy at positions 14–15. The framework predicts that Conservation positions 14 (Surveillance) and 15 (Compensation) exhibit degeneracy — the system can accomplish the same structural function through multiple biochemical routes. This is confirmed across three independent biological systems. Conservation permits functional redundancy precisely because maintaining the system matters more than the specific mechanism used.

Source: The Third Mechanism — Quantitative Tests of the Universal Sequence (2026)

Physics — The Lorentz Invariant and Cosmological Constant

The two operations (hold and cross) act on two axes (time and space). Their products and ratios recover the foundational results of physics with no additional assumptions.

Result	Derivation	Predicted	Known value	Match
Lorentz invariant	s² = H_t × C_t = c²t² − x²	Minus sign derived, not postulated. c is the balance point of the time axis (τ = 1).	Special relativity postulate	Exact
Cosmological constant ratio	Λ_obs/Λ_QFT ~ (t_Planck/t_universe)²	1.54×10⁻¹²²	~10⁻¹²²	0 free parameters

The cosmological constant problem — the 10¹²² discrepancy between QFT's prediction and the observed vacuum energy — is a regime mismatch. QFT computes in the Conservation regime (no temporal coherence constraint). The observable universe is in the Encounter regime (time axis active). Two measured inputs: the Planck time and the age of the universe. No free parameters.

Source: How a Framework for Cancer Treatment Derived the Lorentz Invariant (2026) · Priority: 10.5281/zenodo.19592156

Physics — Five Standard Model Constants from Integers and π

Five experimentally measured constants of the Standard Model, each derived from the cycle geometry with zero free parameters. Every input is an integer derived from the two-operation structure.

Constant	Formula	Predicted	Published	Match
Cabibbo angle λ	π/14	0.22440	0.22530	0.4%
Wolfenstein A	1/(1+λ)	0.8168	0.8117	0.6%
Unitarity triangle	sin(π/8)	0.38268	0.38260	0.02%
CP phase δ	8π/21	68.57°	68.5°	0.1%
Baryon asymmetry η	(2/π)J²	6.04×10⁻¹⁰	6.10×10⁻¹⁰	~1%

The SM prediction problem. The Standard Model's own prediction for the baryon asymmetry (via electroweak baryogenesis) is off by ten orders of magnitude. This prediction is off by ~1%, using no free parameters.

Source: η = (2/π)J² — Baryon Asymmetry and Five SM Constants from Cycle Geometry (2026)

Drug Response Prediction — Generalisation Error

The honest test: leave-one-out cross-validation. Each evidence entry is removed from the calibration table before prediction — the formula predicts it from agent structure alone, without seeing the observed data. Mean absolute error across all held-out entries. Calibrated entries (where M is derived from observed data) are not predictions and are not counted here.

3.6%

LOO-MAE across oncology

0.808

Patient AUC — NSCLC

0.889

Patient AUC — melanoma

Patient-level AUC from FULCRUM CDS — the clinical decision support tool built on the same formula. Two measurements per patient, zero machine learning. Outperforms PD-L1 (AUC 0.64) and oncologist judgment (AUC 0.72) at predicting immunotherapy response.

Loading novel predictions from API...

Source: The Geometry of Intervention (2026) · Run predictions →

Four-State Observer Protocol — Eight Medical Domains

The Observer function degrades through a fixed four-state sequence. Each state has a distinct mechanism, distinct biomarkers, and requires a different intervention. The protocol has been validated across eight medical domains including both immune and non-immune systems.

STATE 1

Attend

Observer functional. System self-corrects. Standard protocols work.

→

STATE 2

Drain

Observer overloaded. Monitoring intact but action capacity depleted. Reduce load.

→

STATE 3

Withdraw

Observer disconnected. Recognition present but action pathway severed. Reconnect.

→

STATE 4

Confusion

Recognition repertoire decayed. System cannot identify targets. Rebuild from scratch.

Domain	System	Observer	States confirmed	Key finding
Oncology	Immune checkpoint	T-cell surveillance	4/4	Withdrawal ≠ Confusion — checkpoint inhibitors work in State 3, fail in State 4
Autoimmunity	Multiple sclerosis	Regulatory T-cells	4/4	Treg exhaustion follows same cascade as anti-tumour T-cells
Autoimmunity	Type 1 diabetes	Peripheral tolerance	4/4	Window for tolerance restoration closes at State 4
Neurodegeneration	Alzheimer's disease	Microglia surveillance	4/4	DAM phenotype = State 3 Withdrawal, not activation
Critical care	Sepsis	Innate immune function	4/4	Immunoparalysis = State 4 Confusion — mortality diverges from State 3
Cardiology	Heart failure	Cardiac autonomic regulation	4/4	HRV decay tracks four-state cascade
Hepatology	Liver cirrhosis	Hepatic stellate cells	4/4	Non-immune proof case — fibrotic cascade follows same four states
Infectious disease	Tuberculosis	Granuloma integrity	4/4	Latent → active reactivation = State 3 → State 4 transition

Recovery must reverse the sequence. A system in State 4 (Confusion) cannot be treated as State 3 (Withdrawal). Recovery requires rebuilding the recognition repertoire before reconnecting the action pathway. Skipping states is the most dangerous intervention error the framework identifies.

Source: The Four-State Observer Protocol (2026)

Structural Physics Derivations

Predictions derived from the geometry that match known physical facts. Each follows from the two-operation structure without additional assumptions.

Prediction	Derived from	Predicted	Observed	Status
Number of forces	2 ops × 2 states	4	4	EXACT
Number of generations	4 − 1	3	3	EXACT
Force pairing (G,S)+(EM,W)	Same-axis coupling	Unique pairing	3/3 tests	EXACT
EW unification succeeds	Same axis	Succeeds	Confirmed	EXACT
Grand unification fails	Different axes	Fails	Not observed	EXACT
Proton decay absent	Cross-axis forbidden	Absent	Not observed	EXACT
Φ > 0 (generative)	Cycle geometry	Confirmed	Brunner 2020	EXACT
Weinberg angle (tree)	1/4	0.2500	0.2312 (M_Z)	8% (tree)
Koide formula	2/3	0.66667	0.66663	0.006%

Open problems. Three dimensionless constants remain unresolved: the fine structure constant α ≈ 1/137 (partial: two-vertex coupling = 1/(5φ⁶) = 1/89.7 derived exactly; gap to 1/137 open), the proton–electron mass ratio ≈ 1836, and neutrino mass squared differences. These are framed problems, not unknown unknowns.

Source: η = (2/π)J² — Baryon Asymmetry and Five SM Constants (2026) · Lorentz Invariant paper (2026)

Prediction Scoreboard

Forward-facing structural predictions — made before confirmation, dated, falsifiable. The geometry either predicts correctly or it does not. Score updates as predictions resolve.

Confirmed

Open

Refuted

Date	Subject	Prediction	Timeframe	Status
Apr 2026	Block — organisational redesign	Speed without maintenance is faster decay. Four structural violations will produce: compliance incidents, model drift, internal/external perception divergence, and eventual quiet rebuild of the conservation function removed in the restructuring.	12–48 months	OPEN

Full record and Substack pieces: raimo.substack.com