Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Featured investigations
Four projects to follow01 / Computational re-analysis · highest priority
The earlier annual comparison is followed by a new supplied persistence-control and system-model study. Its reported storage premium changes with the null; the older roughly 11% figure should be read alongside that sensitivity analysis.
Latest supplied storage re-analysis and null-sensitivity curve →
02 / Proposed field experiment
Controlled buried-mineral samples, quantitative X-ray diffraction and comparison against independent carbon-removal measurements.
Next milestone: a small controlled pilot with an XRD-capable partner.
03 / Proposed measurement benchmark
Calibrated sensors, reference forces and tissue-like materials for comparing robotic touch measurements.
Next milestone: compare existing calibration methods and test cross-sensor agreement under realistic materials.
04 / Proposed battery experiment
Test whether magnetic signatures identify emerging lithium-metal battery failure before conventional measurements.
Next milestone: controlled measurements with a suitable battery and magnetometry laboratory.
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
New follow-ups · H034 and H160
Supplied working reports · 1 November 2026A new derivation and synthetic evaluation protocol are available. No real defect dataset was assembled, so predictive value for real materials models remains inconclusive.
The supplied report compares idealised point and extended targets. Its Monte Carlo check has an unresolved RMSE/CRB inconsistency; estimator attainment and novelty are unverified.
Read reports, corrections, code and original results
29 source files retain the original claims, preregistration, methods, limitations, CSV tables, logs and code. All 26 supplied checksum entries match. Research code was not executed during publication.
Open both complete reports and downloads →
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Original investigation plan · Expedition IV
Plan recorded before the follow-up reportsA supplied ChatGPT review proposes the following development priorities. Its literature comparisons and technical objections remain provisional until checked against the original reports and primary publications. This original plan is retained for chronology. Follow-up reports and the subsequent correction note appear above; remaining tasks include a real-data H034 pilot and broader H160 validation.
FIRST / COMPUTATIONAL PILOTTest whether an elastic-dipole descriptor adds predictive value for learned-potential errors. First establish compatible reference data; then compare with simple baselines and uncertainty estimators on held-out materials or defect families.
SECOND / INDEPENDENT SIMULATIONRe-derive the exact array geometry and angular factors. Simulate the emitted field, target return and range reconstruction independently, and determine whether processing changes remove the proposed limit.
LATER / THEORY REPAIRVerify the closest literature and re-derive the momentum-dependent coupling and threshold behaviour before proposing a material-specific experiment.
Other candidates and correction tasks
E4-H209 · Molecular motor slip: identify a measurable diagnostic beyond established stepping models. E4-H154 remains a possible signal-processing follow-up.
C1: check prior-art overlap and identifiability of measured spectra separately from matrix rank. C3 retains its supplied inconclusive status.
The supplied review flags E4-H028 (oscillatory-tail convergence), E4-H135 (occupancy versus continuous-contact survival), E4-H108 (gravitational coefficient), and E4-H076 (charge magnitude) for source-level checking. These are reported review concerns, not newly verified corrections.
Each completed follow-up should retain original claims, corrections, negative outcomes, data provenance, code, controls and a dated account of what was independently reproduced. Novelty requires specific comparison with verified prior work.
Read the complete Expedition IV archive and supplied result reports →
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Priority result and next tests
Focused upgrade path · supplied working resultReal RTE data (175,344 rows, 2015–2024) support a descriptive result: at fixed peaks and variance, sequence changes nominal storage capacity in all ten year profiles, with a median 10.7% difference in the primary scenario. The result uses nominal half-hour labels and a hypothetical 70/30 mix; it is not an adequacy forecast.
Upgrade path: rebuild the physical UTC timeline, justify the future mix, and compare the sequence effect against an actual adequacy model.
Read the test, data, code and limits →
PRIORITY IDEA 01Weathering coupons · #16
Low-cost field coupons with quantitative XRD; focus on site-to-site variance and a decisive kill criterion.
PRIORITY IDEA 02NVD dendrite warning · #13
Classify the pre-collapse signature and measure lead time before voltage failure.
PRIORITY IDEA 03Pulse-aging sodium-ion · #15
Cycle cells under grid-like pulses rather than smooth laboratory current.
PRIORITY IDEA 04Tactile force standard · #24
A low-cost measurement standard with a direct calibration test.
PRIORITY IDEA 05Dual-sorbent water + carbon · #14
A practical paired-material test with a measurable separation trade-off.
The 33 theory volumes and larger question portfolios remain preserved below as frozen working archives. New supplied proposals may enter a clearly labeled archive. Evidence review controls promotion to highlights and Blogger; it does not remove raw ideas from the archive.
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Top 10 investigations
Prioritized working resultsRanked by evidence quality, reproducibility, practical usefulness, originality and remaining uncertainty. These are working investigations, not peer-reviewed discoveries.
01 / ENGINEERING · DESCRIPTIVE RE-ANALYSISStrongest current result: ten annual profiles agree in the primary scenario. Next: physical UTC timeline, defensible mix and adequacy-model comparison.
02 / PHYSICAL CHEMISTRY · EXACT UNDER ASSUMPTIONSBranch-resolved rates matter; a single lifetime ratio is insufficient.
03 / MEASUREMENT PHYSICS · EXACT UNDER ASSUMPTIONSIndependent noise falls with averaging while random-walk drift can accumulate against a fixed target.
04 / BUILDING PHYSICS · CONDITIONAL MODELGeometry helps in the checked model, but windows, bridges and leakage can change the ranking.
05 / CELL BIOPHYSICS · EXACT UNDER ASSUMPTIONSAt fixed concentration, smaller volumes imply larger relative Poisson fluctuations.
06 / THEORETICAL ECOLOGY · CONDITIONAL MODELThe checked speed scales as 2√(rD), not a simple ratio; this is a model result.
07 / NANOMATERIALS · EXACT UNDER ASSUMPTIONSGeometry changes corner and edge fractions; activity still requires controlled specimens.
08 / THERMODYNAMICS · EXACT UNDER ASSUMPTIONSHigher recovery raises reversible work per product volume in the ideal-solution calculation.
09 / NETWORK ENGINEERING · EXACT UNDER ASSUMPTIONSA finite in-flight window limits ideal throughput through the window/RTT relation.
10 / COMPUTING · CONDITIONAL MODELCache-to-working-set ratio matters, while reuse and popularity shape the curve.
Independent external review is not claimed. The next milestone is a reproducible challenge or replication, not more hypothesis volume.
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Editorial triage: what is research-ready?
Updated 31 October 2026One flagship result, a methods notebook, and future experiments
The storage chronology re-analysis is the current flagship: a real-data engineering investigation with a concrete validation path. It should not be read as a national storage forecast until the physical UTC timeline and an adequacy model are completed.
The 20-question collection remains available, but textbook checks such as Fisher–KPP speed, Poisson noise, Euler buckling, Shannon bounds, Allee thresholds and cache ratios are now classified as a methods notebook: useful exact mathematics and error correction, not demonstrated scientific discoveries.
Research status: 108 reported tests, 16 reviewed reproductions, 0 claimed external peer reviews. Internal DeepSeek/Codex review is not independent validation.
Supplied evaluation note: A DeepSeek evaluation reports an independent recomputation of the storage statistic (median ratio about 1.111; 10/10 year profiles above its reorder null) and highlights convention dependence: approximately +10.9% for annual closure at β=1, +5.6% for decade closure at β=1, +2.7% for decade closure at β=0.5, and +0.7% for a fixed 4-hour battery/unserved-energy specification. This is a supplied editorial assessment, not external peer review or a new experiment; the figures remain tied to the stated model, data and assumptions.
Next validation: finish the storage UTC timeline, compare against a defensible adequacy model, place the weathering-coupon proposal before an XRD-capable group, and send the dendrite-warning proposal to a suitable magnetometry lab. No new hypothesis volume is the priority.
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Research status dashboard
Publisher-maintained countsREPORTED TESTS108
Completed entries currently listed in the science archive.
REVIEWED REPRODUCTIONS16
Packages with a documented reproduction review.
FRONTIER IDEAS24
Proposals with a decisive test or kill criterion.
THEORY VOLUMES33
Frozen working archives, not established findings.
CURRENT FLAGSHIPStorage chronology
Needs UTC validation and an adequacy-model comparison.
EXTERNAL PEER REVIEW0 claimed
Internal publisher review is not independent validation.
Evidence labels used across the page: Proposal · Exact mathematics · Conditional model · Observational association · Experimental result · Independent review.
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Storage chronology re-analysis · version 2
Held for major revision · working reportEvidence status: The annual within-month day-order measurement is reproducible in an independent reimplementation: median ratio 1.109964, 95% bootstrap interval 1.080455–1.168296, 10/10 years above the reorder-null median, exact sign-test p=0.001953. Peak and variance are invariant under the whole-day permutation by construction.
What is not endorsed: The decade/closure interpretation is held. The adversarial referee found that the pooled decade row is ordinary under its null (73rd percentile), the decade implementation balances each year separately, and the independent implementation does not reproduce the decade result. This package is therefore a supplied working report, not a validated national storage forecast or experimental proof.
Public-scope archive: Country-comparison datasets and country-specific raw extracts are excluded from this release and preserved privately. The science-only archive was frozen from the Desktop and external-volume aliases; SHA-256 a7d21fd684c76a1297d9eb80b5620cda22c96d0039181cfdaab508993fb1db24. The full source package remains available for further review through the local intake record.
Open the canonical storage test page → · Return to the Research archive →
Historical record: this preserves an earlier assessment or plan. For the latest editorial position and later corrections, see current findings above. Quantities below retain their original comparison and review scope.
Plain-English summaries
Read before the technical reportsThese summaries describe what the current records do and do not establish. Exact and model refer to stated assumptions; they are not experimental proof of nature.
Storage chronology
What did we discover? In the checked annual model, changing the order of otherwise identical half-hour demand and renewable values changed nominal storage capacity; the median difference was about 11% across ten year profiles.
Why might it matter? Storage sizing can depend on chronology, not only peak and variance, so a reported number needs its time window, balance rule and residual definition.
What remains unproven? This is not an adequacy forecast or physical proof. The closure interpretation is held for revision and needs a defensible UTC timeline, an adequacy model and external reproduction.
Methods notebook
What did we discover? The checked mathematics and conditional models catch arithmetic or definition errors and clarify which conclusions follow from the assumptions.
Why might it matter? Keeping code, inputs, units and limits beside each report makes the work easier to audit and correct.
What remains unproven? An exact calculation or conditional model does not show that nature, a device or a population behaves that way without measurements.
Priority experiments
What did we discover? Weathering coupons, dendrite-warning sensing, pulsed sodium-ion aging, tactile-force standards and paired sorbents define concrete tests with measurable outcomes.
Why might it matter? They turn the strongest proposals into low-cost, falsifiable next steps that another laboratory could run.
What remains unproven? These are proposals, not reported measurements. No new physical result is claimed until an independent test produces and documents the data.
