Research questions

Research questions are documented in research dossiers.

A research dossier links the research question, hypothesis, mathematical model, simulation, experimental test, result, and revision. The order is not fixed: a revision can lead back to any earlier stage, and not every question passes through every stage.

  • Research question
  • Hypothesis
  • Mathematical model
  • Simulation
  • Experimental test
  • Result
  • Revision

A simulation is not physical measurement evidence. Peer review is not empirical validation. Negative and inconclusive results remain visible.

  • Spectral compression

    How does the geometry of multivariate fluctuations change when a stochastic dynamical system approaches a fold bifurcation under explicitly stated assumptions?

    Test method
    • Analytical · applied
    • Empirical · not investigated
    Result
    Analytically derived and published after peer review. The claims hold under the stated model assumptions. Independent empirical validation has not been demonstrated.
    Quality assurance
    Internal review
    By the author
    Independent replication
    No independent replication documented
    Peer review
    Peer-reviewed: AIP Advances 16, 095208 (2026)
    Publication
    Journal article · DOI 10.1063/5.0337045
    Open dossier : Spectral compression
  • Viability corridor

    Under what conditions does an adaptive system retain enough room for action to correct undesirable developments?

    Synthetic three-state model with a limited resource budget: an illustrative parameter experiment, not a societal forecast.

    Test method
    • Analytical · applied
    • Numerical · applied
    • Empirical · not investigated
    Result
    Exploratory theoretical–numerical model; independent empirical validation not performed.
    Quality assurance
    Internal review
    By the author
    Independent replication
    No independent replication documented
    Peer review
    No peer review
    Publication
    Preprint v1.4 · DOI 10.5281/zenodo.23480636
    Open dossier : Viability corridor
  • Experiment A · Stability is not recoverability

    Planned study, execution not authorized

    Can declining recoverability be measured independently of conventional early-warning signals?

    Test method
    • Empirical · planned
    Result
    Open
    Quality assurance
    Internal review
    Not documented on the website
    Independent replication
    None
    Peer review
    No peer review
    Publication
    Study protocol published · not evidence of results
    Open dossier : Experiment A · Stability is not recoverability

Specialized part · empirical experiments

BENCHEWS DUCK LAB · RESEARCH ATLAS

Experiments and their documented status

Each experiment has a card with its research question and status. The dossier is a research folder with seven sections.

Duck Lab
Research project in preparation
Experiment A
Planned study, execution not authorized
Experiment B
Planned study, execution not authorized
  • Experiment APlanned

    Stability is not recoverability

    Can declining recoverability be measured independently of conventional early-warning signals?

    The prospective study protocol is published, but confirmatory experiments have not started. No added predictive benefit has been established.

    Platform:
    Mali (MicroDuck), virtual first
    Protocol:
    published (Zenodo)
    Open dossier : Experiment A
  • Experiment BPlanned

    From detection to decision

    Does a regret-inspired risk indicator (RRI) add predictive value over a strong baseline model with the same raw information when predicting probe-induced falls of a simulated inverted pendulum?

    The simulation protocol has been prepared as a draft. Design freeze and execution are not approved. No data and no results exist; no benefit of the RRI has been shown.

    Platform:
    Simulated inverted pendulum (two states, one input); no robot experiment
    Open dossier : Experiment B
  • Further experiments will be added here once they are documented.
Status levels
  1. 1 Planned – Protocol specified; execution not authorized.
  2. 2 In preparation – Execution authorized; technical preparation under way.
  3. 3 In progress – Data collection under way.
  4. 4 Analyzed – Confirmatory analysis completed.
  5. 5 Independently validated – External replication or independent review available.