01 / QUESTIONResearch question
EcoBackfire researches the evidence, lets you review the assumptions, then runs the analysis.
How to read the results. Findings are model-derived and conditional: they describe a simple
rebound model, not a measured real system. Researched values are candidates found by automatic web search:
a citation exists, nobody has checked it against the value. Values you override are your own
assumptions, never external facts. Inputs without a source stay illustrative placeholders.
evidence_insufficient means “the available evidence cannot yet confirm or challenge this” — it does
not mean the finding is false.
02 / EVIDENCEResearching
Parameters the engine needs
02 / EVIDENCEEvidence review
Choose, for every parameter, which value the engine should use. Where sources disagree, all are shown and you choose. “Evidence insufficient” means no number was found; none is invented.
03 / CONFIRMATIONConfirm and run
04 / ANALYSISAnalysis
- Decisions confirmed and sent to the engine
- Engine running 0 s
The engine executes, in this order:
- Analyzing candidates (discovery)
- Falsification
- Testing robustness
- Classification
- Generating reproducible artifacts
- Reproducibility re-run
The engine does not report live per-stage progress, so this page does not pretend to; the real per-stage statuses and evaluation counts are shown with the results.
05 / RESULTSResults
Parameters used & provenance
Candidates
Backfire hunt (discovery)
Evidence transparency
Computation budget
Pipeline stages
Limitations & warnings
Artifacts
f comes from a real dataset; every other input is illustrative. Takes several seconds.