Evidence policy
Prevent literature, design intent, estimates and preliminary observations from being misrepresented as validated FlavoRotor results.
Evidence classes
Nine levels of evidence — from testable hypothesis to external verification — each with strict boundaries on what can be claimed.
- 01
HYPOTHESIS
A testable proposition or planned biological effect.
- 02
LITERATURE
A finding from an external source used to justify a hypothesis or treatment range.
- 03
DESIGN
A planned system architecture, nominal specification or CAD-level intent.
- 04
ANALYTICAL_ESTIMATE
A calculation that depends on assumptions and is not a physical measurement.
- 05
BENCH_MEASUREMENT
A measured engineering result tied to a method and system version.
- 06
CALIBRATED
A measured result with calibration, uncertainty and declared validity range.
- 07
CONTROLLED_TRIAL
A preregistered biological or sensory comparison with defined controls.
- 08
REPLICATED
A controlled result reproduced across independent cycles or systems.
- 09
EXTERNAL_VERIFICATION
A result independently reproduced or reviewed outside the team.
Publication gates
Conditions that must be met before any result is published.
- the input actuator is calibrated
- the actual environment is verified
- the protocol is frozen before analysis
- the result reports uncertainty and effect size
- the sensory method matches the claim
- at least one independent replication succeeds
Mandatory distinctions
Critical differences that must never be blurred in any communication.
A chemical difference is not automatically a sensory difference.
A sensory difference is not automatically an improvement.
A preference claim requires a separate consumer test.
A motor command increment is not pump accuracy.
A literature result is not a FlavoRotor result.
A recipe for one crop and cultivar is not universal.