Mismatch Simulator

The epigenetic trap, made explicit. When does anticipatory environmental memory stop paying for itself and start costing more than it returns? Framework Question 8, as a model you can push on.

This is a toy model. Its parameter ranges are anchored to real persistence measurements from the Conditioning Atlas (109 extracted experiments); its functional forms are assumptions chosen to be interpretable, not derived from marine invertebrate physiology and not fitted to any dataset. It exists to sharpen the question and generate testable predictions. No number it produces should appear in a manuscript as a result. See MODEL.md for the full specification and limitations.

Scenario

Parameters

One lineage through time

Environment, the induced memory state it drives, and the resulting performance — against the same lineage carrying no memory at all, experiencing the identical environmental sequence.

Environment E(t) Memory state M(t) Performance, with memory Performance, no memory

Where the trap is

Advantage of having memory across rate of change × memory half-life, at the current settings for every other parameter. Blue means memory pays for itself; orange means the lineage would do better with no environmental memory at all. The black line is the trap boundary. Horizontal markers show where real Conditioning Atlas studies sit on the half-life axis — remembering that those are censored bounds, not measurements.

Memory pays Trapped Trap boundary Current settings