← back to timeline

PWS-FACTORIAL

negative

Persistence-Weighted Selection

2026-03-07 L1 level paper2 0
new
premortem
mock
real
metrics
analyze
review
krit
done

In plain language

If you give an advantage to "older" text fragments (those that survive longer), does selection emerge? No — age-weighting does not create evolutionary pressure. The system remains random.

What we found

No selection signal above noise

No selection signal above noise at any persistence weighting level. All 7 predictions refuted or borderline. The text substrate does not support persistence-weighted selection — fragment age has no meaningful connection to catalytic properties. Chemistry-level selection mechanisms (CVS, AMA, PWS) are exhausted. Time to move up a level.

Predictions we made before running

  • ✗
  • ✗
  • ✗
  • ✗
  • ✗
  • ✗
  • ◐

0/7 confirmed · 6 refuted · 1 borderline

▶ Technical details

Research hypothesis

Weighting reaction pair selection by fragment age (persistence) amplifies endogenous selection in the text substrate. Older fragments participate more often, creating a feedback loop between persistence and catalytic connectivity.

Experimental setup

Type: factorial

Condition Parameters
C0_uniform alpha: 0, label: Uniform random selection (control)
C1_weak_pws alpha: 0.5, label: Weak persistence weighting
C2_moderate_pws alpha: 1, label: Moderate persistence weighting
C3_strong_pws alpha: 2, label: Strong persistence weighting
C4_deterministic alpha: inf, label: Deterministic oldest-first

Factors: alpha (0, 0.5, 1, 2, inf)

Parameters

pool_size
40
n_replicates
3
model
gemini-2.0-flash
temperature
1
chemistry selection persistence factorial