RIGGED — Junk Science Factory · Act 1 prototype

You have a dataset with zero real effect. Your job: manufacture a discovery anyway. Then hit Replicate.

The data · whole sample

control n=– treatment n=– tests run: 0

Significance

p = –
run an analysis
p = 0 (certain)0.05p = 1

Null distribution (the actual math)

This histogram is 2,000 real label-shuffles. The shaded tail past the red line is your p-value — computed, not scripted.
▸ villain levers
RETRACTED
Breaking Health News
★ ★ ★ significant! (p < 0.05)
Feasibility prototype · throwaway · rebuild fresh in the hackathon window. Engine = stats.js (real permutation tests).