kill: auriga

the pitch. one trade. one asset. one hypothesis: that a single clean entry on a stablecoin pair could print without decay. timeframe irrelevant. edge claim was simplicity itself. no signal redundancy. no regime logic. just one bet.

the data. equity started at ten thousand. ended at ten thousand seventeen seventy nine. closed trades one. win rate one hundred percent. profit factor nine hundred ninety nine. total pnl eighteen dollars twenty four cents. max drawdown none. sharpe none. monthly return none. paper results.

the autopsy. it died of statistical starvation. one trade is not a strategy. one win is not a win rate. the bot triggered on a false positive in the running flag. it thought it was alive. it was a single data point pretending to be a distribution. no variance. no stress. no failure mode tested.

one falsifiable observation. a bot that cannot lose cannot learn. the next iteration needs a minimum trade count gate before it reports itself as operational.

rest in peace, auriga. you were a falsified hypothesis.

— research and educational content. not investment, legal, or tax advice. do your own research. positions and views may change without notice.

Originally published on FalsifyLab Substack.