How smoothness actually predicts bot performance

Every quant textbook tells you to use the Sharpe Ratio: average return divided by volatility. Sounds rigorous. In practice, it's the most over-fit number in trading.

The problem: Sharpe assumes returns are normally distributed. Polymarket returns aren't. The distribution is fat-tailed — a few really big wins, a lot of tiny losses, the occasional catastrophic loss. Sharpe systematically underestimates tail risk in this distribution. We've seen 'great' Sharpe ratios precede a 30% drawdown more than once.

Our 'Smoothness' label is a more honest signal. We compute the standard deviation of monthly returns, divide by the average monthly return, and bucket it into five plain-English labels: Very good, Good, Steady, Choppy, Wild. Higher buckets correlate with longer time-to-recovery from drawdowns — which is what subscribers actually care about.

When you see 'Smoothness: Very good', it means: this bot's monthly returns clustered within 1.5% of average. Bot dipped? Recovered fast. When you see 'Smoothness: Choppy', it means: monthly returns swing 5%+ either way. You'll have months where you wonder if you should cancel; resist.

The takeaway: don't subscribe to the bot with the highest 12-month return. Subscribe to the bot with the smoothness label that matches your stomach.