Z-Score

Z-Score expresses a data point's distance from the mean in units of standard deviation. For price, Z = (current price − N-period mean) / N-period standard deviation. A Z-Score of +2 means price is two standard deviations above its recent average; in a normal distribution, this occurs only about 2.5% of the time.

Mean-reversion strategies use Z-Score as the primary deviation filter: entry at |Z| > 2 long (if negative) or short (if positive), exit at |Z| < 0.5 or when Z crosses zero. The approach scales across asset classes because normalization removes the need to calibrate absolute thresholds per instrument. Limitations: Z-Score assumes a roughly normal return distribution, which breaks down during volatility regimes (Z-scores that should "never" occur occur frequently in tail events). Strategies that worked at |Z| > 2 during calm markets get destroyed when prices deviate to |Z| > 4 or 5 before reverting. Pairing Z-Score with a regime filter (VIX-based, rolling volatility) dramatically improves robustness.