ExtremeHurst is a quantitative detector of extreme investor behavior. Specifically, strong trend-persistent stock price movements are evidence of positive feedback, such as investors irrationally buying because the price is going up…which drives prices higher, etc. In 1994, Parallax found a way to measure extremes of mean reversion and trend persistency by using multiple measurements of the Hurst exponent. We found that these extremes are unstable points in financial markets that are characterized by discrete scale invariance, accelerating price, and log-periodic cycles. At these points new trends begin and end. We built a filter to identify these events on any time scale.