Comment: The data appears clustered at mean 0 across deviations and the tails are sparse so modeling them is tricky.
Comment: Use the slider below the chart to zoom in on specific signals.
Comment: Across the 18 signals, 11 finished positive at +60 days (~60% hit rate), suggesting limited predictive edge given the Sharpe distribution
Comment: The sharpe ratios across signals are inconsistent; cannot confirm alpha.
Comment: The strategy does not allow multiple trades at once; signals are washed out for the duration of the holding period. A long entry occurs when the calm-vol regime begins, defined as the short BVOL moving average crossing below the long BVOL moving average after being in the opposite regime the prior day. A short entry occurs when the rising-vol regime begins, defined as the short BVOL moving average crossing above the long BVOL moving average after being in the opposite regime the prior day.
Comment: Results are not great - no alpha in the strategy. Signals are inconsistent.
Comment: The stop loss (hyperparameter) can be tuned with simple grid search optimization on mean return and sharpe. The optimization results show that a stop-loss does improve the sharpe but the returns are still negligible indicating this strategy does not yield true alpha. It is probable that once neutralized for factor risk this strategy becomes even worse. Once again note - the sharpe is calculated on the distribution at the end returns since this is an exploration of the strategy we do not have any allocation or calculate overall PnL vector.