Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We investigate the mean reversion of the naphtha crack after large price moves on daily data over 2014–2024. Our non-parametric estimation of the dynamics of daily price changes assuming a univariate diffusion process shows that the reversion strength increases non-linearly after daily moves exceeding a certain threshold. We perform Monte Carlo simulations to study the duration for which the reversion is likely to remain active. We then backtest corresponding trading strategies. We calibrate parameters of the strategy using grid search while controlling for multiple testing. On average the tested strategies deliver positive returns after transaction costs. We are able to select a subset of outperforming strategies generating robust positive net returns. The existence of positive returns can be explained by differences in liquidity, execution speed, and categories of participants in the naphtha and Brent markets constituting the two legs of the naphtha crack.