Abstract
To compute option prices under complex models, Monte Carlo simulation is an important mechanism. Martingale control variate methods are useful for variance reduction. They are suitable to cope with various hedging strategies in order to construct the value of hedging portfolio processes. As a result, the variance reduced from a martingale control variate method reflects the effectiveness of corresponding hedging strategy. Based on these simulating experiences, we propose a model-free hedging strategy for empirical studies. As a limiting delta hedging ratio, the strategy is essentially a stop-loss strategy. Our empirical study documents that this strategy is robust under scenarios of high/middle/low volatility in Taiwan or American equity markets and it works particularly well in high and low volatility environments.