The study looked at different ways to predict changes in financial markets during a crisis. They tested various methods to see if using skewed distributions could make predictions more accurate. The results showed that, except for one model, using skewed distributions did not significantly improve predictions. In fact, models using non-skewed distributions performed better overall. The study also found that a specific model called APARCH with six different error distribution functions was the most effective for predicting market volatility. Overall, having a return with skewness, leptokurtic, and thick tail did not necessarily lead to better predictions in a skewed error distribution.