The article compares different models for predicting daily exchange rate price changes. They looked at models that assume the changes are related to each other (dependent) and models that assume they are not related (independent). The researchers found that none of the models consistently outperformed the others. In some cases, the models that assume independence were just as good as the more complex models. They also found that the volatility of exchange rates can cluster together, but there are also times when rates change independently. This suggests that future research should focus on developing models that can handle both types of behavior.