It does this by evaluating the prediction errors of The 2 styles in excess of a certain period of time. The test checks the null speculation which the two versions contain the very same efficiency on average, towards the choice that they do not. Should the check statistic exceeds a important value,