Each rule is assigned a weight. The score is estimated as the weighted average of the percentage of errors for each rule.

For example, for rules r1, r2, r3 with corresponding weights w1, w2, w3, if the percentage of errors is p1, p2, p3, then the score is calculated by the following formula:

score = (w1*p1 + w2*p2 +w3*p3) / w1+w2+w3

You can find the weights of the various rules used in a test inside its report.