McAfee PASCDE-AB-IA Product Guide - Page 50

Scoring Audits Flat unweighted scoring model Since the maximum possible score can vary from audit to audit and from system to system, it is difficult to compare audit scores. The primary use for this scoring model is for comparing historical audit scores on the same system. Flat unweighted scoring model The flat unweighted scoring model computes the score (the number of rules that passed) and compares it against the maximum possible score. McAfee Policy Auditor is preconfigured to use a normalized implementation of the flat-unweighted score model. The maximum possible score is the number of all applicable rules in an audit. For example, if an audit evaluates a system against 283 rules and the system passes 212 of the rules, the flat unweighted scoring model gives the system a score of 212. Another audit might have fewer rules and yield a lower score. This makes it difficult to compare results from different audits. How McAfee Policy Auditor calculates scores Because of the disparity in comparing audit scores, the software is preconfigured to use the flat unweighted scoring model and normalize the final score to a maximum possible score of 100. This allows you to reliably compare an audit with other audits on the same system or between systems with different configurations, such as Windows XP or Windows 7. The software uses this equation to normalize audit scores: audit score = (rules passed ÷ maximum possible score) × 100 This table shows how scores for different audits can be compared using a normalized implementation of the flat unweighted score model. Audit example Audit 1 Audit 2 Maximum possible Rules passed score 283 212 15 14 Flat unweighted audit score 212 14 Normalized flat unweighted audit score 74.9 93.3 Flat scoring model The flat scoring model compares the system score with the maximum possible system score. The maximum possible score is the sum of the weights of all rules in an audit that apply to a system. Rules that do not apply to a system are ignored when calculating the maximum possible score. The actual score is the sum of the weight of all rules that pass. Since the maximum possible score can vary from system to system, scores from systems that have different configurations, such as Windows XP or Windows 7, may not be directly comparable. This model is useful for comparing a system score with its historical scores. Score weighting The flat scoring model allows benchmarks to use weighted scores for each rule. A common example of score weighting is a school test where one question is worth more points than another question. In this example, an audit has a benchmark with two rules. One of the rules is weighted because the audit benchmark developer considered it to be more important than the other rule. 50 McAfee Policy Auditor 6.0 software Product Guide for ePolicy Orchestrator 4.6