What are weighted signals in spam filtering?

Weighted signals are the individual factors that contribute to a spam score, each assigned a point value reflecting its importance. Not all spam indicators are equal, and weighting reflects that reality.

A missing DKIM signature might add 1 point. A blocklisted sending IP might add 4 points. Matching a known phishing pattern might add 5 points. The weights are calibrated based on how strongly each signal predicts spam.

Weights are derived from analysis of real spam campaigns. If a particular pattern appears in 90% of spam but only 1% of legitimate mail, it receives a high weight. Patterns that appear in both spam and legitimate mail receive lower weights.

Administrators can customize weights for their environment. If a particular rule generates too many false positives, its weight can be reduced or the rule disabled entirely.

Weighted signals are like judges with different expertise. The security expert's opinion carries more weight than the trainee's.