Skip to content
/ drape Public

Detection Reliability And Precision Efficiency (DRAPE) is an index used to assess detection performance

License

Notifications You must be signed in to change notification settings

inodee/drape

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DRAPE: A simple heuristic metric for measuring detection performance

DRAPE stands for Detection Reliability And Precision Efficiency which is basically a unified index (score) that quantifies:

  • Reliability: Are the detections actually finding true attacks? (TP)
  • Precision Efficiency: Are they doing so with minimal noise? (FP)

The algorithm outputs a simple heuristic metric to score detection performance using TP/FP counts. The index score can be either negative or positive with the extremes implying the following characteristics:

  • Negative = noisy, unreliable detection.
  • Positive = precise, efficient, valuable detection.

The index is a composite measure of the individual TP/FP data points which can be used to measure detection/rule performance over time.

It reveals whether a rule is signal-dominant (+) or noise-dominant (–) by inspecting the number of its TP and FPs alerts over a period of time.

How does it work? 

The formula rewards rules that spots a decent amount of threats (TP) with good precision, while penalizing noisy rules, especially those that haven't yet "proven" themselves with enough TPs.

The formula captures the real-world tradeoff between TP and FP figures better than precision or simple raw counts. Most single-number metrics are either:

  • too forgiving of noise (TP-heavy rules look great even when they're a FP factory), or
  • too punishing (precision destroys high-value, high-volume rules).

DRAPE models that tradeoff how detection engineers think in practice:

  • High TP → good
  • High precision → better
  • Extremely noisy rules → bad
  • Slight noise on high-value rules → acceptable

This mirrors actual triage experience so I encourage you to try it out.

Implementations

Outpout Sample (v1.0)

More details

Intro blog post on Detect.FYI website: https://detect.fyi/introducing-the-drape-index-how-to-measure-in-success-in-a-threat-detection-practice-154fd977f731

About

Detection Reliability And Precision Efficiency (DRAPE) is an index used to assess detection performance

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages