Skip to main content
DRiCE™

How the DRiCE score works

You send five control measurements and get back a score report. The model runs only on our servers; its coefficients are never sent to callers.

How the score is built

  1. InputsFive measurements: EDR coverage, MFA adoption, patch cadence in days, external hygiene and vulnerability density.current step
  2. ModelFrequency × severity → expected loss. Both are logistic today; the final model form waits on a cited methodology.not started
  3. ScoreA 300–850 scale where higher means lower risk. The current coefficients are unvalidated placeholders.not started

Sample report

The sample below is the current model's real output for the score form's default inputs. The scoring API is not public: only signed-in staff and holders of a staff-issued API key can call it.

Unvalidated model: not for decisions. This score comes from placeholder coefficients (v0-unvalidated). No cited methodology or validation data backs them yet.

Inspect sample report JSON
GET /api/v1/score/{id}
{
  "actuarial_risk_score": 821,
  "expected_loss_index": "0.05x",
  "confidence_level": "Unvalidated - placeholder coefficients, not actuarially calibrated",
  "peer_benchmark_percentile": "Unavailable - no peer benchmark dataset",
  "key_strengths": [
    "EDR Coverage",
    "MFA Adoption",
    "Patch Cadence",
    "External Hygiene",
    "Vulnerability Density"
  ],
  "key_weaknesses": [],
  "loss_event_frequency_index": 0.12,
  "loss_event_severity_index": 0.45,
  "incentive_recommendation": false,
  "report_id": "rpt_sample",
  "model_version": "v0-unvalidated",
  "schema_version": "v1"
}

Staff sign in with Google Workspace SSO or an emailed magic link for existing accounts.