Construct calibration shared evidence graph

What it's for

Maps which Gitrevio analytics succeed and fail together across operating environments, revealing reusable calibration infrastructure while preserving function-specific proof and privacy boundaries.

What you give it

Inputs split into evidence read from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "bootstrap_draws": 200,
  "minimum_shared_environments": 6,
  "observations": [
    {
      "environment_id": "environment-0",
      "function_id": "forecast",
      "id": "forecast-0",
      "reliability_weight": 1,
      "standardized_loss_improvement": 0
    },
    {
      "environment_id": "environment-0",
      "function_id": "policy",
      "id": "policy-0",
      "reliability_weight": 1,
      "standardized_loss_improvement": 0
    },
    {
      "environment_id": "environment-1",
      "function_id": "forecast",
      "id": "forecast-1",
      "reliability_weight": 1,
      "standardized_loss_improvement": 1
    },
    {
      "environment_id": "environment-1",
      "function_id": "policy",
      "id": "policy-1",
      "reliability_weight": 1,
      "standardized_loss_improvement": 2.1
    },
    {
      "environment_id": "environment-2",
      "function_id": "forecast",
      "id": "forecast-2",
      "reliability_weight": 1,
      "standardized_loss_improvement": 2
    },
    {
      "environment_id": "environment-2",
      "function_id": "policy",
      "id": "policy-2",
      "reliability_weight": 1,

What you get back

This is the actual output of running the example above — computed by the same function the platform calls, not an illustration.

{
  "configuration": {
    "bootstrap_draws": 200,
    "false_discovery_rate": 0.1,
    "minimum_absolute_correlation": 0.5,
    "minimum_selection_probability": 0.8,
    "minimum_shared_environments": 6,
    "seed": 7
  },
  "connected_components": [
    [
      "forecast",
      "policy"
    ]
  ],
  "decision": "shared_calibration_design_candidates_available",
  "guardrails": [
    "Edges describe stable co-movement of standardized held-out loss improvement across represented environments. They do not prove shared causality, exchangeability, identical coefficients or permission to pool raw data.",
    "Use connected components to propose shared calibration infrastructure or hypotheses, then preregister and validate transfer prospectively. Each function retains its own readiness and safety gates.",
    "Environment resampling and false-discovery control limit selection noise but do not repair an omitted common shock, incompatible loss definition, target shift or cross-tenant privacy boundary."
  ],
  "method": "environment_cluster_bootstrap_stability_graph_with_bh_fdr_v1",
  "stable_edges": [
    {
      "bootstrap_interval": [
        1,
        1
      ],
      "fdr_adjusted_significance": true,
      "left_function_id": "forecast",
      "right_function_id": "policy",
      "selection_probability": 1,
      "shared_environment_count": 12,
      "sign_stability_probability": 1,
      "weighted_correlation": 1
    }
  ],
  "summary": {
    "eligible_pair_count": 1,
    "environment_count": 12,
    "function_count": 2,
    "multi_function_component_count": 1,
    "stable_edge_count": 1
  },

How it works

Network & dependency analysis — Trace how load, failure and knowledge propagate through a graph of teams, services or components.

1. 1 Build one standardized, reliability-weighted held-out improvement observation per function and operating environment under a common prospective evaluation protocol.
2. 2 For every sufficiently supported function pair, estimate weighted correlation and resample whole environments to measure practical edge-selection and sign stability.
3. 3 Control pairwise sign discoveries with Benjamini-Hochberg FDR, retain only practically large stable edges, and expose connected components as shared-design hypotheses.

Before you trust it

Every tool in the catalog ships with the conditions under which its answer is meaningful — and the conditions under which it should abstain instead of guessing.

Calibrating it to your org

Calibration workflow

1. 1 Define the management decision, target outcome, aggregate unit, privacy boundary, cadence, and prediction/intervention horizon for this organization.
2. 2 Build a tenant-scoped historical cohort using only information available before each prediction or decision; preserve zero periods, censoring, assignment probabilities, and unresolved outcomes when the method requires them.
3. 3 Estimate statistical parameters on training history, but obtain costs, utilities, risk tolerance, practical-effect thresholds, capacity, and policy constraints from accountable decision owners.
4. 4 Validate on later time windows or held-out aggregate units at the deployment grain, against a simple baseline and the function-specific validation strategy.
5. 5 Deploy only if the returned decision clears evidence, overlap, calibration, robustness, and guardrail checks; warning, unsupported, schema-gap, and fallback decisions are abstentions.
6. 6 Monitor realized outcomes, data drift, coverage, and decision regret; recalibrate at a governed cadence or after a detected regime/definition change, never merely because a stakeholder dislikes the result.

Call it from your AI

You don't wire up 388 tools in your MCP client. The GitRevio MCP server exposes 18 tools, three of which let an agent search the catalog, read a tool's schema, and run it — so the assistant finds this one on its own.

gitrevio_capabilities_search
  { "q": "discover stable candidate relationships for shared" }
  → finds "construct_calibration_shared_evidence_graph"

gitrevio_capability_describe
  { "capability_id": "construct_calibration_shared_evidence_graph" }
  → returns the input schema and agent guidance shown on this page

gitrevio_capability_run
  { "capability_id": "construct_calibration_shared_evidence_graph", "arguments": { ... } }
  → returns the result shown above

Works in Claude Desktop, Claude Code, Cursor, Cline, Continue.dev, Goose and Aider. See the MCP server.

Related tools

See every tool in Analytics assurance & orchestration →