Optimize correlated experiment sequence

What it's for

Moves adaptive experimentation from independent bandit arms to portfolio learning, where one trial can reduce uncertainty about several related interventions.

What you give it

Every tool is a pure function: it computes from what you pass it and reads nothing on its own. Inputs split into evidence you supply from your connected systems, calibration your team owns, and numerical controls that affect precision but never the result's meaning.

{
  "correlations": [
    {
      "correlation": 0.2,
      "intervention_a": "safe",
      "intervention_b": "risky"
    }
  ],
  "experiment_horizon": 1,
  "interventions": [
    {
      "deployment_cost": 0,
      "deployment_scale": 100,
      "experiment_cost": 100,
      "id": "safe",
      "max_measurements": 1,
      "observation_standard_deviation": 0.5,
      "prior_mean_effect": 1,
      "prior_standard_deviation": 0.1
    },
    {
      "deployment_cost": 0,
      "deployment_scale": 100,
      "experiment_cost": 1,
      "id": "risky",
      "max_measurements": 1,
      "observation_standard_deviation": 0.2,
      "prior_mean_effect": 0.8,
      "prior_standard_deviation": 1
    }
  ],
  "quadrature_points": 7
}

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.

{
  "assumptions": [
    "Intervention effects and experiment observations follow the supplied correlated Gaussian model.",
    "Experiments create information only; deployment value is realized after experimentation stops.",
    "Experiment and deployment costs use the same value units as scaled intervention effects.",
    "Production rollout still requires sequential monitoring for harm, interference, and model drift."
  ],
  "computation": {
    "expanded_belief_states": 1,
    "experiment_horizon": 1,
    "quadrature_points": 7
  },
  "deploy_now_value": 100,
  "first_experiment_action_values": {
    "risky": 127.4024,
    "safe": 0
  },
  "first_experiment_learning_spillovers": [
    {
      "intervention_id": "safe",
      "posterior_variance_reduction_fraction": 0.0385
    },
    {
      "intervention_id": "risky",
      "posterior_variance_reduction_fraction": 0.9615
    }
  ],
  "method": "gauss_hermite_correlated_normal_experiment_dp_v1",
  "net_expected_value_of_sample_information": 27.4024,
  "optimal_policy_value": 127.4024,
  "policy_tree": {
    "branches": [
      {
        "next": {
          "decision": "deploy",
          "intervention_id": "safe",
          "policy_value": 92.6448,
          "posterior_deployment_value": 92.6448
        },
        "observation": -3.0247,
        "quadrature_probability": 0.0005
      },
      {
        "next": {

How it works

Constrained optimization — Pick the best feasible option under real limits — budget, headcount, dependencies, capacity — rather than ranking a list and hoping it fits.

1. 1 Sequence pure-learning experiments over correlated intervention effects using conjugate Gaussian updates, Gauss-Hermite lookahead, early stopping, and terminal deployment value.
2. 2 Evaluate the method-specific diagnostics and gates returned by the function, then abstain unless the declared decision clears them under locally governed thresholds.

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 20 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": "sequence purelearning experiments over correlated intervention" }
  → finds "optimize_correlated_experiment_sequence"

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

gitrevio_capability_run
  { "capability_id": "optimize_correlated_experiment_sequence", "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 Causal evidence & experiments →