Stack resolved probability forecasts

What it's for

Lets Gitrevio combine deadline, incident, and execution-risk models only when the ensemble proves better out of time.

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.

{
  "bootstrap_draws": 200,
  "model_names": [
    "delivery_model",
    "reliability_model",
    "weak_model"
  ],
  "observations": [
    {
      "id": "stack-0",
      "outcome": 0,
      "period": 0,
      "predictions": {
        "delivery_model": 0.05,
        "reliability_model": 0.5,
        "weak_model": 0.45
      }
    },
    {
      "id": "stack-1",
      "outcome": 1,
      "period": 1,
      "predictions": {
        "delivery_model": 0.95,
        "reliability_model": 0.5,
        "weak_model": 0.55
      }
    },
    {
      "id": "stack-2",
      "outcome": 0,
      "period": 2,
      "predictions": {
        "delivery_model": 0.5,
        "reliability_model": 0.05,
        "weak_model": 0.45
      }
    },
    {
      "id": "stack-3",
      "outcome": 1,
      "period": 3,
      "predictions": {
        "delivery_model": 0.5,

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": [
    "Every model probability was frozen before its outcome resolved, all models predict the same binary event and horizon, and chronological validation represents future deployment.",
    "Simplex stacking forms a convex probability mixture; it cannot repair shared label leakage, selective resolution, concept drift, or an event absent from every component model.",
    "The benchmark is selected only on training history, and the bootstrap interval measures held-out row uncertainty conditional on the fitted weights and model set.",
    "A validated stack improves forecast accuracy for aggregate decisions; model weight is not causal attribution or evidence about an individual contributor."
  ],
  "decision": "forecast_stack_validated",
  "method": "chronological_simplex_probability_stacking_bootstrap_v1",
  "models": [
    {
      "model": "delivery_model",
      "stacking_weight": 0.5,
      "training_log_loss": 0.37222,
      "validation_log_loss": 0.37222
    },
    {
      "model": "reliability_model",
      "stacking_weight": 0.5,
      "training_log_loss": 0.37222,
      "validation_log_loss": 0.37222
    },
    {
      "model": "weak_model",
      "stacking_weight": 0,
      "training_log_loss": 0.597837,
      "validation_log_loss": 0.597837
    }
  ],
  "sample": {
    "bootstrap_draws": 200,
    "confidence_level": 0.9,
    "models": 3,
    "observations": 400,
    "optimization_iterations": 46,
    "training_rows": 280,
    "validation_rows": 120,
    "weight_regularization": 0.01
  },
  "stack": {
    "benchmark_model": "delivery_model",
    "benchmark_validation_log_loss": 0.37222,
    "improvement_interval": [
      0.007707,

How it works

Forecasting & survival — Estimate when something completes or fails, with censoring and unresolved work handled honestly rather than dropped.

1. 1 Fit convex weights to frozen probability forecasts on chronological training history and require bootstrap-validated log-loss improvement over the training-selected best component on future outcomes.
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": "fit convex weights to frozen probability" }
  → finds "stack_resolved_probability_forecasts"

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

gitrevio_capability_run
  { "capability_id": "stack_resolved_probability_forecasts", "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 Measurement integrity →