agentleFS
Sign inStart free
PthahnixPthahnix/De-Anthropocentric-Research-Engine/skills/calibrate-probability-forecast/SKILL.md

calibrate-probability-forecast

Skill504 starsChanged 49 days agoApache-2.0View on GitHub

Aggregate and iteratively recalibrate probabilistic forecasts, preserving uncertainty, calibration evidence, and unresolved forecaster disagreement.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Update the aggregation or calibration rule only where outcome-linked error supports the change.

    line 18
  2. Distinctive:

    Never collapse unresolved disagreement into a single certainty value.

    line 27
  3. Distinctive:

    Do not score forecasts whose outcomes are not yet observable.

    line 29

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

calibrate-probability-forecast

Purpose

Aggregate probabilistic forecasts, compare them with outcomes, and recalibrate without hiding unresolved forecaster disagreement.

Input contract

required: [forecast_records, outcome_records, calibration_rule]
optional: [forecaster_metadata, prior_calibration, aggregation_method]
constraints: [each forecast has a target, probability, timestamp, horizon, and provenance]

Procedure

  1. Align forecasts with realized outcomes at the declared horizon and freeze the evaluation set.
  2. Compute calibration evidence by probability band and forecaster, preserving sample size and missing outcomes.
  3. Update the aggregation or calibration rule only where outcome-linked error supports the change.
  4. Emit recalibrated probabilities, calibration diagnostics, disagreement intervals, and the next review trigger.

Output contract

produces: [calibrated_forecasts, calibration_diagnostics, disagreement_report, review_trigger]
delta_fields: [findings, evidence_updates, hypothesis_updates, uncertainties, decisions, open_questions]

Quality gates

  • A-class gate: declared universe = all closed forecasts in the evaluation horizon; numerator = forecasts with outcome linkage and valid timestamp/probability; batch increment = one closed forecast; stopping reason = calibration error stabilizes or the closed set is exhausted; source references = forecast/outcome IDs and calibration runs; direction/threshold reason = recalibration moves probabilities toward observed frequencies only when the declared scoring rule shows systematic error.
  • Never collapse unresolved disagreement into a single certainty value.

Failure and counterexamples

Do not score forecasts whose outcomes are not yet observable. A changed probability without new outcome-linked evidence is a model revision, not calibration.

Provenance map

  • resolved: futures-calibration

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free