Monte Carlo Lab — Paste, Auto-Fit & Correlations

Paste from Excel / CSV → Auto-Fit

Paste cells (tab-separated) or CSV. First row can be headers. We’ll fit each numeric column to the best distribution using AIC and detect correlations (Spearman ρ).

First row is header

Corr threshold

Fit

Variables

Use Corr Group + ρ to bind variables together (Iman–Conover rank correlation). Leave blank for independence.

Variable Name	Distribution	Parameters	Corr Group	ρ (-0.99..0.99)	Actions

Trials

Seed

Bins

Result:

Target

Sense

Quick percentiles:

P5 P10 P50 P90 P95

Results

Percentile 50%

Value: —

JSON Import / Export (LLM)

Import or export the full simulation setup as JSON. Use this to interface with ChatGPT, Claude, Gemini, or any LLM.

🤖 LLM / AI Instructions

Copy the instructions below and paste them into any LLM along with your project description. The LLM will return JSON you can paste above and click Import JSON.

Return ONLY valid JSON. Do not include markdown fences, commentary, or explanations.

Schema: pmh-montecarlo/v1

Return this structure:
{
  "schema": "pmh-montecarlo/v1",
  "settings": {
    "trials": 10000,
    "seed": 12345,
    "bins": 40,
    "aggregation": "sum",
    "expression": "",
    "target": null,
    "sense": "ge"
  },
  "variables": [
    {
      "name": "Variable Name",
      "distribution": "triangular",
      "parameters": [10, 20, 30],
      "correlationGroup": "",
      "rho": 0
    }
  ]
}

Rules:
- "distribution" must be one of: "triangular", "pert", "normal", "uniform", "constant", "exponential", "lognormal", "weibull", "gamma", "poisson", "binomial", "zilognormal", "zigamma".
- Parameter meanings by distribution:
  - triangular: [min, mode, max]
  - pert: [optimistic, most_likely, pessimistic]
  - normal: [mean, standard_deviation]
  - uniform: [min, max]
  - constant: [value]
  - exponential: [lambda]
  - lognormal: [mu, sigma]
  - weibull: [shape, scale]
  - gamma: [shape, scale]
  - poisson: [lambda]
  - binomial: [n, p]
  - zilognormal: [p_zero, mu, sigma]
  - zigamma: [p_zero, shape, scale]
- "parameters" must be an array of numbers matching the distribution.
- "correlationGroup" is an optional letter (e.g. "A", "B") to link correlated variables. Leave "" for independent variables.
- "rho" is the Spearman rank correlation coefficient (-0.99 to 0.99). Only meaningful when correlationGroup is set.
- "aggregation" must be one of: "sum", "avg", "min", "max", "expr".
- If aggregation is "expr", provide a custom expression in "expression" using variable names and operators (+, -, *, /, ^, parentheses). Example: "0.4 * CostA + 2 * TimeB".
- "target" is an optional numeric threshold for probability calculation. Set null if not needed.
- "sense" is either "ge" (P(Result >= Target)) or "le" (P(Result <= Target)).
- "trials" should be between 1000 and 100000.
- Generate as many variables as are relevant to the project context.
- Use realistic parameter ranges based on the project description.
- Assign correlation groups when variables are logically related (e.g. labor costs and schedule duration).