Online Tool Store
Monte Carlo Simulator
Define inputs with ranges, write a formula, and run a Monte Carlo simulation to see percentiles, a histogram, and the probability of hitting your target.
This works right here on your device — nothing gets sent out.
Math & Science
Monte Carlo Simulator
One input per line: name: low-high, or name: low-mode-high for a skewed triangular estimate.
Combine your input names with + − × ÷ ^, parentheses, and abs(), sqrt(), min(), max().
Runs entirely in your browser — nothing is uploaded.
How the Monte Carlo Simulator works
- List each uncertain input as a name with a low-high range, or a low-mode-high range for a skewed estimate.
- Write the outcome formula that combines those inputs by name — the exact calculation your model runs each trial.
- Pick a distribution shape, a trial count, and a target value with a direction ("meeting or beating" or "staying at or under").
- Run the simulation. Every trial samples fresh values for each input from the chosen distribution and evaluates the formula once.
- Read the percentiles, the probability of hitting your target, and the histogram — not the average, which hides the shape.
The method
A Monte Carlo simulation samples every uncertain input at random, computes the outcome, and repeats — building an empirical distribution instead of one point estimate.
for each trial: sample inputs → evaluate formula → collect outcome; then read percentiles of every collected outcome
Averaging your inputs and computing once gives one number that may be nowhere near the median outcome, particularly when the model multiplies uncertain values together. The outcomes can be copied as a one-line summary or downloaded as a CSV for further analysis.
FAQ
How many trials do I need?
Enough that the answer stops changing. Run ten thousand twice; if the percentiles agree closely, that is sufficient. Tail probabilities need more trials than medians.
Which distribution should I use?
Triangular when you have a low, likely, and high estimate — which is most business cases. Normal when the quantity is a sum of many small effects. Uniform only when you genuinely know nothing but the bounds.
Why not just use the average of each input?
Because the average of a function is not the function of the averages once anything is multiplied or capped. That gap is exactly what a simulation exposes.
How do I write the outcome formula?
Reference each input by the name you gave it on the left of its range, combined with + − × ÷ ^, parentheses, and abs(), sqrt(), min(), max(). For example, with inputs "cost: 90-140" and "demand: 800-1400", a formula of (demand - cost) * 10 computes one simulated outcome per trial.
Is my data sent anywhere?
No. Every trial is sampled and computed in your browser with JavaScript's own random number generator — nothing you type is uploaded.
How we compare
| Feature | Online Tool Store | Numiqo Monte Carlo Calculator | ScopeCone Delivery Forecast | Share Predictions Trading Simulator |
|---|---|---|---|---|
| Custom formula across multiple inputs | ✓ | ✓ | Resamples one cycle-time input only | Fixed win-rate and risk fields |
| Choice of input distribution shape | ✓ | ✓ | Resamples historical data, no shape to choose | Win-rate probability only |
| Probability of hitting your own target | ✓ | ✓ | Fixed P50/P85/P95 only | ✓ |
| Download raw trial outcomes as CSV | ✓ | Not offered | Not offered | Not offered |
Monte Carlo Simulator reports percentiles rather than an average, because the average is the number most likely to mislead a decision.