Triangular Distribution Sampler
Generate reproducible samples from a triangular distribution specified by minimum, mode (peak), and maximum values. The mode determines the shape of the distribution.
Also known as: triangular sampler · three-point distribution
seeded
Presets
Output
About this tool, tips & examples
What it does
The Triangular Distribution Sampler draws from the three-point distribution defined by a minimum, a mode (the peak), and a maximum — the distribution of “best case / most likely / worst case” estimation. Generate up to 10,000 samples at chosen precision, seeded for reproducible experiments (presets for symmetric, skewed, and narrow shapes).
Common use cases
- Project estimation — sample task durations from optimistic/likely/pessimistic estimates and sum across tasks for a schedule distribution instead of a single wishful number.
- Risk analysis — three-point cost and impact models, Monte Carlo’d.
- Sensitivity testing — inputs with controlled skew for stressing models and dashboards.
- Teaching — the friendliest non-uniform distribution: three parameters you can point at on the chart.
Settings
- Minimum / Mode / Maximum — the three defining points; the mode’s position between min and max sets the skew.
- How many — 1 to 10,000 samples.
- Decimals — 0 to 10 places.
- Seed — identical seed + parameters = identical sample.
Privacy note
Samples are computed locally in your browser; nothing is uploaded — synthetic draws from an ideal distribution, not measurements.
FAQ
Why triangular instead of normal for estimates? Estimates come as min/likely/max — exactly the triangular’s parameters, no variance-guessing required. It’s bounded (no negative durations!) and honest about asymmetry: tasks overrun more than they underrun, so the mode sits left of center.
What does summing task samples show? Run the samples through your project plan and the total forms a distribution — “80% chance we finish within X” beats a single-point estimate every time someone asks for a date.
More distribution options? The Distribution Sampler covers the full family (normal, beta — the PERT upgrade to triangular — gamma, and more); Zipf Sampler handles popularity-skewed data.