About the service
Point estimates lie by default. A single number for revenue, project cost, or launch date hides the entire range of outcomes that actually matters for your decision. I build probabilistic models in Squiggle — a language designed for estimates and probability distributions — so your forecasts show the full spread instead of one convenient figure.
You get a working Squiggle model tailored to your problem. I encode your assumptions as distributions (lognormal, normal, uniform, beta, mixtures, or point values where it's honest), link them with real task logic, and run the model so you see medians, confidence intervals, and the probability of outcomes you care about. Typical uses: budgeting and cost-benefit analysis, timeline and effort estimation, revenue and growth forecasts, risk calculation and expected value, comparing options under uncertainty.
The deliverable includes a commented Squiggle source, a short description of each assumption and data source, plots of key distributions, and a walkthrough so your team can change inputs and rerun the calculation themselves. I can format the model for Squiggle Hub or embed it, and also add sensitivity analysis — which inputs move the result most.
Describe the decision you're making and any data or expert guesses you have — even rough ranges are fine to start. I'll turn them into a transparent model that you can defend, update, and reuse. Clear structure, honest uncertainty, and no false precision.
You get a working Squiggle model tailored to your problem. I encode your assumptions as distributions (lognormal, normal, uniform, beta, mixtures, or point values where it's honest), link them with real task logic, and run the model so you see medians, confidence intervals, and the probability of outcomes you care about. Typical uses: budgeting and cost-benefit analysis, timeline and effort estimation, revenue and growth forecasts, risk calculation and expected value, comparing options under uncertainty.
The deliverable includes a commented Squiggle source, a short description of each assumption and data source, plots of key distributions, and a walkthrough so your team can change inputs and rerun the calculation themselves. I can format the model for Squiggle Hub or embed it, and also add sensitivity analysis — which inputs move the result most.
Describe the decision you're making and any data or expert guesses you have — even rough ranges are fine to start. I'll turn them into a transparent model that you can defend, update, and reuse. Clear structure, honest uncertainty, and no false precision.
