AI Blocks › Track your numbers
Starting assumptions checker
Tell it what you already know about your numbers. Get starting assumptions in plain units, tested to see if they make sense.
Starting assumptions checker is a free AI skill for the job "Track your numbers". Give the instructions to an AI assistant that supports the required files and tools, then review its output.
Free · No account · MIT licence · Made by PyMC Labs · Level: expert · Thinking: high
Made by PyMC Labs. See the original on GitHub; we copied it as it is, with its licence.
What it does
Starting assumptions for a statistical model, stated in real units and tested to see if they make sense. A prior is a scientific assumption, not a default setting, so the skill begins with what the quantity means, its units and its limits, and asks for your quantiles and probabilities rather than inventing them. It turns your answers into a distribution using formulas or PreliZ, then checks the result met your stated numbers. It simulates what the prior predicts for the data in PyMC and compares that to what you consider plausible. Finally it varies the scale and tails to see how much the choice matters. It needs Python.
You give it
- The quantity you are modelling and its units
- Your best judgement of likely ranges
- Your model setup in PyMC (optional)
- Python with PyMC and PreliZ
You get
- Priors written in plain units
- Checks that they match your judgement
- Simulated predictions from the prior
- A sensitivity check on scale and tails
Use it when
- You are building a Bayesian model and not sure what prior to use.
- Default priors give odd predictions.
- You need to explain and defend your priors to others.
- You want to see how sensitive results are to your assumptions.
How to use it
- Add it to your AI once: download it and add the folder to Claude, ChatGPT or Codex as a skill.
- Ask for the job in your own words. Your AI opens the skill when it fits.
- It works best when your AI can run Python code.