How-to guides

Build a sketch

A sketch is a self-contained model with no user inputs — good for one-off explorations.

1. From the analyses list, click + New analysis, select Estimator, then choose Formula (or Spreadsheet if you want to upload a workbook).

2. In the formula textarea, declare your variables — random variables with ~ and derived variables with =. For example:

   revenue ~ norm(100000, 15000)
   costs ~ norm(80000, 10000)
   profit = revenue - costs

3. Click Run simulation. Results appear in the Outputs panel on the right.

4. Optionally configure the output metrics (mean, percentiles, probability thresholds) by clicking the output card.

Create a template

A template is a model with user input fields so the same model can be run against many cases.

1. Build the model as a sketch (see above).
2. In the Inputs panel, click any input card and change its Type to User input.
3. Choose a sub-type:
   • Number — the runner enters a value; set a default.
   • Choice — define options and their values; set a default.
   • Distribution — the runner tunes a distribution; pick a family and param_ui.
4. Optionally set a Test value so you can preview the template from the editor without saving a run.
5. Save the estimator. The analyses list now shows it as a template (folder icon).

Fill in a run as a runner

1. Open the template from the analyses list (click the title, not the chevron).
2. The author view opens. Click Runner in the Template / Runner toggle (top right) to open a new draft run.
3. Fill in the input fields shown in the form.
4. Click Run simulation to see results.
5. Give the run a name by clicking the title area.
6. Click Save run. The URL updates to the saved run's own address.

View and compare runs

1. From the run view, click All runs to open the template detail page.
2. The table shows one row per run with a column for each input and each output metric.
3. Click a column header to sort. Use the filter controls to narrow the list.
4. To download all runs as CSV, click Export CSV in the template detail page header.

Export raw simulation samples

To download the raw iteration samples behind the current preview (one column per variable, one row per iteration):

1. Open the author view.
2. Click Export CSV in the toolbar (beside Share).

This export contains the raw draws the histograms are built from, not the per-run summary metrics.

Share a template

1. Open the estimator in the author view.
2. Click Share in the toolbar.
3. Copy the generated link. Anyone with the link can view the template read-only and clone it into their own account.
4. To revoke access, click Share again and remove the link.

Share a single run

1. Open the run view for the run you want to share.
2. Click Share in the header.
3. Copy the generated link. Anyone with the link can view the run's inputs and outputs read-only.

Use the AI assistant (formula engine)

1. Open the estimator in the author view.
2. Click Ask AI in the toolbar. A slide-over panel opens.
3. Describe what you want to change — add a variable, adjust an output metric, restructure the model. The AI edits the cards and manifests directly.
4. Close the panel with the X button; it stays mounted so an in-flight response survives a close/reopen.

The AI assistant is available in the author view only, not in the run view.

Start from a starter

1. Inside the author view, click New… in the toolbar.
2. Select From starter gallery.
3. Choose a starter model (project cost estimator, monthly revenue forecast, hiring plan cost, conversion funnel, price sensitivity).
4. The editor loads the starter. Modify it to fit your situation.

Create a spreadsheet-engine estimator

1. From the analyses list, click + New analysis, select Estimator, then choose Spreadsheet.
2. Upload your .xlsx workbook or pick a saved data file. The workbook opens in IronCalc — a read-only in-browser viewer. To change formulas, edit the file locally and re-upload.
3. Click a cell in the spreadsheet, then choose Mark as Assumption to make it a sampled distribution, or Mark as Output to record its value after each iteration. Assumption cells highlight blue; output cells highlight green.
4. For each assumption, pick a distribution and enter its parameters in the sidebar panel.
5. To make an assumption a user input (so the runner provides the value rather than sampling it), click the cell and select Mark as User Input.
6. Set the rep count in the sidebar (default 1,000; max 10,000).
7. Click Run simulation. The run is asynchronous — a progress bar shows status while the server processes it. Running requires a Pro plan.

note

Assumptions must be single-cell references. You cannot attach a distribution to a named range.

Cancel a simulation run

While a spreadsheet-engine simulation is running, click Cancel in the progress bar panel. The server stops the job and discards partial results.

Troubleshoot a spreadsheet run

Problem	Likely cause	What to try
Run button disabled	No assumptions or no outputs configured	Add at least one of each before running
Run fails immediately	Rep count out of range	Enter a value between 1 and 10,000
Run never finishes	Circular references or slow formulas	Simplify the workbook; reduce reps to 100 for testing
Results show no data	Output cells are empty or contain errors	Verify the output cells return valid values without simulation
Server timeout	Workbook too large or complex	Reduce workbook size or remove unnecessary sheets

Re-run a drifted run

If a template was changed after a run was saved, the run view shows a drift banner.

1. Click Re-run against latest in the banner to resimulate the run using the current template definition and save the result.
2. Click Keep as-is to dismiss the banner for this session without changing the stored run.