Skip to content
GFunnel
Create A Workspace

The GFunnel Methodology

Forecasting Algorithm

All algorithms

The steps

  1. Identify the variable being forecast. State precisely what is being predicted, by when, with what confidence interval.
  2. Apply the Variable Principle ruthlessly. What is measurable? What are you assuming? Most forecast errors come from filling assumed variables with confident-seeming guesses.
  3. Apply Rhythm. The current cycle phase determines what kind of trajectory is likely.
  4. Apply Layer I.G. What prior similar situations exist? What information do their outcomes contribute?
  5. Per Cause and Effect: identify the causes currently set. Forecasts that ignore set causes are wishful.
  6. Per Pareto: 20% of variables drive 80% of outcomes. Identify the 20% — forecast those, not the 80%.
  7. State the falsifiability conditions. Under what observation would the forecast be wrong? At what threshold?
  8. Document the forecast and its assumptions. Per Database step. Without documentation, forecast accuracy cannot be evaluated and improved.
  9. Per Repetition: forecast accuracy compounds with deliberate review. Each forecast informs the calibration of the next.

Articles on this

No articles on this one yet.

From the GFunnel Methodology v5.1 by Cameron Garlick, licensed CC BY 4.0. Steps are reproduced word for word.