"The accuracy of historical-data Monte Carlo software forecasts improved when at least 20 historical data points were available."
The study found better forecasting accuracy with 20 or more historical observations, supporting the use of workflow forecasting only once enough actual project history exists.
View Source