Referensklassforecasting

Referensklassforecasting, often shortened to reference class forecasting (RCF), is a technique used to improve the accuracy of predictions for projects, investments, or other future outcomes. It is based on the principle that when making a forecast, it is more reliable to start by considering the historical performance of similar, comparable projects (the reference class) rather than relying solely on the specific details of the project being forecasted.

The core idea behind RCF is to mitigate optimism bias and other subjective biases that often lead to overly positive and inaccurate forecasts. By identifying a group of past projects that share key characteristics with the current endeavor, forecasters can derive a statistical distribution of actual outcomes from these past projects. This distribution then serves as a baseline against which the specific forecast for the new project can be compared and adjusted.

The process typically involves several steps. First, a relevant reference class of past projects is identified. This requires careful consideration of project type, scale, complexity, and other significant factors. Second, data on the actual outcomes of these reference projects is collected, such as cost, duration, or revenue. Third, a statistical analysis is performed on this data to understand the range and distribution of past results. Finally, this statistical information is used to inform and adjust the initial forecast for the new project, making it more realistic and grounded in empirical evidence. RCF is particularly valuable in domains where projects are prone to cost overruns or schedule delays.