To assess how the risk profile of a project evolves from Main Gate to ISD, the entire set of over 10,000 periodic ISD forecasts in CMIS were considered. To avoid changes in the mix of projects in each data point demanded a consistent sample of projects. Projects without three-point forecasts across each period up to ISD were excluded. In order to maximise the sample as far as reasonable, and therefore ensure it is as representative of all projects as possible, no forecasts were required in the first 40% of a project's period from Main Gate to ISD. (This would appear reasonable as forecasts would appear to change little during the early stages of a project post-Main gate.)
This necessarily restricts the sample to those projects now in-service that have three-point forecasts of ISD in CMIS for the last 60% of the period from Main Gate to ISD. Each data point therefore represents the average of forecasts for the same set of 25 projects of ISD slippage made closest to that 10% interval.