2.32 PITO's analysis of the value for money of O2's offer used the outputs from the should-cost model to test the reasonableness of O2's cost base and rate of return, as well as comparing O2's offer with the public sector comparator. Although the comparator was subject to a sensitivity analysis, for example by adjusting the number of base stations, the principal output from the comparator was a single number. In this part of the value for money analysis, the basis of Charterhouse's advice that the deal on offer was value for money was that the comparator, estimating the cost of a conventionally procured national network at £1,610 million, was more expensive than the Airwave deal at £1,470 million. Because of the inherent uncertainty of forecasting the future, in this case looking forward 19 years, the reliability of a single figure output from a computer is low, and so a favourable comparison between a PFI deal and such an output does not on its own prove value for money. This weakness was offset by PITO's development and use of the should-cost model to challenge O2's costs.