Estimating Trends

Here are some choices to the question:

1>By comparing with old analoguos data ?
2>Creating and presenting the basis of estimate (BOE) documentation?
3>Illustrating a track back to cost estimating relationships (CERs) and their referenced historical databases?
4>Demonstrate that the estimate came from a reputable parametric tool like TruePlanning?
5> Any other approach?
6> All of the above or some combination of these approaches?
7> You don't have this problem i.e. you never had a need to prove the credibility of your estimates?

If the answer is point number 7, how do would you rate the above criteria over the others?