Price the pending work with the pessimism it deserves

Variations not yet agreed, claims not yet accepted and provisional sums not yet expended are the parts of the forecast most tempting to flatter, because nobody can immediately prove you wrong. Price them the way you would defend them: at the number you genuinely expect to recover, not the one you first asked for, with the optimistic and pessimistic ends both written down. A forecast built on best-case assumptions is not a forecast, it is a hope.