Per Patteson J in Irving v Manning (1847) 1 HL Cas 287, at 306-307:
The principle laid down in these latter cases is this: that the question of loss, whether total or not, is to be determined just as if there was no policy at all; and the established mode of putting the question, when it is alleged that there has been, what is perhaps improperly called, a constructive total loss of a ship, is to consider the policy altogether out of the question, and to inquire what a prudent uninsured owner would have done in the state in which the vessel was placed by the perils insured against.
Per Vaughan Williams LJ in Angel v Merchants Marine Insurance Co  1 KB 811 at 816:
Precise estimates are, of course, impossible, and it seems to me that, unless the insured shipowner is to take upon himself risks which ought not to be borne by him (such as the risk whether the ship will be got afloat at all, or, having been got afloat, will arrive at a port for temporary repairs, and ultimately at home for permanent repairs), a large margin ought to be added to the figures of cost of repair to cover risks of this sort—risks which a "prudent uninsured owner" would certainly take into consideration in determining whether he should repair or sell.
Per Flaux J in Suez Fortune Investments Ltd & Anor v Talbot Underwriting Ltd & Ors Re: M/V Brillante Virtuoso  EWHC 42 (Comm) at paras 92, 94:
92. It seems to me that the effect of this approach is that, in relation to matters which cannot be determined with precision, such as the extent of damage to items of machinery and equipment which were not opened up and tested, the court has to apply to any repair estimate what Vaughan Williams LJ describes as a "large margin". That is by no means the same thing as giving the assured the benefit of the doubt in a manner which reverses the burden of proof, which is always on the assured to prove that the vessel was a constructive total loss. It is simply recognising that a margin of error has to be applied in relation to the extent of the damage where, as in the present case, it was not possible to investigate fully and the assessment of the cost of repair has to take account of the fact that the items which were not opened up and tested might well have required replacement, so that a prudent uninsured owner would have replaced them.
94. It is, as the learned judge said, always a question of fact dependent upon all the circumstances of the case, where the prudent uninsured owner would have carried out the repairs. Whilst cost is always an important factor, it cannot necessarily be determinative, given the presence of other factors, such as are present in this case, including the need for cleaning before any long tow, the costs, time and risks of a long tow, the reputation of the rival yards, the risk of delay in those yards and the difficulties of repositioning the vessel for gainful employment after repairs have been undertaken. I return to consider all these factors in more detail in the section of the judgment dealing with the place of repair.