The first issue is one of "factual causation" (ie but for). What is the overall balance sheet loss suffered as a result of the defendant's breach? If the claimants had not sold the vessel, that loss would have been higher. They would not have sold the vessel but for the defendant's repudiation of the charter. That sale reduced their overall loss because if they had not done so when they did, the market for vessels would have been much lower. 

We don't get into questions of "legal" causation without answering that factual question first. The factual loss suffered was not as high as it would have been absent the sale, and the sale would not have happened but for the breach.

Yes, it was a piece of luck that the ship was sold when it was, and yes there were other factors going into the claimants' commercial judgement to do so, and yes they could have sold even without the repudiation of the charter, but so what? They wouldn't have done it absent the breach.

If the defendants breach had caused the claimants to sell the ship, and the market had moved the other way so that their factual loss was higher as a result of the breach than the mere loss of the expected revenue stream, then a question of remoteness or "legal causation", would have arisen.  Although this additional loss would not have been suffered but for the breach, that would not have ended the enquiry. The additional loss would probably then have been seen as legally caused by the owners commercial decision to sell the vessel, and not by the initial breach, even though it would not have been suffered "but for" the breach.

But I don't think the analysis is symmetrical in the way the UKSC thinks, is it? You have to answer the factual causation question first, and here the sale which would not otherwise have happened but for the breach, thereby reducing the overall factual loss that would otherwise have happened. Why allow recovery for overall consequential losses, that as a matter of fact, were never suffered?

Still, I am sure my mind is dulled by examining.