In short, between phases of a battle the player (and enemy) would be given the opportunity to "merge" the CR of surviving ships. Handwave is that crews, munitions and supplies are re-shuffled to a few ships to keep them viable, effectively mothballing those ships that the CR is sourced from. The efficiency of these operations would be influenced by skills, officers, the presence of logistics ships and spare crew available. This would be less efficient overall during the transition between battle and disengagement attempts, as opposed to fighting another battle.
Calculating it to me seems like it'd be a pain - of course 5% of CR from a frigate would not equate to 5% CR on a cruiser. There would be a cap to how much CR can be restored too, again influenced by hull sizes, crew, skills, logistics etc.
The aim is to make CR a little more elastic, because it's often the deciding factor in whether you win or lose; Who has the most ships that outlast the oppositon. If you burn through the majority of your fleet's CR without decisively winning or at least being able to cleanly disengage, you've already lost, likely to the extent of being wiped. This would allow the player to beef back up a small handful of their ships, crash-mothballing the rest to potentially swing the battle, or at the very least restore some ships back to the point they're able to provide some cover a disengagement.
I was also thinking about the ability to restore a portion of CR to ships at an exponentially higher cost than it would take "naturally", again up to a cap. I've always looked at the retreat scenarios as the pursuing fleet catching up to the fleeing fleet after some time, some of which could reasonably be spent by both sides in patching up what little they can.