@Jint3x, you have a good analysis question here. The short answer in my opinion is that retrieving simple images is less effort to reconstruct than more complex images. General images are simpler.
Let’s go through the complexity of the mnemonic techniques. The first type of association is between a cue and a simple target, often called the link method. One image brings to mind one other image. The target is always one thing with possible enhancements.
The next association adds detail to the target in order to remember more. It’s the cue and compound target. I add keywords and action between the subject and items in the “scene” but it still can be remembered as a mnemonic image. The parts of the compound image are like a sentence including a subject, enhancements, action, item, and a terrain (SEA-IT).
If those images don’t retain enough detail, then you start associating the targets in a story or hook them to a peg system. But of course, the more detail, the more review necessary.
Now about the reverse engineering of the targets, mnemonic images are not one-way associations because you see them as a whole ijmage, sometimes losing a piece of the picture but still having enough detail to retrieve a cue or other targets. Many of my review sessions involve trying to figure out why a particular subject is in the picture or what happened in the story in a previous scene (the A to the B, C, and D I was able to remember) because I couldn’t make my way forwards. I often times do because of the more reviewed and stronger imagery in another part of my story.
The A, or the cue, is likely to be the most obvious thing to retrieve but not always. It depends on how solid the cue is visually and my recommendation is to make the cue as visual as possible adding details that help create a compound image worth remembering adding action wherever possible.
You have a good sense of what might happen in an association and you should practice to see how it all works out for you. Your experience will always give you the best answer.
Doug