You already know the answer to this one, but I'll spray a little anyway.

If they are on their own they do what everyone else does, they know something is wrong so they continue to expirement just making slight changes in what they are doing until the point they either find something that works or give up and move on. Climbers are amazingly good intitutive learners and just by expirmenting they often come up with something that works.

As you know, in the coaching environment video analysis allows us to see these problems that occur in extremely shot periods of time. By video taping repeated attempts and then going through the section where the moves fail frame by frame we will be able to understand the structure of the move.

The most basic level of analysis for a single move is as follows:
1- define the base of support then:
2- define the type of balance present at the start position.
3- define the type of balance present as the move begins to fail.
3- analyze timing between hand contact with next hold and position or movement of the pelvis. does pelvis move downward or away from the rock prior to, or after contact is made?
4- how much movement is there?
5- Also draw the new base of support that will be created if the new hand hold is attained and define the type of balance that will be present if the move is successfully completed.
6- also try to determine how the move is being initiated.

I'm sure I've told you about my experience doing video analysis with Dan Mills before, but one session we did is such a clear example for me that I'll mention it again.

I was doing video analysis of him in the gym, he kept falling on this one V10 move. It was an extreme off-set balance move, a long move up and right on an overhanging wall from a sloper to a worse sloper. The video analysis showed that on every attempt his hips were starting to drop away from the wall about 2 video frames (2/30, or .066 of a second) before his hand made contact with the next hold. There was enough downward momentum and the hold was bad enough that he couldn't stick it. As the climber there was no way for him to figure out that this was going on. .066 of a second is just too short of a duration for the athlete to make an assessment of. Nor could an observer using the naked eye. To both it would just appear that the climber was falling as he reached the next hold. Going through the video frame by frame we were able to see this relationship happening, postulate that this was the problem and then suggest ways for him to prevent downward movement of the hips until his right hand was established on the next hold. We did this by repositioning his right foot on its hold and by having him attempt to consciously maintain trunk extention through the move. He was able to successfully complete the move severl times after making these adjustments.