Overlearning is a technique for enhancing memory. Reviewing and practicing what has been learned are examples of overlearning. Overlearning also has a more specific meaning when the number of attempts to learn something can be measured. Once an attempt succeeds, more attempts are made in proportion to the number of attempts required for the initial success. Normally the number of overlearning attempts is one half the number required for success.
Overlearning may be applied to solving go problems. For instance, if it takes five tries to learn to solve a problem, do it three more times.
Hyperpapeterie: A natural thought is also that the further along you are in the process, the more the trials should be spaced out. So if you wait a week between the first and second attempts, you should wait a substantially longer time before the third. Think of a course in school (math or language study, perhaps): you are taught a technique, practice it for a bit, and then shortly thereafter have a quiz. You don't actively study the technique anymore until you review it for your final, though you may use it or have to think about it from time to time. Then at the beginning of a later course, you briefly review the technique. By this point, if things have gone correctly, the technique is yours for life: you might get rusty, but you'll never again have to do more than a short review to have it available. This matches my experience in repeating Davies' Tesuji, one of the few books I've had available to me in the past (sadly, I can't claim the last step of effortless proficiency).
Also a question to satisfy my curiousity: Bill, I find the idea of spending half as much time on the followup as it took to initially get the problem rather appealing. Is that guideline based on research or just personal experience?
Bill: Research. It was what I was taught in Psych101.
- Force feeding, which discusses much the same technique.