In an experiment in teaching dance it was found that dancers learned better through practicing a simplified version of a routine than the full routine.
The procedure was this. All three groups were taught a phrase by a choreographer. After learning the phrase during a ten-minute period, each dancer was graded individually on how accurately they performed the phrase. Next, each group practised the phrase for ten minutes, using their assigned method: the three conditions of practicing full out, marking, or mental simulation. At the end of that practice period each dancer was graded again, and we calculated how much each had improved. The size of this improvement showed the benefit of practicing in a certain condition. Each group then changed its practice condition and was taught a new phrase. Accordingly, if group one marked when practising phrase one, they now practised phrase two by dancing it full out, and then later they would practice phrase three by mentally simulating the phrase.
When setting up the experiment we were hoping to find that marking was a more effective method of practice than mentally simulating a phrase while lying on the floor. That was certainly true, we did find that both marking and full out practice were better than mental simulation. But far more interestingly, and to our absolute surprise, marking was better than full out performance by a small but significant amount in most dimensions of assessment.
There is some relation here to Deliberate Practice