The question then is, what is the optimal latency to recognise improvement or decline? Could deferring judgement help with combating this observation? Then als

The question then is, what is the optimal latency to recognise improvement or decline? Could deferring judgement help with combating this observation? Then also what magnitude of praise or negative feedback elicits best results? Essentially this could down to finding out the optimal training schedule for humans. (Like for neural nets or classifiers.) And this is similar to adversarial training. You have to consider both the trainee and the trainer (discriminator).