Here's an entire video of my dog making errors finding front, mostly due to loss of attention. I am primarily demonstrating the technique of making the dog "more wrong".     The basic sequence is that he returns to me, drops his toy for another repetition of the behavior (front holding an object) and...then he gets distracted watching his toy roll away. Intead of waiting for him to re-engage, I begin to pivot/move in a way that he will notice when he does return to work - I'm not where he left me! By making the work more difficult (pivoting towards him or backing up makes it easier), I'm re-engaging his brain. "Oh, I need to pay attention and solve this puzzle!". Looking at this, I should have removed the retrieve portion from this exercise, since he was struggling with a decent percentage of his returns - the dumbbell just created delay.