How and when to fade primary reinforcers such as food and toys is one of the hardest things for people and dogs in performance sports. There are a number of levels to this. I will focus primarily on agility training. There are additional nuances for obedience.
The first thing is to have strong cues, criteria and behaviors in a variety of settings. This means that the behavior at the bottom of the contact, such as a two feet on the board and two feet on the ground (2 on-2 off) needs to be able to be performed when cued and done quickly and accurately no matter where the contact equipment is located and no matter where the handler is (ahead, behind or laterally away from the dog). When this is not consistent then it is still training mode. When this happens 90 percent of the time correctly and quickly then the first step is complete.
The second step is to move into what I will call a maintenance mode. This is when I will go from always rewarding a behavior to randomly rewarding the behavior with a primary reinforcer such as food or toy. So if I am doing a sequences with all 3 contacts in it and my criteria is being met 90 percent of the time I will pick one of the three contacts to reinforce.
The third step is happening while I am doing steps one and two. Away from agility I am working on teaching the dog the reward may be away from me or may be delayed in arriving. This starts at home with the dog by the kitchen counter and treats are on the counter. Dog is a bit away from the counter. Ask for an easy behavior, mark with your delayed reward marker and then go to the counter and get a treat and deliver it to your dog. Gradually increase the distance (and hence the time) for the reward. You can decide if you want the dog to learn to wait for you to come back with the reward or if you want the dog to come with you to the counter for the reward.
The fourth step is to teach an end of run routine and pair putting the leash on with getting a reward outside of the ring. Teach a word or phrase that will signal to the dog that reward is coming soon. This will become part of your end of run routine at a trial but you first want to teach it in training.
The fifth step is to bring these various things together. On Saturday Feb. 23rd at 5pm at Agile Canines I will be having a lecture/discussion session on this topic. I will go into more detail about the various steps and the variations you can make in this transition. Check out the "Events" page for more information.