The Runaway is the basis for teaching a dog to leave its handler in order to get a reward. A helper "runs away" with favorite toy or food and entices the dog to leave its handler. Once the dog gets to the helper, the dog is praised and rewarded. What constitutes a reward for the dog must be interpreted by those working with the dog. The reward is most often food, play, praise, or a combination of these. Runaways usually start with a helper familiar to the dog, and are short. As the dog learns this behavior the runaways get longer and unfamiliar people can act as helpers. There are two basic methods for training a dog to search for other people, but both start with a runaway.
A search normally follows this sequence:
- Handler gives search command,
- Dog finds subject,
- Dog returns to handler,
- Dog indicates find to handler,
- Handler gives a "show-me-where" command,
- Dog leads handler to subject.
Dog training requires small steps, and to teach all these steps is a very difficult task. This is where the two methods of training are different. Forward chaining teaches the steps in the order they will actually take place. It requires teaching step one, then having to teach step two and so on. After the first step is learned, the handler must add a new behavior, while at the same time weaning an anticipated reward. Back-chaining is a method that keeps the reward in the same place (with the joining of the subject and searcher). The back-chaining method is taught in a sequence that is the reverse of how the actual search would sequence.
- Have fun with handler, subject, dog and lure, all together; the dog learns that if everyone is together, there is a big reward.
- Have subject run away with lure. Dog wants to follow. Queue dog do a pre-learned alert behavior. This can either be a bark, a bump, a tug on a toy, etc. When dog does alert behavior dog is given refind command and allowed to run to subject and play (just what she wanted to do. She gets her reward.) When this two step sequence is solid with no queuing move to step 3.
- Leave dog with subject. Subject should provide just enough play anticipation that the dog stays with the subject until the handler stops 5-10 seconds away. Subject ceases any activity and ignores dog. Dog wants to play, but subject is unresponsive. Dog then goes to handler, a natural behavior. Handler is unresponsive until dog executes alert behavior, then handler gives refind command. The return to handler and the alert may need to be queued at first, but handler should queue as little as possible. Dog now gets to do what she wanted, run to the subject and reward. The added steps take a little longer, but the dog KNOWS what to expect next, and KNOWS the fun is coming. Reduce the queues to zero before moving to step 4.
- Have subject run away with lure. Dog wants to follow. Same as step 1, so far. When dog gets to subject there is no play - subject is unresponsive, but the dog has seen this before and knows a solution for the problem. Dog goes and alerts handler, sequence is followed thru to reward 4-3-2-1.
- GRADUALLY increase the difficulty of the process. Stop and go back if some part of the process is weakening.
What the dog has learned is that a reward comes only at the end of a set of behaviors.
If the dog is conditioned in this way the dog learns to associate the sequence and the smell of people with its own happiness.