On Continuous Reinforcement
Yes, we discussed this very subject one day. The competition people -- and there were lots in the class -- asked the same question.
Bob [Bailey] asked how long an obedience competition performance is. Then he asked how many times the animal had to do it.
Then he laughed.
He said if you use continuous reinforcement and the behavior breaks down in that period of time, then your behaviors aren't strong enough. I believe I mentioned in an earlier post that Bob hadn't seen but a few competitions, but that of those he had seen, he felt none of the dogs exhibited strong enough behaviors to be performing.
Bob said a video camera is your best friend. He said that you should video tape your dog running through several practice runs. Every part of every behavior should be a carbon copy on each session. Literally. He said if you're getting any variability in the behavior at all, the behavior isn't strong enough.
He also repeatedly added the disclaimer that he doesn't train dogs for obedience, has never trained dogs for obedience, and knows nothing about the sport. His opinions are based merely on his other training experiences.
My own analogy is to think about performing at the highest level of canine freestyle. Every movement -- yours and your dog's -- must be strictly choreographed, a literal dance. That's what your obedience routine should be like.
Realize that the dog isn't without all feedback in the ring -- you can still use praise. You just can't use food for four minutes. When you train with continuous reinforcement, you're continuously reinforcing all *correct* behaviors -- not all behaviors. If you don't reinforce then, the dog must logically conclude that something wasn't right about the behavior, and will vary it next time. In a chain, the reward is either a click and treat or the opportunity to do the next behavior. Never, ever allow the chain to continue if one response is below par.
Variable reinforcement builds strong behaviors. Bob has no problems with beginners using it because they don't tend to train strong enough behaviors. Therefore, without a variable schedule, the behaviors would be too weak and perhaps extinguish, and the person would get frustrated and quit. But competition trainers aren't beginners. They train (or should train) strong behaviors at each step before graduating to the next step.
So continuous (differential) reinforcement enables the competition trainer to build strong, excruciatingly precise behaviors. A variable schedule builds strong behavior, but it also add VARIABILTIY, which is the bane of a competition trainer.
Bob likes continuous. Continuous is simple and continuous works. Why make things unnecessarily complex?
Copyright of all posts is the property of the original author. Please obtain permission from the original author before copying, quoting, or forwarding.
List and Site Owner: Melissa Alexander, mca @ clickersolutions.com