For whatever reason, two-fers and three-fers have become a rote part of clicker training methodology. Okay, I'm getting the behavior I want, now quick let's start doing two-fers and three-fers. They're being added with no understanding of what they are or how they affect training.
Let me ask you -- what are your training goals for your dog? Are you looking for a well-mannered pet? Or are you training for a competitive sport?
Two-fers and three-fers and such are what's called a variable schedule of reinforcement (VSR). This is the slot-machine principle. An animal will continue working because it MIGHT get reinforced. A properly used VSR will create a STRONGER behavior -- the animal will try harder and harder to earn that behavior. It will also cause a VARIABLE behavior.
Let's look at a couple of examples. Pet owners want a dog who will, when asked to sit, plop his butt on the ground. Competition obedience people want a dog who will sit tucked, straight, square, and fast.
The pet owner's dog is offering sits. Sometimes they're straight, sometimes not, sometimes they're square, sometimes they're cockeyed -- it really doesn't matter. All the pet owner wants is for the dog's butt to be on the ground. When the behavior is strongly on cue and the clicker has been replaced with a release word, THAT'S the time to begin a VSR, if you're so inclined. In my technique, I begin varying the type of reinforcer, but I ALWAYS use some sort of reinforcer. If you were to use a VSR, you would give a reinforcer after some repetitions and not after others. Just as I phase in praise-only reinforcement gradually (the closest to VSR that I get), you would phase in a VSR gradually too. Remember, you want to motivate the dog to try harder, not frustrate him into quitting.
Now let's look at how a VSR can affect a precise behavior. Competition obedience people want a dog who will sit tucked, straight, square, and fast. To get that sit, they shape one criteria at a time. I want a tucked sit. I click ONLY tucked sits. If I don't click, that's a signal that the dog has done something wrong and should try something different.
Uh-oh. Bad news for VSRs, eh?
Variable schedules of reinforcement create variable behavior. I train precise behaviors on continuous reinforcement. If the dog does a behavior correctly, he will either be reinforced OR he'll be given another cue. I won't give a dog a chance to perform another cue unless the first behavior is correct.
List and Site Owner: Melissa Alexander, mca @ clickersolutions.com