Ixas_girl
Posted : 5/18/2007 6:11:45 PM
ORIGINAL: Liesje
So maybe I'm confused or getting side-tracked, but is there really one method of "clicker training"?
The clicker is used for different purposes. Here's a few methods for using it:
you can
lure a reward, then click/treat to reinforce it,
you can
capture a behavior that you didn't ask for, but that you like, by click/treating it
you can
shape behavior, by luring or capturing it in increments and fine tuning it with click/treat markers
Generally verbal cues or hand signals are added
after the desired behavior is reliably given for a number of c/t's. The idea is that the cue is meaningless untill it is attached to something (a behavior) that already has meaning. This is why you don't teach spin left and spin right in the same session,nor do you teach "spin" then "right" and "left", you teach each separately. You get the dog clear on the behavior, attach cue, then proof it.
Clicker is convenient for agility, for example, you can click when dog enters tunnel to mark the behavior, then you can treat when he gets out.
I think the thing you'll notice is that timing is huge. You don't actually click after the behavior, but during it, or even during a specific part of it. When starting with "come", you click when the dog first shows a sign of coming, like looking at you, shifting weight toward you, or actaully approaching you. You want to mark the moment he's enlisted himself to do what you ask, then build upon that. Again, in agility, you wouldn't click after the dog has come out of the tunnel, or when he's halfway through, you click him for entering it.
This also sets a limit to clicker training. Like Brookcove mentioned in one of her recent posts, when a dog is performing a complex set of behaviors simultaneously, the dog can't know which behavior is correct. The click is given to mark a discrete behavior. This is why I don't buy it as a means for solving so-called psychological issues, which are complex sets of behavior (ya, it's JMNSHO).