***** & Buddy
|Our longer-term goals:||Wait, Drop It, Leave It, Heel and loose-leash walking. Anything else?|
|Our immediate goal:||to achieve 3 trained behaviors at the 80% level of success. We’ll start with Look, Come and Leave It.|
|Recommendation:||We agreed to meet again next Monday at 2:30.|
|Resources:||Check out the resources listed on our webpage for books and dvds that will support our work together. http://www.thedoghousenyc.com/training-resources/|
On our first visit, I introduced you to the use of a marker (clicker) to mark and reward behavior (click and treat, or C/T). We talked about how long it generally takes to train a behavior and the difference between teaching a behavior (generally a few minutes or, at the beginning, a little longer) and bringing it to a wholly reliable level (several months).
I told you that the first three or so behaviors take the longest to teach because both you and Buddy are new to the game, so to speak. But after that, the training tempo should pick up quickly.
I also suggested that going forward you should ignore any unwanted behavior completely (on the principle that even negative attention is attention and therefore reinforcing) and amply C/T any behavior that you liked.
Using the behavior Look (his eyes to your eyes), I took you through the stages of training any behavior: catch it, wait until he offers it deliberately, give it a name, proof it. I felt Buddy “got it” with me and I moved on to name the behavior Look. But you said it was a little difficult for you to tell if he was really looking in your eyes. (So maybe we rushed it to give this behavior a name because we were working within a training hour but, if so, we’ll fix it next time.) We chose to start with Look because giving attention is the foundation of other behaviors.
We also talked about clicker mechanics, ie. the timing of clicking and the timing of treating. Next time, we’ll talk about the placement of your treat hand.
I mentioned you should only ask once for a behavior. If Buddy doesn’t understand or doesn’t deliver, simply “reset” yourself by moving an inch or two and asking again. (This only applies if the behavior already has a name.)
I also mentioned to you that, if possible, it’s a good idea to name a behavior (“put it on cue”) as soon as possible. Otherwise, Buddy will deliver the behavior when we’re not thinking about it and will not receive any reinforcement. This will confuse him.
But of course we cannot name the behavior (whatever behavior) until the light bulb goes off for him and he starts producing the behavior repeatedly. We are looking for him to give the behavior 80% of the time. A good way to measure is 8 times out of ten.
We then moved on to training Come. We sat opposite each other at a distance of 4 feet and waited for Buddy to move toward either one of us. At first he was confused and stood there barking and wagging his tail. We lowered our criteria – we were setting him up to succeed. We did this in two ways: first, by C/Ting a mere look in the direction of the person we wanted him to come to and, second, by moving closer to each other so he had less far to travel. We did not get to the point of giving this behavior a name (“putting it on cue”), but we did both notice that toward the end Buddy seems to be moving back and forth between us. We decided to quit on a positive note.
Look: (a) I asked you to see if you could distinguish the exact moment Buddy’s eyes meet yours.
(b) Practice your timing of the click so that you clicked the instant Buddy met the criterion.
(c) Practice the time of the treat delivery, so that the treat came within a second or two of the click
Come: We agreed you would work by yourself, rewarding Buddy for coming to you when you sat on the floor in the manner that we did.
Basic training arc for any behavior:
1. Catch him doing it and click as if you’re shooting a picture (timing is critical)
2. When you see the light bulb (he offers the behavior deliberately), wait until he is delivering at an 80% rate (say, 4 times out of 5 or 8 times out of 10)
3. When the rate is right, give the behavior a name (“put it on cue”).
4. At this point, you will stop rewarding the behavior if you have not asked for it.
5. Build this stage up to an 80% success rate. ie, he is offering you the behavior when you ask for it 8 x out of 10.
Generalization: This refers to a dog’s need to perform a behavior in many different contexts for him to understand it fully.
Some examples: sit in the bathroom or in the living room; sit inside or outside; sit from a down position or from a standing position; sit at your side or in front of you; sit when you’re sitting or sit when you’re standing; sit when you’re wearing sunglasses or a funny hat; sit when someone besides you asks him to sit.
Fluency: This refers to several (six) different dimensions of any behavior. We talked only about two: Distance and Distractions. Any behavior is more difficult to perform the farther away it is and the more distractions there are. If you are setting your dog up for success, you will train him close in and in a quiet environment to the point where he succeeds 80% of the time before raising the distance or the distractions in small increments. We did not discuss this, but you should work with one dimension at a time, either distance or distraction, for example. (The other dimensions for another time include: duration, precision, speed and latency.)
Setting up for success: This refers to helping your dog do well at every training level. If he is not performing, the chances are good that (a) he does not understand what you want or (b) the fluency requirements are too high for him.
Ending on a positive note: Resist the temptation to keep drilling. When your dog performs well on the last attempt or in the last session, quit! C/T and offer to play with him.
Definition of a session: Generally 10-70 treats long. Maybe 2-10 minutes. Short is good!
The Click: The click does two things: (a) communicates information and (b) promises something good is coming.
Clicker mechanics: The click happens at the precise moment the behavior happens. The reinforcer happens a second or two later, not more. But not less either, lest the dog fail to perceive the lesson communicated by the click (“that’s right!”) and focus his attention on the treat. So separate the click from the treat, but by no more than 1 or 2 seconds.
Also, find a parking spot for your treat hand. Your hand should always return to that spot and it should be away from your treat bag so that your dog’s attention will not be on the food.
Capturing a behavior: This is the process of C/Ting when the dog does what you have in mind. He does not know you have this in mind but that doesn’t matter. He will come to perceive that whenever he does that thing, he gets a C/T. In a short time, he will be doing that thing again and again, on purpose.
Capturing also happens when you are shaping a longer behavior. In this case, we could say you are capturing approximations. And example was when we were teaching Buddy to come, and we C/Ted a turn of his head in the desired direction. It was not Come, but it was an increment (or approximation) on the path to Come. And we captured it.
Putting “on cue”. This refers to the point when you give the behavior a name. First, you capture it, then you put it on cue.
Reinforce things you like, ignore things you don’t like.