This blog is one of those good news-bad news - situations because some of what I did was appropriate, and it worked, but the training methodology was correction-based because, at that time, I knew of no other way to train. This video is VERY difficult for me to watch because of the stress I place on Blush. Just after this video was recorded, I began my journey into clicker training.
Never again will I train a dog in the way you see me do it in this video.
If I had used positive reinforcement training with Blush here, she would have been happy to go back to her bowl. I probably would have needed to redirect her several times, as I did here, but not as often because she would have understood what I wanted, not what I didn't want.
Before I talk about what actually happened in this video, I want to share with you what I should have done instead.
In this video, I use a verbal, distinct sound to interrupt Blush's behavior. I happen to say "nope" most of the time, or "no," said distinctly. I had never told Blush nope, or no before, so I could have said "plink," and it would have had the same effect on her. The interruption is fine, but what I do next, every time, isn't.
My distinct words "marked" her behavior; it identified her specific behavior that I didn't like. She stopped; then I told her she was a good girl - so??? Why was she a good girl? What was she supposed to do next? How was she supposed to know that?
I had never done this kind of training (redirecting) with her because she always stayed with her own bowl. I lost track of this in my mind because there are a lot of dogs in training here, I was thinking about what I was trying to talk about in the video, and it seemed obvious to me that she should just go to her bowl and eat.
Well, as the saying goes, "You train the dog in front of you," meaning that you make training decisions in the moment as it relates to the dog in front of you. This dog was becoming increasingly frightened. I noticed that and tried to talk her out of feeling that way by comforting her. It didn't work because I still had not clarified for her what I wanted her to do. She only knew she was making mistakes; dogs don't like to make mistakes any more than humans do. It is entirely possible that this was the first conflict I had ever had with this young dog. Again, poor Blush, and I hate to watch this video.
Shaping!
Shaping is what I should have used to train this behavior. I had no idea what shaping was at the time this video was recorded, but I do now, so let me explain the concept before we go forward.
Most complex behaviors dogs are taught are pieces of behaviors all added together to complete the finished behavior. In shaping, the dog is marked and reinforced for the first pieces of the finished behavior. Over time, the criteria to click for is raised as the dog learns to link the pieces together. If the dog is having trouble giving the trainer the full behavior, often clicking for a beginning piece of the finished behavior will give a clue to the dog of what is wanted, and the dogs will then be successful.
But I didn't know any of this:
Dog walking away from bowl = wrong.
Dog eating from bowl = right.
This is what I should have done:
After I interrupted Blush, I should have marked and then reinforced any tiny piece of her behavior that was part of what I wanted her to do. Interrupt, eye contact, mark, and reinforce. A glance toward her bowl, mark and reinforce. A step toward her bowl, mark, and reinforce. If I had taken a step toward her bowl, she would have followed me; mark and reinforce. If I had stood near her bowl and she stayed with me even a second, mark and reinforce. If she had put her nose in her bowl, mark and reinforce. If she didn't put her nose in her bowl, if I had placed my hand near her bowl, her nose probably would have followed my hand, mark and reinforce.
In watching this video, I can see what feels like a million "shaping points" that Blush offered me. In trainer's terminology, a shaping point is a piece of a behavior that is the beginning of heading in the right direction toward the finished behavior I am trying to teach the dog; it is something I can mark and pay the dog for.
Over and over and over, Blush offered me behaviors that I could have paid her for. She might not have understood why I was doing that in the beginning, but she would have been happy and engaged with me. In my experience, it usually takes a dog less than a minute to recognize that the click (or mark) will always be followed by something they like. Once they understand this, they try to figure out what it is going to take to make a human pay them. It is so much fun for the dogs and the humans. I use a verbal marker only with pups, or dogs, that have learned the clicker game well. To teach a new behavior, the use of a clicker will be much more clear to a dog than a verbal marker.
One of the few things I did well in this video was to use my own body repeatedly and purposefully to block Blush's advances toward Yeti. In doing so, I also "broke the sight line," meaning that she could not give Yeti the evil eye while he tried to enjoy his dinner. I have done this training with Yeti many times, always protecting him from what Blush is trying to do. Me moving around is of no concern to him. In fact, because of the shaping I had done with him previously, though I didn't know what I was doing had a name, I am able to verbally tell him to go back to his bowl, that I've got this, and he believes me.
Notice that in the very first part of the video, Yeti is calmly eating. Then he starts to shift around a little. When he raised his head and looked past me, I realized what had changed, and that Blush was headed his way. Somewhere in the middle of things, when Yeti stays with his bowl and eats, I verbally reinforce him for doing something right. I knew to do that before my KPA training, but I didn't fully understand how to create a more developed behavior by marking and reinforcing pieces of the behavior, which is shaping.
In this video, I did a beautiful job of supporting Yeti and really didn't support Blush at all; please view the blog below for more information about that. In a blog to follow, you'll see this pair of dogs again at a point in my education after I had learned the principles of shaping and had become good at it.
I use shaping with my Maremmas more than any other training tool in my toolbox, in part because I can use shaping even when I am a great distance away from my dogs and when they are with other dogs. I have a short video that describes what I am doing at that moment while shaping eleven dogs to eat their dinner as I sit in my Kubota, parked in the lane and not near any of the dogs. It is an incredible tool!
Appeasement Behaviors – Canine Conflict Resolution:
This is a blog I wrote to show the difference for the dogs in switching from correction-based training (what the dog did wrong) to positive reinforcement training (what the dog did right). Notice that the first Rosie experienced a lot of frustration and confusion as she learned this lesson. The second pup, also named Rosie, was taught the same behavior through the use of positive reinforcement training. She loved every minute of our training games, and learned the next behavior quickly and easily.
Training Rosie – Then & Now – Shaping
Comments