Pyg's Recall

Look! I have trained him!

Shedd Animal Training Seminar: Other Advanced Concepts

Picture is unrelated. But awesome.

There are four more advanced concepts that Ken discussed. Each is interesting, but there’s not quite enough material to flesh out separate posts. So… here’s the quick run-down.

Recall

It surprised me that Ken put a recall (which he defined as a behavior that requests an animal to return to “station”) as an advanced concept. After all, every puppy or beginning class I’ve seen teaches “come!” I think the reason Ken classifies it this way is twofold. First, he considers it a safety behavior, either for the trainer or the animal. And second, he outlined a number of errors that novices often make when teaching a recall.

Really, the errors people make with the recall all boils down to one thing: reinforcement. Ken said that a recall should be reinforced often and well. He recommended practicing it daily with your animal. If you use a recall in an emergency situation, you should always reinforce the behavior, preferably with a high-value item. And finally, he noted that novice trainers often use the recall only when the animal has done something wrong, or when good things are going to end (like when you call your dog to come at the dog park and then leave).

Behavior Chains

A behavior chain happens when the completion of one behavior cues the start of the next. Each subsequent behavior reinforces the previous one. Ken discussed two types of behavior chains, technical chains, where the trainer gives ONE cue and the animal then performs a series of behaviors, or common chains, where the trainer gives a series of cues, but there’s only one reinforcer at the very end.

There are two ways to teach a behavior chain. With back-chaining (which is the preferred method) the last step of the chain is taught first, and then the trainer teaches backwards to the first step. This is more reinforcing for the animal because he is always moving towards something he knows well. Forward-chaining (teaching the first behavior, then the second, and so on) also works, but there are usually more errors.

It is best to teach each individual component of the chain successfully. Each should be maintained separately as well. Ken shared that the animals that are the most successful at learning chains are ones who have learned about reinforcement variety because they already know that a behavior can be reinforcing.

Keep Going Signals (KGS)

A KGS is a signal that tells the animal that he is on the right track. It’s encouraging feedback that literally says “keep doing this and I’ll reinforce you!” It’s technically a tertiary reinforcer- that is, it’s a signal that a secondary reinforcer is coming (and thus, a primary). It’s sort of the way you get a poker chip that can be traded in for money which can then be used to by an actual reinforcer, like food.

Most people don’t purposely train them; they tend to happen along the way naturally. That said, you can purposely teach them by introducing them as a secondary reinforcer and then approximating longer periods of time before giving the primary. Interestingly, Ken said he doesn’t use them; he just doesn’t need them.

End of Session Signals

An end of session signal tells the animal that the training session is done and there will be no further opportunity for reinforcement. While it might be helpful to let your pet dog to stop bothering you, Ken pointed out that he doesn’t particularly want a 700 pound animal realizing the (fun!) session is over and thus refusing to let the trainer leave. That can be dangerous. In fact, the staff at Shedd are pretty careful that they don’t create an accidental signal that might lead to chaos, like picking up a bucket of fish and walking away.

That doesn’t mean you should use them. Ken simply doesn’t think they are important enough to argue about. If what you’re doing is working, great.

Training Tuesday: The Joy of Training

Maisy and I continue to work hard. Well, "work" is probably the wrong word. Maisy actually stands by the front door (we only use that door to go to the park across the street from our house) and whines if I haven't trained with her yet that day; she loves the heeling game we play. For my part, I feel exhilarated after training sessions. Although I expend a lot of energy during our training sessions, they are so much fun that it doesn't seem like a chore.

Of course, it certainly helps that I see measurable progress every time. Part of that is because I've taken Denise Fenzi's advice to challenge my dog: I have raised criteria fast and furious since I've begun working on Maisy's heeling. Okay, it wasn't really me raising the criteria- Denise has coached me through each step. I'll admit, I was initially skeptical with how far and how high she wanted me to raise criteria. I really didn't think we could do it. Maisy has proved me wrong, though. This dog is capable of amazing things!

Take a look at this video. I've posted it before, less than two months ago, in fact. Then watch this:



Amazing, no? As the video says, my task was to vary my pace and direction frequently. I was striving for some kind of change roughly every three seconds. At the same time, I drastically reduced the frequency of ball-reward, and used verbal praise in its place. This was to teach Maisy that my changing speed and direction is the reason to pay attention, not just the ball. It totally worked, too. She went for almost 90 seconds without a ball throw, and she doesn't show any dampening of enthusiasm!

I've also started adding in some non-toy play, which you can see it at the beginning of the video. This wasn't something Denise suggested (although I don't think she disapproved), but I wanted to start building some reinforcers that were based on me, not an external object. Right now, this really only works before we start training- she's so amped up then that she'll play back. If I try to do it during a session, she will sometimes play back, but usually she licks her lips, or shows me some other stress signal. We'll keep working on it.

Another area where I've been raising criteria like crazy is with Maisy's dumbbell retrieve. As you may remember, we were stuck at the stage where Maisy would pick it up and hand it to me. Click here to see our last retrieve video.

A couple weeks ago, feeling empowered by our success with heeling, I decided to get out the dumbbell again. I started easy, asking Maisy to just pick it up off the floor. Then I moved it about a foot away. Then another. I kept tossing it further and further until she was retrieving it from six to ten feet away. The video below was taken at my aunt and uncle's cabin- a new environment- and Maisy absolutely rocked it.



But I don't think that simply asking more from Maisy was the key to our success here. In watching the two videos, the part that struck me most was how much I was talking to her, encouraging her, and praising her. In the first video, I was doing "proper" clicker training- being silent and letting the dog think. That works for many things, but it didn't work here. Once I started helping her with my voice, though, things just took off.

Yes, I broke the rules and repeated my cue. So what. She wasn't going to respond to the cue anyway (she was too distracted by those treats on the counter), and by gently reminding her what I wanted, she was able to be successful. I praised when she got close, and celebrated when she grabbed it. In the video, you can see her whole demeanor change when I do. You can also see that in each successive retrieve, she trots out to the dumbbell more confidently. She understands what she's supposed to do now, thanks to a little verbal support. 

Finally, I'm going to leave you with one last video. If you follow us on Facebook, you've already seen it, so I'll just link to it. Click here if you want to see what happens when the training session is over and we're just screwing around. That's right- she can do a drop on recall! What an awesome dog! Training is truly a time of joy for us both. 

Training Tuesday: Three Things

On Retrieves and Jackpots
Maisy still doesn’t have a formal obedience retrieve, and while we may never get the opportunity to use one, I still want to teach it. It’s been a good exercise for me- it’s really helped me pay attention to my timing, criteria, and rate of reinforcement, but that’s all beside the point. I’ve been using Shirley Chong’s method to shape a retrieve, and overall, it’s been going well, but I’ve struggled to add duration to the hold, so that’s one thing we’ve worked on lately.

Now, when I work with Maisy- on any task- I often toss treats on the ground away from me to help reset the exercise. I’ve also mentioned before that I use jackpots while shaping, and my typical method is to click, toss a treat, and then verbally tell her how smart she is as I continue to toss treats, one at a time, on the floor for about ten seconds. Then we return to the exercise.

Last week, while working on holding objects for longer periods of time, Maisy did a lovely four second hold. I clicked and tossed a treat, which Maisy found, then looked at me, eager to start the next rep. I belatedly realized that, hey, that was pretty good, and murmured, “Nice!” The second I said that, Maisy immediately began searching the floor for more treats.

Apparently, I have a jackpot marker.

The Come! Go! Game
I’ve been wanting to write about the Come! Go! Game for a long time, but it’s better with video. It’s also been hard to get video of it, but today, I finally have some! It’s not the best example, but here it is:



The Come! Go! Game is basically the Premack Principle at work. Premack says that you can reinforce a low-probability behavior (coming) with a high-probability behavior (running away). Interestingly, as you do this, the low-probability behavior gains value, and the high-probability behavior loses value.

You can totally see this happening in this video. At first, Maisy runs far away, and quickly, when I tell her “Go!” But, as the game goes on, she not only quits running as far, she also takes several cues to take off running. This happens every single time we play the game. The first few reps are enthusiastic, and then she decides it’s more fun to stay near me, which in the end, is exactly what I want anyway.

However, it does mean that if I ever train an obedience go-out, I can’t use “go!” as a cue.

The Stuff I’m Supposed to be Working On
Last week, at our re-check with the vet behaviorist, we agreed that I’d start working on counter-conditioning Maisy to everyday noises around the house. I have failed miserably at this. It seems like I never have treats handy when I need them, and at the end of the day, I’m too tired to get up and grab some.

I’m posting this publicly in an effort to embarrass myself into doing it. I've already put a glass jar with treats (glass so neither canine nor feline can chew it open) and put it in the living room where I spend most of my time. I've also stashed a clicker there, because even though a clicker isn’t the best tool for counter-conditioning, there are times where it can be helpful.

Next Training Tuesday, I want you all to shame me if I don’t report progress on this, okay?

The joy of a well-trained dog

Last weekend, my husband and I took Maisy to a local state park. It was a beautiful day: warm, but not too hot, mostly sunny, and perhaps best of all, while the clouds were ominous, it didn't actually rain. I call this the best part because it meant we were virtually the only people there, and as such, there was very little risk to letting Maisy off leash.

Maisy had a great time sniffing new scents, investigating critter holes and fallen logs and such, and just generally getting a chance to be a dog, but despite the awesomeness of being in a new environment, she was really good about staying close. She never got out of eyesight, and rarely went more than 20 to 30 feet before she would stop and look at us. Most of the time, she'd either wait for us to catch up, or wait for me to acknowledge her before running off again.



I also used the opportunity to practice recalls. I purposely chose times she was focused on something else, and would call her. I was thrilled to see a "whiplash" response where she would turn the second she heard her name. When she arrived, I'd give her a treat, and then send her out again, thus using both positive reinforcement and Premack.



All the recalls paid off, too, when I saw a cluster of people heading toward us on the path. Maisy saw them too, and she did hesitate before responding, but she came. I clipped her leash on, asked her to sit in heel position, and waited for the people to pass. As they did, they commented on what a good dog Maisy is. She is, of course, but I tend to take it for granted these days.

Anyway, it was a really fun afternoon. I bought a year-long pass, so I see many more hikes in our future!