Delayed Reinforcement

Beanwood

Administrator
I think it was a comment from @Joy that brought this topic to my mind. This is something I am exploring right now with Bramble, not saying I am doing it right, think I need a bit more work in this direction. An example is when we are working on heelwork, I don't really want to be streaming treats, as it seems to encourage an unnatural walk, although initially, it was important to use very high reinforcement to reward the right position by my side. Coming back to delayed reinforcement and how I use this for heelwork. The advantage with Bramble is that she knows and trusts the fact that BIG rewards are coming! :) The excerpt below discusses confidence, and that a confident dog is a patient dog, this goes a long way in considering using delayed reinforcement. I use a lot of positive words and happy upbeat body language to let Bramble know that she is doing a super job.

In this video. Yes again I have made a couple of mistakes with my cues (actually I am pretty awful!!!!) but you get the idea! Bramble is happy to work, and I am asking for a lot from her. She is not quite sure in the beginning, you can see her hesitate when I ask for a heel at some distance. This is unusual, and she recognises the fact. I use verbal praise and upbeat body language. She relaxes, gives me her happy face and enjoys the game! We have a quick party at the end, high fiving which she loves as well as hunting for treats! :)




Excerpt from the podcast -
"Empowerment and Choosing Delayed Reinforcement"

A full transcript is available HERE


Melissa Breau: That plays into your approach when it comes to teaching dogs the concept of delayed reinforcement too, right? And for those who might not be familiar with the term, can you briefly explain what the term "delayed reinforcement" means?

Julie Daniels: Oh sure. That was a very clever segue, Melissa. Well done. The empowered dog — I'm going to expand upon your segue because I think it really works here — the empowered dog is self-assured enough to wait. The empowered dog has a measure of patience that the insecure dog does not have. For the dog to be able to maintain an optimism that reinforcement is coming, in other words, you could say he trusts the trainer is good for it, even though it's not there and he can't see it, that's delayed gratification.

We're not … nobody, not just dogs, but we are not hardwired to choose delayed gratification. This is not the innate measure that curiosity is. Delayed gratification doesn't really play that well in the wild, and it's not an attribute that our dogs are likely to take to naturally. There are dogs who stand back and watch. There are dogs who are more passive and not more active. But nobody chooses delayed gratification. It's learned.

The way I teach dogs to choose delayed gratification is to set up games, of course, because you're talking to me, but it's all fun and games, but I'm setting up situations that allow the dog to make choices that are going to show him via A-B comparison that the better deal for him is in delay. And it doesn't have anything to do with our telling him what to do. He chooses what to do.

It's very important to me that the dog himself say, "I'm just going to stand back and watch, and then I'm going to go on cue." That's very different from the handler saying, "Ah-ah, you can't go until I tell you." Do you see what I mean? Or "Yes, you can break your wait, but I'm going to cover the bowl full of cookies." I just don't do that. I think there's a better way, if I can just be that blunt. I don't play it that way, and I think, too, in a certain small way we're lying when we do that. "You want to go check out the cookies? Oh, can't have them yet!" I don't like that game.

I like the game where we start … for example, in my game Cookie Bowl, the dog is encouraged to investigate the bowl before we would cue the dog to wait and also even after I've cued the dog to wait. He's invited at any time to go and check the bowl. Nobody's going to mention it, nobody's going to say that he shouldn't do that, nobody's going to rush in and try to beat him to the bowl. There's just nothing like that going on. Well, why not? Because the bowl's empty. He can check it any time.

If he waits till I cue … I have such wonderful videos of this of my little baby Koolaid. I'm so glad that that dog was able to go through my Cookie Jar Games class as a very young dog, before she really was mature enough to learn all of the games. The games that she could play, which actually was most of them, taught her so much about the value of delayed gratification, and so she is pretty darn good. She's not a patient dog by any means naturally, but she knows the value of delay and she is apt to choose the more patient route simply because she has faith that it will pay better, and it does. That's how it is working with me. It's a good deal for the dog when it's a good deal for me. It's not a win-lose training situation. It's always a win-win.

I'm just going to expand a tiny bit on that game that I call Cookie Bowl that I love so much, and it's now going to take a big role in Advanced Cookie Jar Games as well, because I think it's that useful. When you use the dog's own food bowl as the cookie bowl, that's a powerful stimulus operating on the limbic system of the brain. You don't really need to fill that bowl with cookies in order to invite the dog to make a mistake, which is what I would say people are doing when they fill it up with cookies and then cover it if the dog moves. That's not necessary.

The very stimulus of using a stainless steel feeding bowl, if that's what you normally feed your dog in, or whatever bowl you would normally feed your dog in, or a dinner plate, if that has a very special meaning for your dog, whatever it be, the very stimulus of that bowl being placed on the floor is huge. It's challenge enough. It doesn't need cookies in it in order to strengthen the dog's resolve to wait.

So the way I do it, the dog is invited to check out the bowl at any time. However, if the dog chooses to hang back, then I'm going to give initially a short time, of course — obviously we start with short duration and then increase it; that goes without saying, and yet I should say it — if the dog chooses to hang back, I'm going to see that little flicker of self-control and I'm going to quickly, quickly give the cue that cookies will now be available in the bowl.

Keep in mind: the dog already knows there are no cookies in the bowl, the bowl is empty, and the bowl is empty when I cue the dog that cookies are in the bowl. Therefore the onus is on me, the trainer: that cookie had better hit that bowl before the dog gets there, or I'm lying again. So I really try hard to develop the order of events just so, and develop the timing and set up the situation so that everybody is successful, nobody is lying, everybody wins, and the dog gets trained to hold back in an empowered way as opposed to a coerced way.

Melissa Breau: I think for most people this next one may be an obvious question, but I wanted to ask it anyway because I think it's important to connect the dots. Why is having a dog who understands the concept of delayed gratification so critical if you want to do a competition sport?

Julie Daniels: I think all the competition sports that I know about require an element of hurry up and wait. I call it "hurry up and wait" because speed is usually a good thing while we're executing our drills and exercises in competition. Snappy responses are valued more than lethargic responses. Everything should look like a happy game and not look like a chore.

So we want that eagerness, and yet there's always an element of "not yet." That element of delay is critical to competition, and it, by itself, is not what the dog would choose to do. Again, not one of those innately easy things if you want the eagerness and yet you also want the patience part, the waiting part.

That dichotomy, I think, just develops beautifully in classes like Cookie Jar and Empowerment because the waiting is self-imposed. The way I train it, it's embraced by the dog as a good deal, and yet the eagerness pervades every single game that we play. So you get the whole picture, you get it all — the dog who just is so happy to have his turn and so eager that he's right there playing that game in the moment, and yet he's solid as a rock on his waiting.

I'd love to give you an example from my own life, a dog whom I adopted as an adult, and his name was Colt. I had him for about four years and then lost him to epileptic cluster seizures, so just a heartbreaking story I won't go into. But his training then began as a dog who was over a year old, and he was lovely and loved in his previous home but just not trained to do anything. He had no sport. Of course my favorite sport is agility, and he took to training. It wouldn't have mattered what I wanted to do with him, I think he could have excelled.

Within the first week or two that I had him, I was doing a seminar a few states away and I took him with me, particularly as a demo dog for the foundation group because I wanted to show them something, and that was especially that here was my brand-new, wonderful dog Colt just embarking as an adult on his agility training. He had no stay, and no concept of stay, and he really couldn't stay, but he had an unbelievable start line. He just had a fabulous concept of when to go — when to hold back and when to go. He was a big boy, so it was just so easy to see in him the eager self-restraint coupled with the shining bright confidence in his eyes.

If you can picture it, leaning forward, solid as a rock, trembling with anticipation, and going on cue, but couldn't hold a stay for anything. I hadn't had him long enough for him to have developed up to step four of mat work, so I had no stay on the mat. Yes, he loved his mat. He was just where anybody else might be after a couple of weeks of training on his mat work, so it was going to be fine, but already he had that concept of the advantages of holding back and the fun of exploding forward on cue. So there's my beginner sports dog story of the value of teaching the way I teach.

Melissa Breau: I know we got into this a fair amount already, but is there anything else you want to add about how you approach the topic, or maybe even another game you want to share with us?

Julie Daniels: Well, the classes blend so well together. Would you like an Empowerment game or a Cookie Jar game?

Melissa Breau: I will leave that one up to you, what comes to mind.
 

Joy

Location
East Sussex
I have some reservations about the article. They suggest that you use an empty food bowl and that when your dog chooses to not to go to it, then go to it together and you put food in it for the dog to have. I can see that this may work but I don't see how it's better than having some food in a bowl with a lid ( I don't mean rushing to cover it before the dog can get there, I mean showing the dog that the food is there but not available at that moment). In both cases you are building value in the human, as in both cases the human has to go to the bowl to either put the food in or remove the lid.

There is also the problem that if you are going to put food in the bowl after the dog has done what you want, then you still have to either carry food on you to put in the bowl or go to another place to get it to put in the bowl.

They say that by using the dogs normal feeding bowl, "that's a powerful stimulus operating on the limbic system of the brain" - I presume they mean that the dog has already built an association between the bowl and a primary reinforcer. However that is not necessarily the case. Molly receives most of her food either scattered on the ground or in Kongs and I quite often forget to pick up her food bowl after it's been used, so I dont think she has especially strong positive feelings with it. It's also easy to quickly build positive associations with any container that the dog receives food from -after only a week of using the same lidded container with Molly she is showing signs of excitement when I produce it.

Bramble is lovely in your video, but I don't think it is what the article is suggesting. In your video the food is acting as both a distraction and a delayed reward. I have done this with Molly, but having food in an enclosed container seems to be working better for us. For a start we are mostly training in public places where other dogs could get the food, and even if we are in a place with no dogs if she's consciously having to resist the open food bowl I dont get such good engagement- if it's lidded she seems to accept that she needs to do whatever I'm asking first.
 
Top