Raising Criteria Without Shutting Down

[brody_smom]

Does anyone have tips on how to do this? I have been struggling with this with Brody forever! There are quite a few tricks and behaviors that he has learned, but none of them are really sharp. I have actually stopped adding verbal cues to new tricks because I am afraid of poisoning them. I would like to sharpen up a few things with Brody, but I can't seem to raise criteria slowly enough(?) so that he understands that I am looking for something more that what he just did. If he offers something completely different from what we had been working on (like rolls over when he had been crawling), then I know I have to take it back a step, but then it seems we just get stuck at the same spot for a long time.

One example of this is his retrieves. We play a game with 2 identical balls and a basket. I put one ball in the basket and throw the other ball. I give a cue, "go" and Brody runs for the ball. Meanwhile, I take the other ball out of the basket and wait for him to put his ball in. Sometimes, he runs quickly back and drops the ball directly into the basket. I say yes, or just say go, throw the ball, he runs... Sometimes, he stops and chews the ball a bit before he comes all the way back. Sometimes he tosses the ball toward the basket, but misses. I won't throw the ball until he gets it in, but sometimes he takes 2 or 3 shots without really looking at the basket. Finally, he will actually place it in, then he gets the yes or go, and he's off. Can I get it so he understands that I want a quick return every time, that I want him to place the ball directly in the basket every time? I don't know how to communicate to him that, although he doesn't mind that the whole process takes so much longer his way, I want it to be quicker and more precise. I have tried picking up the ball and basket and walking away when he has done either of these things. I don't know if he makes the connection, but it hasn't made a difference.

 

[running_dog]

I'm not sure the problem with the retrieves is a criteria issue. If I was you I would look at what you think is a reward and what Brody thinks is a reward. And also look at whether you can predict when the problem will happen, is it when Brody is getting a bit bored? It sounds to me like the whole spitting a ball at the container, chewing a ball on the way and so on is actually self rewarding and your rewards are not trumping the self rewards. If he does these things you could try calmly removing the ball (cutting down his self rewarding behaviour) then give him a few calming tricks to perform before asking him to try again, if he messes around with the ball on the way in you could give him verbal praise, if he puts the ball straight in you could reward with tug or a jackpot of treats (if he isn't too excited for food, it doesn't sound like he is).

The problem with raising criteria is one I'm working through too, we have many unpolished tricks in our repertoire. I try to take the time to work out a training plan before I start, so I know what I am looking for at each stage and so I avoid looking for too much at once. I think it is Sue Ailsby who uses the terms "lumpers" and "splitters" to show how we normally lump too much together for our dogs to understand in one go when we need to split it right down into tiny increments.
eg/here's part of a training plan I might use with Gus for cross paws, obviously the suggested jackpots are only indications and the whole thing would be carried out over several sessions, the idea is to move on to the next criteria when the dog is 80% successful at the previous one:

• sitting/lying down
• lying down
• any movement of chosen paw/leg including tensing of muscles and shifting weight
• any definite movement of paw
• any definite lift of paw (click lift to cement the idea of lifting rather than moving)
• any definite lift of paw landing in parallel with or pointing towards or touching other paw (click landing to give idea that location of landing matters, jackpot pointing towards & touching)
• any definite lift of paw landing pointing towards or touching other paw (click paw in air when possible as this encourages height and height is needed to get overlap, tricky because need to predict landing. Jackpot touching and overlapping)
• any definite lift of paw landing touching or overlapping other paw (click paw in the air, jackpot overlapping)
• any definite lift of paw landing overlapping other paw (click landing)
• after this there is a divergence depending on whether I would be building duration, repetition, or increased overlap.

The level of detail may seem excessive but it saves me having to really think while I am training because I've already visualised what I am looking for. If there is no progress I stop and rethink the plan. I sometimes find it useful to do a training session when we are out walking as Zac often offers more the next time we try at home.

 

[brody_smom]

I'm not sure the problem with the retrieves is a criteria issue. If I was you I would look at what you think is a reward and what Brody thinks is a reward. And also look at whether you can predict when the problem will happen, is it when Brody is getting a bit bored? It sounds to me like the whole spitting a ball at the container, chewing a ball on the way and so on is actually self rewarding and your rewards are not trumping the self rewards. If he does these things you could try calmly removing the ball (cutting down his self rewarding behaviour) then give him a few calming tricks to perform before asking him to try again, if he messes around with the ball on the way in you could give him verbal praise, if he puts the ball straight in you could reward with tug or a jackpot of treats (if he isn't too excited for food, it doesn't sound like he is).

I don't know that he's bored when he does this. Sometimes it can be the first or second throw. I have gone back and got my clicker and treats and clicked for putting the ball in the basket. I have also moved the basket around, and at times he will put the ball where the basket was before, so we have worked on trying to solidify the criteria. Chewing the ball is definitely self-rewarding. That's one reason why I make sure the balls are identical, because there are some that are much more fun to chew than others, and sometimes he won't bring back the ones he doesn't like chewing as much. This was much more of a problem when we were playing in a large field, as I had to run out to him. Right now we're just working in the house or the yard, so it isn't as far for me to run! I will try taking the ball from him when he chews it.

The other thing he does in this game is try to roll the ball down the stairs. When I toss it for him, I am usually down in the living room and I toss it up to the next floor to the end of a short hallway. There are only nine steps including the top one. He runs up the stairs to get it, but brings it to the top of the stairs, chews it a few times, then drops it. These are fairly large indoor balls, and the stairs are carpeted, so it doesn't roll very fast and sometimes even stops. He then pounces on it to make it go, then chases it the rest of the way before picking it up at the bottom and bringing it to the basket. The game is still fun for him, but if I am trying to tire him out with some intense running, it doesn't always happen.

The problem with raising criteria is one I'm working through too, we have many unpolished tricks in our repertoire. I try to take the time to work out a training plan before I start, so I know what I am looking for at each stage and so I avoid looking for too much at once. I think it is Sue Ailsby who uses the terms "lumpers" and "splitters" to show how we normally lump too much together for our dogs to understand in one go when we need to split it right down into tiny increments.

I've done this type of thing as well, starting out with the full intention of being a splitter, but then he will go ahead and lump himself. Quite often, he will catch on very quickly to what I am planning to shape with him, and all of a sudden, my training plan is out the window. This makes it very difficult to go back one step, because he just leaped ten steps! He has done this with the crawl challenge already. There I was ready to click at the slightest movement of a front or back paw, and he crawled right toward me, nearly sliding on his belly all the way under the piano bench. He did it when I was working up to backing up onto a box in order to develop muscles for eventually doing a handstand. I had a pile of 2 or 3 books in a corner between 2 footstools to keep him from swinging his back end out of line. He was just supposed to be backing onto the books, but he bumped into the footstool and lifted his back legs right up onto it. From then on, he would just back up onto anything.

 

[running_dog]

I don't know that he's bored when he does this. Sometimes it can be the first or second throw.

Then I think maybe either it he gets bored really quickly (Zac does ), or his own self rewarding variations are trumping your rewards, or you need to explain it again .

I've done this type of thing as well, starting out with the full intention of being a splitter, but then he will go ahead and lump himself. Quite often, he will catch on very quickly to what I am planning to shape with him, and all of a sudden, my training plan is out the window. This makes it very difficult to go back one step, because he just leaped ten steps!

If he consistently offers the behaviour after he's skipped all those steps that is fine, he's just an intuitive learner. If you need to explain part of a trick to him again just go back to that element of your training plan - even is you didn't have to train that stage.

So supposing in my example Gus went from basic movement to realising he had to touch the other paw intuitively in the first couple of tries but he wasn't lifting enough to overlap the paws I'd just skip the other steps and click him as he was lifting his paw so he realised he had to lift higher - even if that meant at first his paw stopped landing touching.

In your case you want a retrieve without mouthing. One way is to go back to basics of hold, give, reward. give him no opportunity for mouthing, build duration, then build distance. The fact he can already do a retrieve if he feels like it is irrelevant, you're explaining to him how you want him to retrieve.