The Shaped Pick Up and Hold
Gunn Anita Fossli
The concept of shaping isn't always the most efficient way to teach or modify behavior, but for some things it can definitely be very beneficial. Armin and I have found that the retrieve is one of the things where shaping can go a long way to make learning easier for the dog and teaching easier for the handler in contrast to other traditional approaches to teaching this skill. While I was formally trained and certified as a clicker trainer and instructor, I do not consider myself a "pure" clicker trainer anymore. In the broad spectrum of sport and real world dog training a trainer can never limit him/herself to one singular approach. The mark of a true expert is one who can utilize all the great functional approaches and concepts out there and make them work to make training fair and easy to learn for the dog and simple and enjoyable for the handler. Together Armin and I are exploring and discovering ever more effective and efficient ways to introduce, teach, solidify and proof important skills and techniques.
Generally, a "pure" clicker trainer will teach the dog to repeat behaviors and then chain them together backwards. This is a concept that can work wonderfully, but it requires (like most dog training techniques) good timing and correctly set criteria. Although I am not a pure clicker training anymore (using solely positive reinforcement and negative punishment – easily explained reward or withhold of reward), I like this particular approach to teaching the retrieve.
I like to start in one of two ways: Either by me presenting the object from my hand, or by having the object lying on the ground. The general idea is that the dog should figure out that showing interest and finally picking up and holding the object, will lead to reward. I like to use treats for rewarding, at least to begin with. It ensures you more repetitions, there is no potential outing conflicts and the dog isn't heavily in prey drive, which might complicate things. Later, as the dog has grasped what leads to reward, I might use toys to reward with, and also to make sure the behavior is solid. When I start chaining behaviors together I might also bring a toy into the equation, but that will of course depend on the dog.
The basics that need to be in place before even trying this approach is that your dog wants to be with you and train. If you have to struggle to hold your dog's attention, this approach will not function. So I like to ensure that the dog can stay with me when I bring him out to train, that he truly wants to interact with me. In any case I always keep the dog on a leash in the beginning, so that I can stop the dog physically, and without having to get stern or "personal" during training.
Shaping is easy, and then not so easy. It is based on building on whatever behavior you get to begin with, keeping the criteria for reward low and simple at first, and steadily working towards the behavior you want. Remember that "the click ends the behavior." This means that when you have clicked, the dog can stop whatever he was doing you were rewarding. In the beginning let's say sniffing at the object, the moment you click he can stop showing any interest. Later when he starts putting the object into his mouth, he is allowed to let go of the object when you click.
From your hand:
Present the object you want the dog to eventually retrieve. It shouldn't be the dumbbell, but some object that is similar. The dog can stand or sit in front of you, but he should not be under any kind of command at this point. If the dog looks at it, sniffs at it, shows any interest towards it, click and treat.
Do this several times and make sure the dog actively tries to get towards the object. It is crucial that you don't try to make the dog interested in the object, tease the dog with it, or try to push it at the dog. It is supposed to be the dog's initiative to try to get to the object. Repeat this step until the dog reliably tries to get at the object whenever it is presented. I typically do 5-8 repetitions per session. I like to keep the sessions short, and rather do more sessions, rather than fewer which run too long.
When the dog is interested in the object, and tries to get at it in order to get the reward, we heighten the criteria. Now we want the dog to not only show interest, sniff or nudge, but open his mouth and grab the object. Some dogs do this very easily, with others it may take a little more time. If the dog grabs the object immediately, great. Click and treat, and make sure YOU CLICK WHILE THE DOG HAS THE OBJECT IN HIS MOUTH. It doesn't matter that the dog lets go as soon as the click comes, but it is very important that the timing is good and that the click comes when the dog still has his mouth on the object. To begin with, the dog may hold on to the object only a second, and that is ok.
If the dog is more hesitant and won't grab the object straight away, we must work on getting the dog to understand that it is grabbing the object which will lead to reward. Any kind of opening of the mouth, gaping over the object (though not biting down) must be rewarded. When the dog starts grasping the concept, we heighten the criteria again. The next step from there can be actually grabbing it and biting down. Make sure the click comes as the dog has the object in the mouth. It can sometimes be easier to have someone else on the side do the clicking, so that you can concentrate solely on the dog.
When the dog grabs and bites down on the object as soon as it is presented, it is time to start working lengthening the time the dog is supposed to hold. To begin with, we're talking seconds. It is also important that the dog doesn't chew. Work your way up from 1 second to 20, making sure that the dog doesn't chew. Make sure you click only when the dog holds calmly and without chewing. If the dog drops the object while you are working on the hold, do nothing apart from trying again. Sometimes it is necessary that the dog makes that mistake, it is part of the process for the dog to figure out what leads to reward and what doesn't. It is, however, important that the dog doesn't fail too many times in a row, as this can lead to big set backs. As a general rule we say that if the dog fails more than three times, it is a sign that the criteria for reward are too high. In such cases it is usually enough to make it a little easier for the dog (if, for instance, you are working on the dog holding the object for 15 seconds, but it keeps dropping the object after 12, you want to lower the criteria a little. Let the dog hold the object for 10 seconds, click and reward, and do this 3- 4 times, just to remind the dog what works, what behavior leads to success.) and then, try again.
From the ground:
The approach is very similar to the approach from the object in the hand in regards to timing and reward. The difference is of course that the object lies on the ground. Any interest towards the object should be rewarded, then the criteria are heightened until the dog picks up the object and holds on to it. From there, the work begins to focus on lengthening the time the dog has to hold the object.
In the beginning it doesn't matter what position the dog is in, whether it sits or stands. When the dog can hold on to the object 15 - 20 seconds, I like to have the dog sitting in front of me during the work. I also like to reward the dog a few times for coming into front position without the object in his mouth, and then introduce the object again. What often happens then is that the dog repeats the behavior which was last successful, which was of course coming into front position, only now he also has the object in his mouth. What has happened is that a small backwards chain has been formed.