It's easy to be a "motivational" trainer when everything is going right. Feed cookies, play with toys and watch the behaviors accumulate. It's all good!
What do you do when it's not going so well? When you have a problem and you are struggling to fix it? When all of the common solutions offered resort to traditional methods that involve compulsion? When you've run out of ways to make it clear to your dog what you want? When your dog knows what you want, but chooses not to perform because disobedience is more fun than obedience?
Well, that's pretty much where the rubber meets the road, isn't it?
I'm in the midst of this right now. I came very close to failing a trial a few months ago, because my dog didn't much want to to let go of the protection decoy when I called her. Biting is fun. Recalling to mom - not so much.
I went into the trial knowing I wasn't properly prepared, so my results weren't a surprise and at nine years of age, it didn't seem wise to pass up the trial opportunity when I knew no harm would be done.
You can see the biting exercises here; note the multiple whistles where she goes into a barking guard rather than returning to me: http://www.youtube.com/watch?v=cMlc6kIhQ6U
I'm going to explain the required behavior, where my dog is failing to perform, the traditional solutions, the traditional downfalls, and then I'll go into the decisions I made and where I'm at now. Many of you will want to stop reading right now, because this is quite detailed and specific - unless you are interested in ringsport or my approach to problem solving with my own dog, you can move on.
The required behavior is that Raika release her hold on a decoy who is fighting with her when I blow my whistle, and then immediately return to heel position. I will be anywhere from five yards to fifty yards away, depending on the exercise.
The problem is that Raika almost never releases on the first whistle. She often releases on the second whistle, but doesn't necessarily return to me. On the third whistle, she almost always returns correctly to heel. The further away I am the worse she is.
The traditional solution would involve either a prong collar or an electronic collar. Any failure to release and recall would involve a collar correction. If the collar were a prong she would either be corrected by myself (holding a long line) or the decoy (reaching for a tab on her collar). If the collar were an electronic collar, then I would supply the correction from across the field.
These methods would likely work on Raika. She's a sensitive enough dog that she would respond if she were compelled through pain compliance techniques. Or at least they would work for awhile; it's certainly possible that the pressure would have to increase over time or under high stimulation levels as she adapted to the training, or that I could inadvertently create other problems which would then need to be fixed.
The downfalls to this method could include the following:
1) Dogs learn when they can or cannot be corrected - wearing the special collars or being attached to a long line signals to the dog that they must be obedient. That leads to the classic problem called "trial wise" behavior. The dog learns that when the place "looks" like a trial - complete with a nervous handler, a judge on the field, no rewards visible, no collar on the neck, and all of the other cues that signal an "event", then the corrections will not happen.
Another downfall includes the potential development of nervous or hectic behavior on the part of the dog. Instead of focusing on the job, they remain in conflict between what they want to do (bite) and what the handler wants them to do (obey). Note that I didn't say the handler wants them to "release" because that is incorrect. We are not trying to extinguish biting, we want them to release AND bite. Corrections can create a dog who refuses to bite at all, or who becomes very nervous and frantic in their work.
Both of the above scenarios are more or less avoidable by a very sophisticated handler with good timing and a clear headed dog that isn't easily stressed or worried. Then again, very sophisticated handlers with good timing have fewer problems in the first place.
Regardless of whether pain compliance would give me my recall, I'm less interested in solving the problem than in seeing what I can accomplish using motivational techniques. For all I know she'll never compete again, but I still want to figure out the puzzle.
Solving the problem:
I started by teasing apart the problem - determining what she knew, what she might not know, and under what circumstances could she perform correctly, if ever.
I believe she knows the meaning of the whistle when she is not in drive. If I blow the whistle while she is running about, she comes immediately and with a bright attitude.
I believe she knows the meaning of the whistle when there is a toy she wants nearby. If I blow the whistle she will come back to heel position and wait to be sent to the toy.
On the other hand, if I whistle her back when she is engaging another other person with a toy, then her rate of return suffers quite a bit, even if I also have a toy that is equal or better to use as a reward for her return.
If I have a toy and a second person is wearing a bite suit, then her rate of return is even lower.
From these data points I believe that she is able to respond to a whistle cue under simple circumstances, but not when her drive levels are brought up higher through the addition of a second person or a higher value bite toy (suit or sleeve).
Unfortunately, I rarely have an opportunity to work with a second person, so testing my hypothesis can be tricky. Only so many people enjoy playing tug with my dogs, and even fewer want to wear a bitesuit.
I started experimenting. She returns on the first whistle the vast majority of the time when she is not actively biting, even if she is facing the other person. Good! This reaffirms my belief that she understands the whistle.
Next, I had a friend play with a low level toy and then I whistled. Failure. Sometimes she comes back. Sometimes she lets go but stays with the person. And sometimes she thrashes all the harder. Hmm. She either doesn't understand when in higher drive or she simply doesn't want to return so she refuses to recall.
I decided to try negative punishment. I allowed her to play tug with my helper. If she returned on one whistle, I immediately sent her back to play some more. If she failed to return or went into a guard, I placed her on a "time out" in a crate or alone in a training room for a minute or so - no more biting. She didn't like this at all.
After several sessions, I saw no improvement. Indeed she appeared worse; becoming more hectic, barking, and even less likely to return at the whistle.
The hectic behavior suggests that she truly doesn't understand what I want or is unable to perform in her current drive state. She gets hectic because she knows she is about to have a time out, but she is unable to make a good decision to avoid that scenario.
If she is not learning from punishment, then I need to find a way to make her right often enough that she can keep a clear head and make the decision that will get her sent back to her bite as soon as possible. But what exists between facing a dead toy on the ground and being whistled off (at which she runs close to 100% success) and a dead toy that my friend is holding (at which she runs no more than 30% success and declining?)
Logically, if she understood why she was getting the time outs, she'd simply change her behavior and cooperate. She is not. Now what?
Who could I ask for help that would understand the problem and possibly have some solutions? Very very few people have experience with this sort of challenge and motivational problem solving.
I called my friend Shade Whitsel, a motivational trainer who competes primarily in Schutzhund. The first ten minutes was a thorough review of what I had done along with a description of Raika's temperament and training background. After this initial discussion, Shade suggested the following directions to explore:
1) separate out the whistle/recall from the release. Whistling off of an object requires two behaviors; the first is letting go and the second is returning. To begin, I was to play tug with my dog. On the whistle, I was to ask for a release, until the whistle automatically meant release. That part was easy since I could work on it as much as I like - it didn't require another person. After very little time, Raika was cleanly releasing her toy on the whistle command. Her reward was re-engaging her in the game. So far so good.
But I still had the recall to work through. How to do that without a second person?
Shade suggested hanging a bite suit to a strong object. Send Raika to bite the "dead" bite suit. Whistle off - and immediately send back if she succeeded. I haven't tried this yet because I have to hang the suit somewhere. Also, I'm not sure how enthusiastic Raika will be about biting a suit with no one in it. Regardless, it is a path to try.
Finally, she mentioned another ringsport trainer who trains alone. This person taught her dog that the whistle meant release the suit and return to a spot behind the dog (the start line). Basically, teach the dog to go back to the line of departure where the handler would normally be standing on the whistle cue. Hmm. Interesting. Now the whistle would no longer mean recall as much as run away from the direction of the bite. if a second person were available, then I'd be standing at that spot (start line).
I thought about this last option. My biggest concern was teaching her that my whistle meant to leave me (the handler) rather than to recall. On the other hand, once I'm wearing the bitesuit I'm no longer the handler, I'm the decoy. There is no handler. Play tug, whistle, and send her back to a spot. Over time, work to the point where the whistle was the cue to go back to the spot. The reward for cooperation? Let her bite again.
Then I asked Shade the classic question that I get asked all the time by traditional trainers. What do I do if she does it wrong? And she gave me the answer that I should have known - do nothing. The same as when she learned the out in the first place - be patient. Wait, and let her be right enough that she can succeed. Focus on what I want, not what "might" happen. Ignore the mistakes in the training process and work hard to set her up for success.
And we've begun. If this program doesn't work then I'll try something else. Eventually one of two things will happen. Either I will decide it is a viable training option (in which case I will suggest it or variations of it to others struggling with this problem), or it will fail, in which case I will not suggest it to others, and I will go back to finding a solution. Regardless, I have won.
There's nothing like having someone to brainstorm with, so thanks to Shade for talking me through this. And thanks to the many folks and decoys who are intrigued by my challenges with no force training and who give a good effort to implement whatever plan I come up with - whether they agree with my choices or not.
Note: Nine days later: It's clearly working: http://www.youtube.com/watch?v=jhudpkbBXJE
Note: Sixteen days later: Tested with a new decoy working Raika - not perfect but a huge improvement in whistle recall. *happy dance*:).
Working through this has been a huge high for me - applying what I know, asking for help as needed, staying true to my beliefs, and making progress. I'd guess that we will kick this in a couple of months - then we need another trial.