Training focus including eye contact is not something new to me. It’s woven into most things my dogs and students work on. There’s a huge history of my dogs being reinforced for looking AT ME to get the thing they want, and away from the thing they want (you know, to exhibit impulse control and not just grab whatever looks good, but rather to go “through me” to get their rewards). For scent detection, you need to remove yourself from the picture. Us humans are so olfactorily retarded (if I may make up a word) that we ignorantly might try to “help” when we are pretty clueless about exactly what is happening to the odor in the environment. So training a good alert is quite a different game for us. There is an agility exercise to train a dog to focus away from the handler at a particular obstacle, but it is cued by the handler. The stare for an alert needs to be cued by the odor, not the handler.
In order to get that intent stare at the highest concentration point and source of the odor, I needed to experiment. There’s a lot of work to do before even introducing the search component. Otherwise, you get something like this:
After talking with Jen and Steve, I decided the idea of developing the alert without the odor present made sense, so that I could start proofing the behavior and complete the understanding of it before attaching the cue. If you don’t know, the cue (the odor) is only supposed to be attached when you are getting the terminal behavior (otherwise you’re naming something else, right?). So, next to figure out how to get the dogs to sit and stare strong and with duration, while trying to distract them.
Let me save you a whole lot of time and recommend that you figure out a way to make the reward appear to come from the source. Even throwing the reward onto the odor target (though that is better) isn’t as good as it appearing to come from it. Here’s how to attempt to throw the reward correctly.
Building the behavior with duration will be close to impossible if you noticeably bring the reinforcement in. Even if you mark exactly when the dog is alerting, if he breaks the alert and looks at you, you have a lot more work to do. I wasted over a week trying to train a stare with a laser, using a desirable item (toy/food)…and then I got a remote treat dispenser. I needed a good reason to get one, and it’s so much fun!!! I should mention that I think a remote treat dispenser in the wrong trainers hands can teach a dog to ignore the person, which is normally not recommended for a companion dog’s training. But that’s exactly what I needed here. In a remarkably fast time I have multiple dogs sitting and staring, with such determination that proofing can begin. Once a dog has a strong enough focus in the correct alert position, I’ll attach the odor. Then to quickly change from the giant treat dispenser to less visual tubes. Eventually there will be no dispenser at all, but to get the alert built up I can’t think of a better way. Experience instructs me to prevent any dependence on aids, so containers and dispensers will be changed as often as possible (and early on).
So to summarize and give you some great links, the goal is to get the dog to stay glued to the odor in spite of you. Check out Randy Hare’s excellent version of “box hides” that makes a lot more sense than setting cardboard boxes in the middle of the room.
They’re running with that for SAR, too:
Check out a super cool device anyone with a ball dog would love to have! They call it a BSD, in my mind that’s a ball shooting dealy, but yes, it definitely is a behavior shaping device. There are a few videos in that series. I can’t get enough of that BSD! Super fun!
Do you think that I should put the alert on a verbal cue before adding the odor? Or just let the odor be the cue?