It doesn't matter what you're training, whether it's the basics or complex competition skills, there are only three reasons why your training goes wrong. It doesn't matter who your learner is, a new puppy or a trick dog champion - the reasons are still the same. Here's one of my favourite meme versions.
It all comes down to timing, criteria or rate of reinforcement. There could be a combination going on, but it's usually just one at a time. Let's have a look in a bit more detail.
Timing - this could be the timing of your cue, a lure or your marker. As a practical example, there is an exercise in Rally called 'sit, send around, sit'. If I am too late saying 'sit' as Ripley goes around me, then he may end up sitting in front of me. Of the cue is too early, he may sit behind me. Timing of the cue here is critical to make sure that he sits at the right point, as is knowing what works best for my dog (early is usually better than late for Ripley, given the speed he lives at!)
With markers, it's generally better to click too soon rather than too late. When training a nose touch indication for scentwork, if you click too late then you are likely to be clicking the dog's nose coming away from the scent. If you're using a verbal marker, then sooner is definitely better than later because it takes a fraction longer for them to process a verbal marker than a click.
Criteria - another one that can go two ways. We either stick to lower criteria for too long, or jump our criteria up too high too quickly. How your dog reacts will likely dictate how you deal with the situation. If you're shaping something, and haven't upped criteria then you may need to add a lure briefly to get your dog to progress to the next step (and try not to get stuck there too long!) When you've pushed for something that your dog isn't ready for, don't be afraid to abort the session. Take a break and come back later, starting at where you were successful, or even further back.
Taking the scentwork indication example, don't try to jump from one second of duration to five seconds - your poor dog is likely to fail! You're far better using a ping ping approach, where you jump up and down between shorter and longer durations. Make sure to keep plenty of repetitions of the short duration to keep them successful.
Another example would be shaping your dog to step into a box. You might have got to the point where they're putting one foot in consistently, but now you're stuck there. With shaping, you only want a few successful repetitions before increasing criteria otherwise your dog will think that they have the behaviour nailed - they're being paid well for putting that one foot in the box, why do more? In this case, placement of your reward will help to get them moving on again.
Rate of reinforcement - in a shaping session, your rate of reinforcement needs to be HIGH. When you're working on something difficult, you need to pay well. The rate at which you reinforce is possibly more important than WHAT you use. Think about a young dog that you're getting ready for competition, in pretty much any sport. They know all of the individual exercises or moves, and are performing them confidently one at a time. One exercise = one reward. Trying to do ten exercises for one reward pretty much guarantees failure! Of course, you could class this as increasing criteria too soon as well (I did say there may be overlap!)
Now that you know about these three reasons, can you see how they apply to a training problem that you're having? Head on over to the post about this blog on my Facebook page if you want to discuss this further.
Happy training
Have you got your dog's attention? For five fun games to super-charge your relationship with your dog, click here.
Comentários