The Game Of Know, Not No!

Posted by on May 5, 2016 in Fido Blog: Life in the F-Lane | 0 comments

The Game Of Know, Not No!

Many years ago when I was first learning this “new” method called positive reinforcement training, even though my dog was learning in huge leaps to do all the behaviors I asked him to do, I had this nagging question: But how can I tell him when he does something wrong?

Humans seem to be wired that way and are more apt to point out someone’s errors more often than looking for something correct. And unfortunately for our canine friends, we take the same punitive stance. This makes sense to some degree, because traditional dog training has its roots in the military as opposed to the science of behavior, so these trainers, being of the human species, have done what came natural to them and that was to look for the wrong and punish the dog. Sure they might toss in a “good dog” to tell the dog when he was correct, which has led many (especially those still using traditional training methods) to believe that dogs work for praise. The truth is that against a harsh background of “NO!” (and leash popping, collar jerking, hitting, etc.) it is a huge relief for the dog to hear “good dog”. The operative word here is relief from the anxiety and pressure of being punished. This also means that praise functions as a negative reinforcer (removing pressure or pain), as opposed to a positive reinforcer.

But even some positive reinforcement trainers, who would rarely if ever, engage in traditional methods, are still human, and as such have the inclination to let the dog know that he is wrong. Of course we have to provide clear, clean feedback to our animal or he won’t understand. But to my way of thinking you can choose to outright tell the dog that he’s wrong or you can do something that is totally counter intuitive and let your silence speak.

In any kind of learning situation there are phases to transition through; among these phases are when the behavior is brand new and when the behavior has been learned. So let’s take a look at some examples, beginning with a new behavior.

When the behavior is new we should be working in many repetitions for brain-muscle memory. Much like going to the gym to work out, you cannot grow a muscle with one or two repetitions.

Here’s an example of the dog learning SIT with a food lure in hand:

Repetition one: Lure dog’s nose upward Dog sits Click & treat Reset

Repetition two: Lure dog’s nose upward, dog sits, click & treat, reset.

Repetition three: Lure dog’s nose upward, dog sits, click & treat, reset.

Repetition four: Lure dog’s nose upward, dog jumps up to get the food (Incorrect response), withhold click & treat.

Repetition five: Lower hand to lure to prevent him from jumping. (Trainer makes the necessary adjustment), dog sits, click & treat, reset.

Repetition six: Lure dog’s nose upward (Trainer tests dog to see if he understands), dog sits, click & treat, cue, “Free” (End the set)

In this example the trainer is working a set of 6 repetitions. The dog jumped up in repetition four, so the trainer made the adjustment in repetition five to lower her hand, and then she went onto a sixth repetition to see if the dog understood the lure. Notice that this is a smooth, silent, quick flow from repetition to repetition. There are no verbal insertions of “Nope”, “No”, “Too bad”.

Here’s an example of the dog doing SIT, using the new hand-signal:

Repetition one: Give hand-signal, dog sits, click & treat, reset.

Repetition two: Give hand-signal, dog sits, click & treat, reset.

Repetition three: Give hand-signal, dog stands (Incorrect response), withhold click & treat, (Trainer makes the necessary adjustment).

Repetition four: Give hand-signal, (Trainer tests dog to see if he understands) dog sits, click & treat, reset.

Repetition five: Give hand-signal, dog sits, click & treat, reset, cue, “Free” (End the set).

In this example, there is a smooth flow of 2 successful repetitions (clicks and treats) and then the animal misses the behavior so no click & treat. But we continue with the next 2 successful repetitions. This was a clear communication to the animal. And more importantly, the mistake was not followed with an intrusive verbal sound, which in my opinion, at the very least breaks the animal’s concentration. Much better to have a “distinctive silence” so to speak which keeps the animal in a positive frame of mind and has him looking forward to the next repetition! Remember if you’re working with a powerful motivator (in this example high-value food), it is in the best interest of the animal to try to get it right, so no need to do anything to spoil what should be a fun game by using the verbal no-reward-marker, “No!”

Imagine if you’re the learner and every one of your incorrect answers was followed by NO, you’re WRONG!!! Actually if you’re learning to use the clicker you ARE a new learner! It’s not particularly motivating to be criticized, embarrassed, or in some way feel inadequate. While we don’t know what the dog is feeling, we do know that by disrupting this flow of repetitions with extra verbiage or even punishing tones in our voices is confusing and unnecessary, as it creates a learner that may not want to try new behaviors; this is the opposite of what we want. In positive reinforcement training, it is necessary to create a learner that will offer all sorts of behaviors and we withhold the click, sometimes because he missed the signal or sometimes because we are trying to shape a behavior. The absence of a click is his signal to try again!

Here’s an example of a dog who is fluent in the SIT behavior and it answers the question many have of what to do if the dog doesn’t do the behavior:

Repetition one: cue, “Sit”, dog sits, click & treat, reset.

Repetition two: cue, “Sit”, dog sits, click & treat, reset.

Repetition three: cue, “Sit”, dog “Spins (Incorrect response), withhold click, count silently 1-2-3-4-5.

Repetition four: cue, “TOUCH”, dog touches target, click & treat.

Repetition five: cue, “Sit”, dog sits, click & treat, cue, “Free” (End the set).

In this example, unlike the previous ones, this behavior is well known to the animal and for some reason he didn’t respond to one of the SIT cues. The appropriate course of action is to do as little as possible. In training terms this is called an LRS (Least Reinforcing Stimulus). We need to protect our valuable cues, so we don’t dilute them by cuing, “sit, sit, sit”, but rather we cue only once. Now we separate in time the missed cue from a new cue, so we use 5 seconds (in this example), and then we choose another cue that is easy for the animal such as an old well established cue (TOUCH in this case). This provides two things: one it gives the animal an opportunity to be successful and it gives him another opportunity to respond to the missed cue without any repercussions that might sour him to training.

This type of training, not only creates an animal who loves to learn, it changes the human to want to play on the same team as their dog. Any kind of “no, nope, too bad”, even given without a harsh tone, is a negative thought. By contrast, the absence of the crisp sound tells the dog, “Try again” and by giving the dog an easy cue to do tells him that you’re going to help him succeed! That’s a win, win for the team!

The bottom line is when you want Fido to “know” the training game, keep “No” out of it!

Leave a Reply

Your email address will not be published. Required fields are marked *