Webinars

Mistake protocol, hard/easy and Rewarding the Try

What do you do when the dog either does the wrong behavior or doesn’t respond to your verbal cue?

I have a couple of strategies that I employ when I’m working on cue distinction.

Mistake protocol:

At its simplest execution, this means:

• If dog makes a mistake,
• Repeat the circumstances that made the mistake
• If dog makes 2 mistakes in a row, simplify the task

Let me take 2 examples and show you how I would modify the training

Down stay:

• I’m teaching the down stay and the dog gets up when I move 4 steps away.
• I reset the dog in exactly the same position, and go 4 steps away again.
• If the dog is successful, I go back and reward, and then do 4 steps away again (not 5!)
• Repeat a couple of times and then end the training session
• If the dog gets up that 3^rd time, then I reset, but make the next rep much easier, say 1 step away, reward, 2 steps away, reward. The dog just told me that he can’t do 4 steps away at that time.

“But, wait a moment,” you say, “He did 5 steps away multiple times yesterday!” And I reply, “Yup, but today, right now, he’s telling us he can’t do 4 steps away.” Listen to the dog. Because when you do, he’ll be able to do 6 and 7 paces away the next day.

Tuck sit:

I’ve lured my tuck sit and I’ve named it. In the home environment, my dog can do the sit entirely on verbal cue with no physical signal needed. I’m in a different environment, I’ve done the threshold test and my dog is paying good attention to me and not the distractions.

• I cue sit, dog looks at me and doesn’t sit
• I move a step to the side, encouraging the dog to move with me
• Cue sit again
• Dog sits! And I’m happy. I’d probably cue a couple hand touches or something the dog finds easier before asking for sit again.

Let’s say the dog doesn’t sit for that 3^rd time?

I’d make it easier by physically cuing as well. If that didn’t work, I’d abandon sit for the time being and move on to something way easier. Like the down stay example above, he’s telling you he can’t do it right now.

Hard/easy

It’s good to always have this ratio in mind when thinking of upping criteria and sequencing behaviors. Teaching and practicing skills in a hard/easy ratio keeps the rate of reinforcement up and prevents us from constantly making things hard, harder and hardest, which can be quite hard on our learner!

Let’s go back to our down stay example and our 4 steps away:

Instead of going from 2 steps away, reward,3 steps away, reward, 4 steps away, and the dog gets up and makes a mistake, it would be kinder training (to your learner!) to teach,

2 steps away, reward, 1 step away, reward, 3 steps away, reward, 2 steps away, reward, 4 steps away, reward, 1 step away, reward, etc… The dog would be much less likely to make that mistake at 4 steps.

Reward the try

Often times when I am working on cue discrimination with my dog, I’ll get into a rhythm of Get it/reset treat, feed in position, reset treat, feed in position, reset treat, skip the feed in position if the dog doesn’t execute the correct position.

That looks like this:

• Get it
• Sit-click and feed in sit for sitting
• Get it
• Down-click and feed in down for downing
• Get it
• Stand- dog sat
• Skip that treat, cue get it
• At that point, I can either cue stand again, or cue something easier

Believe it or not, when you establish this rhythm of 2 treats for each correctly executed position, the dog is well aware that he missed out on a treat!

I also tend to “reward the try” if my dog gets stuck in shaping, by not clicking but throwing a reset treat with a Get it cue. Usually by that time I’m on a “reward in position, throw a reset treat” ratio so again, the dog doesn’t get fed a treat and the ratio is interrupted. I’ll also reward a recall even if using the recall to reset a behavior or call the dog off the wrong behavior, like when I’m working on sending to multiple target areas.

No Reward Markers:

I don’t formally use one any more. I’m sure that on occasion words slip out of my mouth (I’m human!) but in general, I don’t tell my dog that he’s wrong when he tries something. I tend to interrupt with a recall if it’s possible, or withhold the reward and wait and see if the dog will adjust into the right behavior, or cue again, or move the dog out of position with my body language. It depends at what stage of training we’re at, as to what I will do at the time.

Thoughts on mistakes:

One of the things that really strikes me nowadays in my training is that when you cue a behavior, and the dog does something else, he thinks he did the right behavior! And if you have a confident sure of himself dog, he REALLY thinks he’s right. So you have to be careful how you treat that scenario, because I don’t want to erode that confidence and the happiness that comes with it. Many dogs are quite generous with our training, they figure stuff out in spite of our bad timing, our lumping skills, and general confusion. But they still show the stress. Sniffing, being “distracted”, vocalizing, jumping up, nipping, and leaving us are all signs of stress in our learner about the training session. I try my utmost to avoid those in my training!