Click your heels: Analyzing complex behaviors

This year has been incredibly busy for a variety of reasons!

In addition to my dog training business, I was in charge of planning the 8th annual Art and Science of Animal Training Conference and, for the first time this spring, I taught an undergraduate behavior analysis class at the University of North Texas. Unfortunately, this all has meant very, very little time for blogging. However, I’m planning to get back in the habit of posting updates on a more regular basis.

A sleepy pup at the Cincinnati ClickerExpo

A sleepy pup at the Cincinnati ClickerExpo

I’ve been going back through my notes from the March ClickerExpo Conference in Cincinnati. This year, I was lucky enough to get to attend both of the US ClickerExpos because Dr. Jesús Rosales-Ruiz and I co-presented a lab session about the shaping game PORTL.

At the Expos this year, I got to hear Hannah Brannigan speak for the first time. Hannah is a great trainer, and I loved her lecture on “High precision, high scores,” because of her extreme attention to detail and her creative solutions for reteaching behaviors. You can find a few notes from that talk in my Reno ClickerExpo notes blog post.

So, at the March ClickerExpo, I naturally wanted to see the other talk Hannah was giving at the Expos this year. The talk was titled “Click your heels! Heeling for performance,” and was all about the behaviors that a dog and handler must master to correctly perform the heeling exercises that are required in competition obedience, rally-o, and other dog sports. To succeed at these exercises requires much training and teamwork, as heeling is a pretty complicated behavior. To score maximum points, the dog and handler must move completely in-sync together, with almost invisible cues.

Hannah’s lecture broke down the behaviors and skills needed for heelwork. I don’t compete in dog sports, but I still really enjoyed this lecture and learned a lot! Hannah talked about a lot of things that apply not just to heelwork, but to training any sort of complex behavior or chain of behaviors.

Know your criteria and know what you like

What is heeling? Heeling involves all sorts of different skills and component behaviors on the part of both the dog and the handler. Put them all together, and it is a pretty complicated skill set. To train and compete successfully, the handler needs to understand all of these different component behaviors and how they fit together.

Many times, it is easy to want to jump in and start training, without really understanding the final behavior you are hoping to train. So, whether you are training heelwork or something entirely different, spend some time doing some carefully analysis BEFORE you start training. As Hannah said in her presentation, it is really hard to train a complex behavior if you do not fully understand the behavior you are teaching.

You should consider what various movements the dog will do, what movements you will do, what cues you will use, how all of these behaviors and cues will fit together, and what criteria you will use to decide if you and the dog are doing each part correctly. Find examples of teams performing the behavior smoothly and accurately, and study how they move at each step. In her presentation, Hannah went through some of this analysis that she has done for heelwork, down to details such as whether the handler is better off leading with the right or left foot when making a turn.

Ultimately, though, for most behaviors there will always be a subjective element to what exactly is the “correct” behavior. For example with heelwork, certain judges will prefer a slightly different head position from the dog, a slightly different overall balance in the dog’s body, and so on. Hannah said that every judge has a slightly different picture, and you’ll go crazy if you spend all your time trying to train for what a specific judge might like. Instead, figure out what variations you like best, then train for what you like and what you want.

Consider your cues

Much of heeling is context cues. The dog stays next to you in position, turns as you turn, stops and sits when you stop, and so on. The handler uses very few verbal cues or hand signals.

What if your training accidentally results in the dog doing an incorrect or sloppy version of the behavior? Your body movements and context cues are going to be associated with behavior you don’t want. This can make it difficult later on to clean things up and re-train the right version of the behavior.

Hannah suggested that many of the component skills that are required for heeling, such as teaching the dog to pivot in place or practicing position changes, can (and should) be taught with the dog not in the heel position. If you mess up the behavior or don’t like how it turns out, no sweat, retrain it. Then, when you have the behavior exactly how you want it, you can transfer it to the heel position.

I think this is a great idea any time you must use a certain cue for a certain behavior. Shape the behavior initially with a different cue or in a different context. Use a “training cue,” rather than the cue you will use for the final performance. Then, once the behavior is just the way you want it, transfer it to your final cue. This method is a smart way to train because if you do get errors or the incorrect version of the behavior, these mistakes don’t get associated with your final cue.

Using reinforcement strategies to your advantage

Before you start training any behavior, it can be important to consider the emotional state and energy level that you want the animal to have when performing the final version of the behavior. For example in heelwork, trainers usually want their dogs to appear happy, alert, focused and confident. They don’t want a sluggish, disinterested-looking dog!

The dog’s emotional state during training will become part of the behavior. So, it is important to consider how you will induce and maintain the correct emotional state both during initial training and then during practice sessions. Different reinforcers and different reinforcement strategies can be used to encourage certain emotional states.

Hannah discussed four different reinforcement strategies that she often uses when teaching and practicing heelwork. This included:
Hand delivered – The dog takes the treat from your hand. The treat is given high, so that the dog’s head tips up, but not so high that the dog’s feet come off the ground.
Jump up – The dog jumps up to get the treat. This can be used to increase energy and arousal level.
Tossed – The treat or toy is thrown so that the dog can go get it. You can use the direction you throw the reinforcer to influence the dog’s position and movement.
Stationed/placed – The reinforcer is off to the side somewhere, and the dog gets to go to it when you give a certain cue. This can be used to add distractions and add an element of self control. This can also be used to reset the dog, so that he has to then find a certain position again.

So, when you are training your dogs (or other pets) don’t pick your reinforcers or reinforcement delivery strategies arbitrarily. Instead, consider the behavior you are training, the energy level you would like from the dog, and how the particular reinforcer and reinforcement delivery strategy will influence how the dog performs the behavior.

In conclusion

Anytime you are training a behavior, whether teaching a dog to heel or training something entirely different, there are always lots of things to consider! But, I encourage you to spend a bit of time thinking about your criteria, cues, and reinforcement strategies the next time you start training a new behavior.

If you liked this post, take a moment to share it!

, , , , ,

Don't miss out on great information about animal training! Subscribe now to the Stale Cheerios newsletter and receive email updates when new posts are published.

Disclaimer: StaleCheerios posts occasionally contain affiliate links. Affiliate links are one way that StaleCheerios can continue providing top-quality content to you completely for free. Thank you for supporting our hard work! Learn more here.