Clicker Training for Dogs

My No-Pull Recipe (a.k.a. "Penalty Yards")

The Birth of this Technique

I came up with this method after watching people pulling on their pulling dogs. The dogs looked to be enjoying the game of lunging forward only to be jerked back to the owner and then allowed to lunge again and again. Dogs do what is expected of them, but only if they know what is expected.

Their limited ability to generalize requires that each lesson be simplified in a manner they can understand; it must include a picture for them to refer to. That's why pulling back on a pulling dog does nothing to convey to him that the pulling behavior is undesirable, but rather reinforces the pulling behavior each time the leash is released.

It was just a few days before Karen Pryor and I were to conduct a training seminar in Tukwila, Washington, that I had this revelation and I quickly tried it out on a client’s dog, a huge Golden Retriever with a severe pulling problem. It worked! When I described to Karen what I observed and explained my no-pull recipe idea, she recommended demonstrating it at that seminar. I was a bit nervous about going “public” since I had tested it on only one dog.

That Saturday during lunch at the seminar, I conducted an “audition” of the most serious leash pullers. I chose a 10-month old Siberian Husky as my test subject. The results, in a nutshell, were that after a few minute’s work on Saturday and then just a few more on Sunday, the subject puppy heeled with attention up to and then past the target she'd previously been pulling toward. That target was I, lying on the floor while verbally and visually tempting her to pull right on over and dive into a huge bowl full of treats.

Since that day this method has been used to solve pulling problems all over the world. It works well with all sorts of targets (other dogs, children, vehicles, livestock, and toys to name just a few) and even with the variations other trainers have added to make it their own. Deb Jones renamed it “Penalty Yards,” a catchy name, since the aversive of being hauled back to a target in exchange for pulling, is indeed a penalty.

Game Overview

The goal of this game is to teach your dog not to pull on the leash without being constantly reminded not to pull. This is not heeling, but is called "loose-leash walking" by most trainers.

Applications

Walking your dog along the street or in the park where other dogs, children, and squished food on the ground will become possible without having to avoid these distractions and doing it on a loose leash.

The Set Up

You will need the following:

A starting line (can be a crack in the sidewalk or a line in the carpet)
Leash and non-corrective collar
Bowl of very desireable treats
Clicker

The Game

This is literally a step-by-step lesson and it is trained a step at a time. By starting out with the mind-set that only one step will be accomplished at a time, the trainer will be quick at issuing both aversive and reinforcement feedback to the dog. By starting small, big rewards will come quickly. Trainers who try too quickly to achieve 10 non-pulling steps will end up with a pulling dog, period. Trainers who take it slow to begin with and work diligently to achieve those first few non-pulling steps will end up with a dog that enjoys the freedom of a loose leash and will work to keep that freedom.

Establish a starting line. The starting line is the "marker" pointing out where the dog and trainer will begin their walk. A crack in the sidewalk or parking space paint will do. I used to carry chalk so I could make a starting line anywhere, and would deliberately choose “problem” areas; areas where the dog I was training showed a lot of interest in his surroundings.

Put the on a 2' to 3' leash; any longer and he tends to sway from side to side, causing the trainer to experience difficulty in quickly deciding what to do next. Keeping the dog pulling in only one direction during this shaping exercise eliminates the need to train it for each and every direction of the 360 degrees available to him. A short leash helps to accomplish this task.

A target in the form of a favorite toy, a pile of food, or an enticing human is placed about 20' away from the starting point. We want the dog to pull toward this target, so it must be something he wants badly enough not to give up too easily. What happens between the starting point and the target is the means by which the trainer measures success, so I strongly recommend food and human. The target also keeps the dog somewhat focused on walking forward, helping him through this learning process.

A verbal cue to alert the dog that the game has begun is given just before the trainer takes her first step forward. I say, "Let's go,” and nothing more. The dog will pull in about two steps at the beginning of this training, so the trainer must be ready to back up quickly. As soon as the dog pulls toward the target, the trainer changes into reverse, backing from where the dog pulled until both dog and trainer are again behind the starting line.

Care should be taken here not to add a leash pop, as there is nothing to punish the dog for… after all, there IS a motivating target in his face causing his “bad” behavior. Care should also be exercised in reversing direction – the dog is not to see the trainer’s back. He needs to see the trainer's face while realizing he is no longer traveling toward the target.

NOTE: The actual aversive the dog experiences is not the act of being pulled backward -- he's the one who chose to pull. There is more than one penalty here. The most obvious penalty is that he is returned to the starting line; a subtler penalty is the realization that no ground was gained; thethird and most powerful penalty is looking at the trainer's face while experiencing the other two penalties.

Once behind the starting line the trainer waits quietly until the dog is not pulling, sniffing, jumping up, etc. before repeating the verbal cue and moving forward. Most dogs will begin to sit or get into a sort of heel position after only a few forward/reverse actions. I have found that the magic number for a dog to catch on to a new behavior is four. Watch for changes in the dog after each three pulls and reverses. It’s quite an amazing thing to watch.
Repeat the above until the dog can go some distance without pulling, and then quit. Put the dog up, giving him some time for overnight latent learning, then perform the following steps.
If the target is a human with fairly good timing, they may click the dog for not pulling. Otherwise, the trainer, who is in a far better position to determine whether the leash is slack, should click, stop, and treat. More on this treat maneuver in a minute. The trainer then returns to the starting line to begin the game anew.
As before, the game starts only after dog and trainer are both well behind the line and the dog is not pulling, sniffing, or jumping up. As soon as the dog is quiet the verbal cue is given and the walk starts again, with the trainer being ready to either click or reverse direction.
Repeat the above steps until the dog walks all the way to the target without pulling - usually fewer than 10 clicks, and often with total attention on his trainer.NOTE: The dog NEVER eats or plays with the treats or toys that make up the target, or the food or toy a human target is offering.
Increase the distraction criterion of the target by using human targets who will gradually become more aggressive by calling the dog, touching him, and offering him food.

Reinforcing Without Interfering

A common problem experienced is stopping the dog’s forward motion with the leash, even to the point of jerking the dog back so he can be treated. This is counter-productive, since pulling on his leash is exactly what we are teaching him not to do. The dog’s first reaction when he hears the click, as anyone having done clicker training knows, is to run to the treat. The target is foremost in his thoughts and he will more likely than not try to drag the trainer toward it after hearing the click.

Pulling is now allowed behavior because he has been clicked. There must be no punishment for pulling after the click. So…the trainer must be prepared to click and then quickly take one or two giant steps to catch the dog and reach out to literally put the treat in front of his mouth.

When the dog can travel back and forth in a familiar area without pulling, take it "on the road" using anything as a target, and work the same steps as previously listed.

Anytime the dog has mastered not pulling for some distance, going back to the starting line is no longer treated as an aversive, but becomes a welcomed restart of the game, because:

He has been reinforced for correctly moving a predetermined distance without pulling.
The exercise has been aborted by his success.
If the dog views returning to the starting line as an aversive at this time, his trainer needs to let him know it is not by skipping back, praising verbally, and having a lot of fun on the way back to the starting line. At first the dog might be reluctant to return, so the trainer might have some work to do in convincing him otherwise.

Before hitting on the method I have described above, I tried every recommended way to stop leash pulling; becoming a “tree,” leash popping, prong collars, squirt bottles, having an assistant perform some heinous action upon the dog, and so on. The above ”recipe” is the only one I have found to work fast and work for all dogs, even those that are not food motivated.

There is one more part to this lesson because there is no such thing as a dog that never pulls, even after mastering these lessons.

Whenever the “trained” dog pulls, the trainer will gently stop and slowly back one or two steps. Most dogs at this stage of the training hop into something near heel position to get the trainer to resume walking (it is the most fun thing to watch).
If the distraction is really tempting, backing a longer distance until the dog is trying to get into a heel position before resuming forward motion is recommended. Random clicks whenever the dog attempts heel position will do much for his desire to remain there.
Waiting for the dog to get into a heel position before stopping backward motion will teach him that he can keep this game going by simply not getting out of position. As the dog advances, as the weeks go by, he realizes that it is he causing both the aversive and the reinforcement in this game.

About Eating the Target

One could say that being allowed to eat the target as reinforcement for not pulling toward it makes not pulling a self-reinforcing behavior. If that were true, allowing a dog who is learning not to pull to eat the very target that is wholly designed to entice him to pull, would be a really good idea.

However, think about that for a moment, and picture in your mind how that training scenario would go. Then answer this question: How long do you think it would take for a dog to start pulling again if the target that caused him to pull in the first place became his only reinforcement? My guess is about four times for the average dog.

The target is a target and not a reinforcement...never the primary reinforcement.
The dog earns reinforcement for controlling his behavior while approaching the target.
He learns that each time he gets close without diving for the target he is reinforced. The by-product of this knowledge is that he will be able to better ignore just about everything going on around him.
This is another form of "attention.” When another dog runs up, when someone else calls a command, or when he sees a dumbbell thrown in the ring next door, ignore it and pay attention to the trainer, because only the trainer has reinforcement that is obtainable.