equine clicker training

using precision and positive reinforcement to teach horses and people

Food for Thought

These are two short articles that I wrote in 2005. I’m not sure I would say everything exactly the same way today (2021), but I do still love to watch horses learn and think there’s a lot of power in allowing them to have some less structured sessions. The second article was prompted by a new clicker trainer who seemed stuck because she felt she had to maintain a high rate of reinforcement and her horse had a limited repertoire. Sessions were not as relaxed and thoughtful as she would like.

On Babies and Horses…or… the Power of Free Shaping and Creativity

I was sitting in my living room the other day with my youngest daughter just having a quiet moment and thinking about how much fun babies are at her age. She just turned 5 months old and is working really hard on controlling her hands. For those of you who aren’t familiar with the developmental stages of babies, they really blossom between 4 and 6 months. They learn to control their heads much better and can start to sit supported on the floor. Even more important, they start to really get control of their hands. They go from wild flailing and just random waving of arms to really concentrating on how to move their hands. Sometimes I catch Geneva just studying her hands very intently.  Two weeks ago, if I held up a toy, she would very carefully work her hand (one) over toward the toy,  and if she was successful, she would grab it. Now she is starting to do that with both hands at the same time. It takes a lot of concentration and I can see her determination in her face.  But, interestingly enough, at this age, she doesn’t get frustrated. If she doesn’t get it, she either stops trying and does something else or she continues until I help her.

While watching her the other day, I realized that there are a lot of similarities between my fascination with babies and my fascination with clicker training. I love to watch horses think.  I love to watch them try to figure things out.  I truly believe that if you give them a creative task, they go beyond just working for the carrot and actually start to enjoy the task.  This is easy to see with my yearling and the other young horses. But I can also see it with Willy, who is 19 and still loves to try new things.  I used to play with the horses a lot in the evening after the kids went to bed and since it was dark, I did a lot of tricks and free shaping in either the aisle or the stall. This space limitation meant we had to be pretty creative about coming up with new things and I did a lot of little games.  Sometimes I would just put a new object out and see what they did.  The great thing about free shaping is that it is just like watching my baby learn. I can sit back and watch the horse thinking. If I keep the reinforcement rate high enough, the horse will continue to work out the problem. If he or she gets frustrated, I can step in and help or change the game a little. In addition to teaching my horses a lot of new tricks, I learned a lot about the horses as individuals. I now know who is more bold in their offered behaviors and who needs encouragement. I think you can learn a lot from taking the time to just sit and watch a horse work through a puzzle.

Earlier this year, I attended a clicker clinic where we played a game that Alex called the “creativity game.” She had gotten the idea from another clicker trainer (Kathy Sdao, I think..) and the idea was to set up a training session where you encouraged the trainee to offer new behaviors.  This was done by using the following rules. You could only click a behavior once and you had to click any new behavior.    There was some discretion involved if the trainee got into repetitive patterns. So, for example, if the trainee touched an object, click, touched the same one, no click, touched a new one, click and so on. Except if they really got in a rut and just touched everything, you could withhold the click to get them to offer a new behavior.  The trainer would start the session with no goal behavior in mind, but just watch the trainee interact with the environment. It was really a chance to see what the trainee might do.  We played it with people and it was fascinating. I had expected the trainee to get frustrated at not being clicked for repeating behaviors, but the rate of reinforcement was so high that that frustration never developed. She said later that as long as she was getting clicked, she was ok.  I also found it very interesting that the trainee really was driving the game. In our game, the trainee decided that we wanted her to make hot chocolate. Since each step was a different behavior, she was clicked along the way, even though that chain of behaviors was not what we had in mind. 

This whole session made me think that we sometimes limit ourselves and what we do with our horses, because we just don’t have ideas for things to try. I know that I have periods where I spend a lot of time doing “serious” stuff with my horses and some of this is because I am working through something or getting ready for a clinic. Other times, it is just because I can’t think of anything new to do with my horses. So now I feel that it’s ok if I don’t have a new toy, or a plan. I can still go out and play with the horses and see what they can come up with if I let them do what they want with all my regular stuff.  It would be a nice break from our usual training routine where I am directing the training.

I also think that if you don’t spend much time free shaping, your horse will be more hesitant to offer behaviors. I can’t say this definitively, but I do know that my horses seem to become braver about looking for and offering possible answers when I have spent recent time on free shaping. I love offered behaviors. For me, they are one of the great pluses of clicker training. With some behaviors, I use the horse offering the behavior as both a sign that we are ready to move on and add a level of complication, and also as a way for me to make my job easier. I have been teaching Willy how to do haunches-in from the ground. He found this really confusing as he was willing to move his haunches toward me, but only in the wrong bend. When I asked him to move them toward me while keeping the same bend, he would get confused. We struggled with this for a long time, until I got him to the point where he just knew the right answer was to move his haunches toward me and offered it very freely. Then I could start to adjust him. Before he got to this point, he would back off as soon as I said “well, not quite that.” I had to have that piece really solid before I could move on. I had worried that he would consider the bend part of the haunches moving, but he seemed to just learn that stepping his hind end over was the right answer and we could adjust things from there.

In addition to being a nice exercise for the horse, I think free shaping is a great way to introduce novice clicker trainers to clicker training. If you pick a task that has no “wrong” answer, you will find that as the trainer, you don’t get trapped into feeling frustrated if your horse is doing something wrong. I always recommend that new clicker trainers just spend some time teaching their horse’s simple tricks where it really doesn’t matter if the horse gets it or not. You can use it to practice your own timing and problem solving skills and might even teach your horse some fun party tricks to impress visitors. 

So, for those of you who haven’t done much free shaping, give it a try. You might be surprised at how much fun both you and your horse can have. 

Katie, August 15, 2004

You Get What You Click: Looking at the bigger picture.

This was originally posted to clickryder in response to a question about how to decide what to click and what to work on. The person who posed the question had a mare who was a quick learner but also impatient.

Hi Jill,

I think I will stick by my earlier suggestions of working on head down and mat work because these are two exercises that are both calming and good for working on duration, and that sounds like what Zuni needs.

When I am training a horse I try to be very aware of their mental state and choose exercises that will help them maintain an appropriate level of interest and eagerness without getting too excited. The easiest way to do this is to alternate exercises that get the horse excited with calming ones. This is similar to the good strategy of training both sides of each behavior (if you train head down, train head up too), but I think of it more as training exercises that help you balance energy.

In addition, you can help balance your horse’s energy by adding duration to many exercises which can promote calmness, so it doesn’t mean you are stuck forever with alternating between head down and other things.

In your specific case, it sounds like you have given her a really good start and she is catching on and eager to play. That is great, now you just have to get to the next stage which is where she is eager to play but understands that she has many opportunities for reward and she doesn’t have to be so excited about things.

I have done a lot of thinking this fall about clicking while riding and how to take advantage of the power of the clicker without getting myself in trouble (which I have done, but luckily clicked my way out of it – phew!).

One of the things I realized is that when I click a horse, I am clicking for one of three things: a change in response to a cue, duration, or an offered behavior. This is oversimplified, but you will get the idea. How much I click for each of these (compared to the others) affects the kind of horse I will end up riding. There is no right answer here. It is a matter of personal preference, and might vary from horse to horse, even for the same rider/trainer. With my two riding horses, one is higher energy than the other so I choose different exercises for him and I also structure my lesson so that I am clicking for different reasons.

I am wondering if this way of looking at how to choose clickable moments would be helpful to others, especially new clicker trainers, so I will throw my ideas out here and you can use them if they are useful…

I think that when you start out clicker training, it seems so simple. You click for what you like and build from there. And that works, but once you are beyond basic work (and with some horses, even within the basic work), there are layers of complexity and it is important to be aware that you are influencing your horse’s behavior not just by what you click at each individual moment, but by what aspect of the behaviors you are working on and the general pattern of things you click.

So, for example, if you click a lot for offered behavior, you will get a very eager horse that does lots of wonderful things, but maybe not on cue. Or maybe your horse will end up using visual aids for cues (such as a concrete object (mat, ball, cone..) . This is great if you want to do a lot of free shaping or tricks or you want to have a creative horse. I have to confess that I love offered behaviors and I think while most of us might say that we want behaviors on cue, the line between asking for something and having it offered is not always clear. And most of us have default.html behaviors, so even though we might think most of our training is under stimulus control, there are still parts that are not.

If you click a lot for changes, you will end up with a horse that has a large repertoire but is quick to move on to trying something new, and you will probably find it difficult to get duration. I think everyone goes through this stage because you have to get any new behavior started before you go for duration. But I think that if you get too caught up in the details of improving the quality and clicking for every improved effort without adding a little duration, the horse doesn’t get the message that sometimes they just need to do the right thing a little longer. You can use clicking for changes to build duration but only if you don’t click every effort.

If you click for duration, you will get a steadier horse but one that might be less quick to catch on when you want to add the next level of refinement to the current behavior. They might also have spent more time practicing the behavior in a less polished form. But we need duration. If you don’t have duration, then you don’t see enough variation to know what your horse can do. One of my common errors is to stay with clicking each and every effort. I might make the horse do it a little longer before clicking because I am waiting for a specific detail, but I don’t tend to stop the horse and restart. I guess this is more true for some types of exercises than others but I think it is important to praise the horse for trying and ask again before the horse gets in the habit of being rewarded for each effort. You don’t want the horse to get frustrated, but you do want the horse to become persistent about continuing to try.

This leads to one way to build duration which is to just decide not to click every effort. So you might click randomly for head down until the horse starts to show that she is thinking a little more. I find that when I start head down, the horse is in such a yo-yo phase that sometimes there is no significant difference between tries. So, as soon as I am reasonably sure the horse will offer some form of the behavior, I start clicking less often. Sometimes I just randomly omit a click here and there and see how the horse responds. I say randomly although I am, of course, watching the horse’s body language and trying to click the better efforts, but I am not worried about it. If I miss a click, I just ask again and sometimes I let the horse put her head back up and ask again so that I can show her that head down is still the right answer. I also like to click if I can see the horse change her mind by starting to put her head up and then dropping it again. In this case, I don’t wait for duration on that effort, but click her for thinking and decided to put her head back down without me asking again.

Another way to teach duration is to just extend the amount of time/steps before you click. Or you can click and allow the horse to eat while continuing the movement. So, you mentioned clicking Zuni and feeding her low. This is a great way to build duration in head down because the horse figures out there is really no point in putting her head up.

Well, I think I am rambling a bit here and it is getting late. I hope this makes some sense. I guess what I am suggesting is that you work on getting Zuni to do head down for longer periods and use head down to have her settle when she gets excited about other exercises. I think that some of the initial excitement is often just because the horses can’t believe we are playing this great new game and they chill out a bit once they realize we are going to keep playing. But, in the meantime, if you are working on targeting and she gets grabby or upset, just have her put her head down for a minute.

…I would add to this that I didn’t include captured behavior in my list of possible things to click.  I tend to think of capturing a behavior as capturing a complete behavior instead of the beginning piece. For example, I did capture the behavior of lying down with my mini. I clicked for the final laydown instead of the steps leading up to it. My mini was not offering to lie down because he thought I wanted it. He was just lying down and when I clicked it, he thought “wow, what did I just do?” In that sense it is different from an offered behavior. But since I don’t find myself capturing complete behaviors very often I didn’t include them in my list of reasons to click.

Katie Bartlett,  February 1, 2005