As an aside, if you ever wonder what it's like to take a lesson from me, this is it. I talk incessantly, don't give my riders a single moment to understand what I've said, and I do laugh at my own jokes. Sorry, not sorry.
Link here if the below video doesn't work.
TC's Background
Every rider has that one horse, the one that makes you question everything you thought you knew about riding horses. My current horse, TC, is that horse for me. He’s been an absolute blast to develop and he’s in a great place currently, but he was a very tricky young horse.
I began riding him in late 2014 after earning my Gold medal on my retired Grand Prix horse, Rico. I needed something fun to ride and TC was bred by a family friend who I reached out to in search of something fun to ride for free while I finished up graduate school. I had client horses to ride, but I enjoy having my own horse very much, so TC would be a surrogate until I found a horse to buy after I entered the workforce.
Baby TC was a blast and I loved his personality and his work ethic. Sooo I bought him in the fall of 2015. And he promptly climbed up on his panel fence at his new barn, fell off of it, and ripped his suspensory out of the origin site. He went to a rehab facility until he was trotting for a good amount of time and then came back into my life.
Rehabilitating a young, powerful horse is tough. TC had to stay in a small paddock so he wouldn’t run around, he couldn’t be turned out, and he had to stick with a small amount of work each day as he continued to heal. This led to a lot of undesirable behavior, the most undesirable of which was bolting.
Baby TC had shown a desire to bolt, but his gangly body didn’t let him do anything really aggressive. Five year old TC was a different animal, add into it that he was cooped up and unable to let off any steam. There was no longeing him down or turning him out before riding. And winter was rapidly approaching too.
I began clicker training him on the recommendation from a good friend who also uses clicker training in her program. I found that clicker training was incredibly helpful to get the bolting to stop. I also used it to reduce his spooking. As we went into winter, I realized that I had no way of exercising him effectively without riding him, he’s never taken to side reins very well, so I clicker trained him to reach for contact on the long lines. I moved on to teaching him to engage his hindquarters and lift his back while long lining.
I’m using clicker training currently to continue to improve the connection, to put the piaffe on, and to reinstall the medium walk because all he wants to do is piaffe now.
 |
| Baby TC being skeptical about sprinklers |
The basis is conditioning and there are two types of conditioning, both of which are utilized in clicker training.
Classical Conditioning
Pavlovian conditioning, or classical conditioning, involves the pairing of a primary reinforcer with a secondary reinforcer. Pavlov actually was studying dog’s salivary glands. He fell into classical conditioning when he noticed that his dogs started to salivate not when they were given food, but prior to then when they knew they were given food. He called this psychic secretion and wanted to know why it happened. And that’s where the bell came in to the operation. Essentially he would ring the bell and then feed the dogs, them eventually associating the bell with the food, and would salivate to the bell.
Now, dogs aren’t born thinking that bells sound delicious, but with enough pairing of the bell, a secondary reinforcer, to the food, a primary reinforcer, the dog then responds to the bell as they would respond to food, by salivating.
To relate this back to clicker training, take the bell and replace it with a clicker, take the dogs and replace it with your horse, and take the food and replace it with a treat or some grain. Eventually the clicker means the same thing to the horse that the treat does. And the clicker is a much easier reinforcer for the rider to use to pinpoint the exact moment the horse does something correctly rather than trying to quickly run over and shove cookies in the horse’s mouth.
 |
| Not spooky at this lol |
The other kind of conditioning is operant conditioning, which we all use every day with our horses. Operant conditioning was fleshed out by B.F. Skinner. Essentially operant conditioning is when a behavior is shaped by what follows it, but it’s more complicated than that. It’s a four part process, but I’m going to start with three parts because it’s easier to understand this way.
The first part is the antecedent, but for our sport, let’s just change that to aid. The aid is a stimulus that indicates that you want the horse to perform a behavior, which is the second part. The behavior itself. Following the behavior, as we all know from growing up, we have consequences. Consequences can either reinforce a behavior or punish it. Reinforcing a behavior causes the behavior to occur more frequently and/or with greater strength. Punishing a behavior causes the behavior to occur less frequently and/or with a weaker strength.
There are four types of consequences: positive reinforcement, negative reinforcement, positive punishment, and negative punishment. They can get confusing because the words positive and negative don’t refer to good and bad, they refer to whether you added or removed a stimulus. Reinforcement is anything you do that increases the frequency and/or strength of a behavior and punishment is anything you do that decreases the frequency and/or strength of a behavior.
So positive reinforcement is the addition of an appetitive stimulus after a behavior has been performed that increases the frequency or strength of the behavior. Think of a horse finally nailing a clean flying change and getting a walk break and scratches on the wither. The addition of an appetitive stimulus, the scratches and break, after a behavior, the clean flying change, that you want to encourage to occur again.
Negative reinforcement sounds bad or weird, but we see it everywhere as well. This is the removal (negative) of an aversive stimulus after a behavior has been performed that increases the frequency or strength of the behavior. Think of a rider who puts their spurs in their horse’s side, then when the horse moves forward (desirable behavior), the rider takes the spurs away from the horse’s side, the removal of an aversive stimulus that increases the likelihood that the horse will move forward.
Positive punishment is another weird one, but it’s what you think of when you think of punishment. The addition (positive) of an aversive stimulus to decrease the likelihood of a behavior. For example, when TC crowds my space, I tap him with a whip on his chest. This is the addition of an aversive stimulus that will hopefully reduce his desire to crowd me. I can tell you that hasn’t worked all that well.
Negative punishment is one we dont typically use with horses unless we’re eight years old and through tears tell our pony that just bucked us off that he is NOT getting any carrots later. It is the removal of an appetitive stimulus after a behavior that discourages the behavior. We use it rather infrequently with horses only because it’s tough to set it up where it actually discourages a specific behavior. TC isn’t going to think that I’m not giving him a cookie before I leave the barn because I was unhappy with something he did earlier in the day, horses just don’t think that way.
Conditioning for Clicker Training
Now that we’re through conditioning, hopefully you can see how clicker training combines the two types of conditioning. We use classical conditioning to make the clicker an appetitive stimulus that you can then use as a consequence of a behavior in operant conditioning to make the behavior more likely to occur.
To clicker train an animal, we first have to acknowledge what primary reinforcers are effective reinforcers. To classify a primary reinforcer, think of what can reinforce an animal without any conditioning history, or learning history. For example, I do my job in exchange for money, a reinforcer. But I had to be taught to value money, because children raised in a society without money will not see any inherent value in having it. Children raised in a society without food will… well they’ll die. That got dark. But children are born wanting food and water and shelter. So are horses, they aren’t born wanting praise or clicks, but they are born wanting this list, among other things. One thing to note on this list is the wither scratches. There have been studies that show that scratching a horse’s wither lowers their heart rate and therefore relaxes them, and that is also a primary reinforcer. I’ve used wither scratches as a reinforcer for horses I ride who I don’t use the clicker on, it’s a great reinforcer.
 |
| Lots of exposure for this weirdo |
Bridge Training
Our first task involves classical conditioning and it’s called bridge training. We are essentially creating a bridge between the clicker and the primary reinforcer we are using. I chose to use TC’s grain because I wanted to not worry about how many cookies he was eating in a day and I also don’t use clicker training every day, so didn’t want to have his diet changed so randomly. I can adjust the amount of grain he gets after his ride based on how much clicker training we did during the ride.
One of the worries people have about clicker training is that it will cause the horse to mug you for food or to start biting, but the bridge training I used with TC helps eliminate that response while also bridging the clicker to the primary reinforcement of grain.
I started with TC in his paddock and he saw that I had grain and was very interested in me. I didn’t let him come after the grain, just waited for him to get bored and realize I wasn’t going to give it to him. He was fairly rude, so it took some time. Eventually he heard a horse whinny and looked away. And that’s when I clicked and gave him a small handful of grain. He was suddenly really interested in me again, mugging me for more grain and getting in my space. I remained patient and he got bored again, looked away and then I clicked and gave him grain again. We did about five minutes a day for a few days and soon he had figured out that he only gets grain when he looks away, so he would see me coming with the clicker and my bag of grain and would not mug me, but rather look away because he knew that was how he got the grain.
By teaching it this way, I’ve taught him to not be mouthy or mug me for grain but rather to give me space and wait for the click. [at this point I showed this video of Shawna using this bridging technique]
At this point I was finished with the classical conditioning part of the training, TC knew that a click meant grain and was happy to work for clicks because he knew grain would follow.
 |
| Remember when he used to rear himself calm at the trailer? |
Eliminating TC's Bolting Behavior
Now, a lot of people start target training at this point, but I had a very specific goal in mind, so I didn’t go that route. I’ll eventually target train TC for fun, but I moved straight from the bridge training to addressing TC’s bolting behavior.
I started to then begin to solidify the whoa aid to address TC’s undesirable bolting behavior. If you remember from Kate’s talk a few weeks ago, she went over Andrew McLean’s work on behavioral issues. He has an entire chapter on bolting and his main point was that bolting, at its root, is caused by the breakdown of the whoa aid leading to the horse no longer being under stimulus control, which is a fancy way of saying that I couldn’t stop him. Basically, the reinforcement he got from bolting, such as getting away from perceived danger, was way more effective than the reinforcement he got from not bolting, which at the time was just to not annoy me.
So my goal was to increase the reinforcement history of the whoa aid so that it could override the reinforcement history that he had from bolting. Bolting is very self reinforcing, if the horse bolts and doesn’t get killed by the thing they were afraid of, no matter whether the thing actually posed a threat, the horse feels reinforced by the lack of death and that strengthens the bolting response. But TC likes food, sometimes more than he likes avoiding dangerous situations, so I began to reinforce the whoa aid.
I worked with him on the ground first, using my aid, saying whoa and taking back on his halter, to produce a behavior, a halt, that I then reinforced with a click and some grain. When he stopped, he was not allowed to immediately climb in my lap for the grain, the whoa needed to be appropriate. At first, he didn’t get it easily, but after doing it every day for a while, he started to slam to a halt more aggressively, even offering it when I even touched his rope or when I said the word “no” because the whoa meant grain so he may as well do it as much as he could.
From there, I schooled it on the longe line, getting it to the point where if I just shifted my weight a little or touched the line a little, he’d stop. Then I took it undersaddle, which I’ll talk about later.
 |
| Tarp de-spooking unnecessary |
Reducing TC's Spooking Behavior
TC’s bolting behavior was getting way better with the clicker training. I noticed two things about the bolting. First, I noticed that when he bolted, it was far less aggressive and easier for me to stop. He didn’t go nearly as fast and didn’t go nearly as far before I could stop him. Second, I noticed that the frequency went down, he didn’t bolt as often as he used to. For a long time, it seemed that he bolted out of the blue, without any warning. At this point it started to only happen if something scared him rather than just randomly.
His spooking was a secondary issue, but when I noticed that his spookiness was what was keeping the bolting alive and well, I began clicker training for curiosity and relaxation.
TC’s spooking has always been interesting to me. I took him to two bombproofing clinics and he loved them, wasn’t spooky at all. You can thwap him with plastic bags and pool noodles, ask him to walk across mattresses and water, he’s fine with all of that. He was even fine with the cats running across the front of him, dogs barking and running at him, strollers and bicycles passing him on the trail, and the resident chicken taking a dust bath in the arena. If there’s a clear reason for what is going on, he’s fine with it. But if the wind picks something up and it flys around, or if something has changes at the barn, or if a wheelbarrow is sitting by itself vs being handled by a barn worker, he is terrified of it.
Because TC’s spooking was not just a startle and would always lead to him putting more space between him and the thing he was afraid of, which was naturally reinforcing to him because getting away from a scary object meant that the object would not attack him, I began to address it by trying to override that reinforcement history with clicker training. Now, he could choose between two reinforcers: the reinforcement that he got from running away from the object vs the reinforcement that I was going to provide to him for taking a step closer to the object.
I started out by leading him around my barn waiting for him to become weary about something. I found a tarp that he didn’t like and made him stand still near it. Then I coaxed him forward and he took one step, so I clicked and gave him a treat. We worked on it until he was standing on the tarp and then we moved on. I did this technique for a while, until every time he saw something that made him nervous, he would approach it rather than run away from it.
Once I got him to drag me toward things that previously scared him, I noticed that he still would stand stock still with his heart beating really hard, so I began to click for relaxation. It wasn’t good enough to stand near the object, he had to relax. It took a while at first, but soon he took a deep breath and I clicked for that. I noticed a huge decrease in his spooking once I started only clicking for both approaching the spooky object AND relaxing near it. He can still be spooky, he still hates wind, but he’s way easier to console now and definitely the spooking has both decreased in frequency and strength. He’ll startle or if it’s really bad, scoot, but gone are the moments where he’d get truly out of control.
 |
| Happy to hang out in this tunnel |
Improving the Connection on the Long Lines
I mentioned using it for long lining, and here’s the progression for that. TC has a very friesian neck, his tendency is to go the way he’s going in this first picture, with his neck retracted and his hind legs stepping short. He didn’t like long lining at first because he can be rather claustrophobic, so I had to clicker train him to reach to contact. I would aid him to bend a little more but most of this beginning work was a waiting game. He soon blew his nose, leading him to reach down a bit, so I clicked that. Then I got a lot of voluntary nose blowing, I clicked at first but soon waited until he stopped blowing but kept a bit of stretch to click. Slowly I shaped the behavior until he reached down all the time.
I loved this because it meant that I could work him effectively in hand without having to ride him, very helpful when he would act up in the winter or when I was in a rush or when he stamped on my foot and I couldn’t ride him. I started clicking for all sorts of things, coming through with the whip meant he would flex his lumbosacral joint and step further under for a few steps and I clicked for that. I continued our work with the whoa aid there, all kinds of things. Just shaping his behavior on the long lines so I could continue developing him when I couldn’t ride him or just didn’t want to.
Clicker Training Undersaddle
As I mentioned earlier, I’ve used it extensively undersaddle as well with TC. I started doing it with the bolting, reinforcing the whoa aid from my seat, rein, and voice. I then moved on to the spooking undersaddle, encouraging him to walk up to things. Then even did some work on reaching for contact, reinforcing any weight he put in the reins.
Clicker training undersaddle can be tough to coordinate as a rider. I will often grab small cookies instead of a bag of grain because it’s just easier to carry around, as the grain sometimes finds its way out of the bag. Holding the clicker is tricky too, as is coordinating the click. But once the coordination is there, it’s incredibly useful. What it does for TC is provide him with a discriminative stimulus, something I wanted to throw in here at some point.
A discriminative stimulus is an indicator to an individual that a reinforcer is available. Think of it this way, you’re walking down the street and you’re hungry, you see a restaurant, that restaurant is reinforcing to you because it means you’re about to get food, it’s the bell the dogs salivated to. But then you see that the restaurant is closed and it becomes far less reinforcing, you’re still hungry but food is not available to you. But if you see the sign is flipped to open, the restaurant immediately becomes reinforcing. The open sign is a signal to you that the reinforcer, food, is available.
Think of a horse who acts up in the show ring because he’s figured out that his rider doesn’t have as much ability to correct him. The show ring becomes a discriminative stimulus for bad behavior, the horse can quote get away with bad behavior in there.
What I noticed when I would grab the clicker and ride with it is that TC would be more likely to offer me things that he’s been reinforced for with clicks. Me holding onto the clicker became a discriminative stimulus. So even though I hadn’t done much work undersaddle with the clicker, he saw that I had it and knew that clicks were available, so he worked harder for them. I didn’t have to click very often undersaddle to keep him working hard for clicks.
We’re nearing the end so I wanted to share two videos of things I’m doing with clicker training right now. Here is a video of me teaching TC to piaffe with the clicker.
 |
| Hilarious, definitely doing more bombproofing clinics in the future with future horses and with TC just for funzies |
Changing my Philosophy
Lastly, I wanted to talk about how this has changed my training philosophy. I’ve ridden with trainers in the past who emphasized that when a horse is doing something bad, they are meaning to be bad. There’s a lot of language used that just isn’t correct: my horse is an asshole, he knows what I want but he just doesn’t want to do it, he’s not actually afraid of that thing he just wants to get out of working.
Clicker training for me put everything in a different perspective. TC didn’t bolt to be an asshole, TC bolted because bolting had a stronger reinforcement history than not bolting. TC didn’t spook because he was trying to get out of work, TC spooked because spooking had a stronger reinforcement history than working. TC didn’t refuse to stretch to contact because he didn’t feel like doing it, he refused to stretch to contact because there was no reinforcement history for it.
With clicker training, TC has become an incredibly easy horse to train for me, he’s only been doing the piaffe for two months and it’s already a higher quality than my Grand Prix horse’s. He’s not got the conformation to reach to contact easily but he does it becuase he’s been reinforced for doing it.
And I notice a difference in him too. He is very playful undersaddle, excited to work every day. He knows that I’m going to be fair to him, that if he gets an answer wrong, my first thought will never be that he is being deliberately bad.
The End!
I've been meaning to write this post for a long time and I know this is a bit different than my normal posts because I literally just posted my webinar in words. But it was so fun! Make sure you check out the Facebook page to find out when the other webinars are happening and look at the playlist of all the ones we've done so far on youtube.
It was a good talk!
ReplyDeleteThank you, thank you, thank you. I have been so intensely curious for so long about the details of doing clicker training with a horse. I tried years and years ago with Grif, but I struggled. (It was so much easier with my hyper-food-motivated dogs.)
ReplyDeleteAs you shared success after success with clicker training TC, I wondered more about how to do it, but was still a bit hestiant/stumped. So I was REALLY excited about this presentation. After listening, I feel like I'm going to be able to accomplish this with my horses.