Magazine Banner Image

THE CARROT OR THE LIQUORICE? POSITIVE REINFORCEMENT

BY DR KERRY MACK

New research shows that rewarding good behaviour gets better results than punishing wrong conduct. Animal behaviour consultants agree that positive reinforcement should be the foundation of teaching and training.

Mayfield Farm’s horse of the month

Mayfield Farm produces quality young sport horses, and currently has a number for sale that are under saddle and ready for you to enjoy. Stud principal Kerry Mack offers the featured horse of the month, Mayfield Diamond Surprise.

Surprise is a beautiful three-year-old gelding with three quality paces and an outstanding canter. He is by dressage sensation Diamond Hit (Don Schufro/Ramino) who is the sire of Maree Tomkinson’s Diamantina, Hawkins Delicato (shown by Carl Hester), and more than 65 licensed stallions. His mother, Mayfield Be Brazen (Mayfield Pzazz x SPA Whitegril X Lauries Crusador), showed great promise for dressage but has proven to be a talented jumper; she was sold to an Olympic eventing home as a high-level eventing prospect. Her sire, Pzazz (by Jazz), was a wonderful Grand Prix dressage competitor with talent, trainability and a good jump as well.

Surprise shows a great forward attitude, and is a really willing worker looking for his forever home. Easy to float, shoe, catch and handle. He has started some easy gymnastic jumping exercises and shows a willing, careful attitude to this as well.

View the Mayfield Farm website for more young horses for sale.

www.mayfieldfarm.com.au

“Animal trainers are becoming
interested in the principle of
LIMA — Least Intrusive, Minimal
Aversive training.”

Scientists have been studying how animals learn for 130 years, and yet horse riders can be slow to use what has been learned. Pavlov did his famous dog experiments in the 1890s. Skinner formulated his theories about operant conditioning from the 1930s. Operant conditioning is learning behaviour by different types of reinforcement. Positive reinforcement is essentially rewarding the behaviour with something the trainer adds. For example, a food reward can be given when the horse demonstrates the behaviour you want to train. Most horse training is based on negative reinforcement where the trainer removes something when the behaviour is shown.

This is otherwise known as pressure and release training. An aid to trot is given with the leg and when the horse trots the pressure of the leg is immediately withdrawn. The aid to stop is given by pressure with the rein, and when the horse stops the pressure is immediately released. So the horse learns that legs mean go and rein means stop. The rider can increase the pressure with the leg by squeezing, using the spur, kicking and then add more pressure with a light touch of the whip. If the whip is used as punishment that is not pressure release training.

I think all horse people are familiar with the principles of pressure release training even though they may not recognise that this is operant conditioning, or negative reinforcement. Of course, the timing of the release is really important, so the horse knows which behaviour the release is referring to. It can be easy for an inexperienced trainer to reinforce unwanted behaviour inadvertently, for example by taking pressure off when a horse does the wrong thing, such as gets nippy or threatens to rear. The horse threatens the rider, the rider releases the rein and the horse learns a good way to avoid work.

Increasingly, animal trainers are becoming interested in the principle of “LIMA” — Least Intrusive, Minimal Aversive training. The Intenational Association of Animal Behaviour Consultants recommends that “positive reinforcement should be the first line of teaching, training and behaviour change programme considered. Positive reinforcement is associated with the lowest incidence of aggression, attention seeking, avoidance and fear in learners.” Recent research has shown that young horses trained with positive reinforcement remember the lesson better six months later and choose to be closer to humans, compared with horses trained conventionally. They showed a preference for their trainer over other humans.

So how can we use positive reinforcement in our horse training? Essentially this means that you give your horse something he likes when he does the thing you want. You want your horse to stand still while you saddle him up? You give him a food reward when he stands still. If he is really restless it may take a while to see a moment when his legs are still, but as soon as they are, you reward him. Give him something he likes. A piece of carrot or a small amount of pellets. You don’t push him around or smack him for moving, you reward him for standing still. You can use a clicker to mark the behaviour you want so he can understand more easily what he is being rewarded for.

For example, he is moving around in the cross ties, maybe pawing with restlessness. Then for a moment he is still, maybe something catches his eye and he stops to look at it. So immediately when his legs are still, you click and offer a treat. You might just say “good boy” and scratch him in that favourite place he has. Most often that will be a scratch on the wither. After a few repetitions of the clicker or of saying “good boy” and scratching him or feeding him he will come to learn that the clicker or “good boy” marks the behaviour that you are wanting.

It is very easy to train him to put his head down to be bridled by using positive reinforcement, or to stand still to be mounted, or to step forwards into the horse float. You can use positive reinforcement to get him to accept a medicine paste, or an injection, or to lift his legs up, or to be caught.

You should use the principle of shaping the behaviour. This means that you start by rewarding any attempt at the behaviour that you want. He stands still for a moment, immediately you reward. As he starts to understand, then you reward him for sustaining the behaviour. We use the clicker to mark the behaviour we want to shape. While he stands still, we click and reward. We may continue to repeat the click and reward while he continues to stand still. This is different from using food as a bribe. When you use food as a bribe or a distraction, it goes more like he is restless in the cross ties so you feed him to keep him still. When you run out of treats, he paws the ground and is restless, so you bribe him with more feed. In effect he is training you to get him treats. You are positively reinforcing the restlessness and pawing. It is better to wait until he is still and then reward. If he gets restless you stop feeding him. You can add a verbal cue, “stand still” for example, before he stands. It won’t take him long to recognise the verbal command if you are quick to reward him when he does. Giving him food at the end of a ride after you have gotten off and loosened the girth and done a few other things is not positive reinforcement of the work sessions, as he will not understand what it was you thinking you are rewarding because he won’t associate the reward with his behaviour in the ride.

“You don’t push him around
or smack him for moving, you
reward him for standing still.”

There are some principles about choosing positive reinforcers. Paralympian and qualified animal behaviour consultant Georgia Bruce points out: “The most important thing is that the positive reinforcer must be something he likes.” Of course, young foals don’t like pellets or carrots and have to learn to like scratching. Youngsters who have not had carrots and apples won’t find them reinforcing. They may be motivated by pellets or lucerne, or handfuls of fresh grass. But once your horse has developed a taste for carrots or liquorice, these can become high-value motivators that really help him want to do what you want him to. Personally, I don’t like to use sugar as a treat. If it’s bad for my teeth it must be bad for his too. If you are wanting to get him to do slow or easy things like stand still, you might want to use hay that takes him a while to chew. If you are wanting to repeat an exercise quickly in order to teach something, like walking onto the float or lifting his legs up, you might want to use something small of high value that he can chew quickly, like a few pellets or finely chopped carrot. That way he gets his reward, quickly chews and swallows it and is looking for the next reward, so may offer the behaviour. Quick repetitions of the behaviour will teach him the quickest.

If you are teaching him something new, you might have a range of treats that he likes and use them differentially. If he makes a bit of effort, he gets pellets, but if it’s really good he may get carrots. Mix it up to keep him motivated. If he seems to be losing interest you might offer his absolute favourite, e.g. liquorice. You can increase his motivation by planning a teaching session when he is hungry. He will be less motivated when his tummy is full.

Of course, some horses are more motivated by food than others. My Grand Prix horse Limelight has never really been a horse that is that interested in food. But he is interested in pleasing me and appears to love the clicker training. He is much more interested in food when we are using it in training than when it is in his feed bin. Kingsley, my young stallion, on the other hand just loves carrots and will do his best to earn them, even if he has just had his breakfast. He has just turned five but is learning the beginnings of piaffe with the clicker and positive reinforcement. He gets excited when he sees me pick up the stick, which I use to help cue the steps I want, as he knows that carrots will follow. He certainly is not afraid of the whip. I especially like to train the piaffe with positive reinforcement. This way the horse is motivated to want to do it and you can explain to him clearly what to do with his legs and carry weight, rather than the old-fashioned way of using tension (so-called “positive tension”) by telling him “go, don’t go” until he works it out.

“You can train your
horse to have good manners
around treats.”

If you are teaching him something and he makes a really good effort, you may want to “jackpot” his effort. This is telling him “really good”. To jackpot, give him several small portions of the reward quickly. Each portion is like saying “good, good, good”. This is more effective that giving him one big handful, as he won’t discriminate it as being better than just “good”.

Generally, it is good to make sure that the behaviour is connected to a specific cue. However, with something like piaffe you may be very happy for him to be self-motivated and not so reliant on you saying “now and now”. Rather you want to shape it so you say “now” and he keeps going until you say “enough”.

Some people are reluctant to use food treats as they say it encourages horses to nip or bite and be pushy. Good training will ensure that this doesn’t happen. We certainly don’t reward nipping and pushiness. You can train your horse to have good manners around treats, for example, by starting by teaching him to turn his head away from you to receive the reward. If he gets pushy you should say “no” and move him out of your space before you repeat the exercise you are training. You may have to stop the lesson you are working on and go back to basic manners. However, researchers have demonstrated that horses learn best with food treats. Scratching in a favourite spot will not strongly reinforce behaviour even in socialised horses (Keison and Abrahams, 2020).

Researchers have shown that horses trained with positive reinforcement remembered their lessons better after an extended break, and remembered their trainers as well (Leste-Lassere, 2010).

It can be tricky using positive reinforcement under saddle. You can scratch the withers easily. You can have a helper with treats or your pocket full. You can tape the clicker on to the top of your whip so as to not have too much in your hand. I have had the experience of a horse eager for treats develop the habit of a sudden and unbalanced halt when he heard the clicker, and sometimes if I say “good boy”. This unbalanced halt became a problem that I then had to fix. You may want to ride forwards for a moment after the behaviour you want before you click. Don’t click every time. For example, if you want to use positive reinforcement to help train the flying changes, you don’t want him to do the change and then stop.

Have fun playing with the positive reward.

(I am grateful to Georgia Bruce for her assistance in preparing this article.) EQ

YOU MIGHT ALSO LIKE TO READ:

Submission or stress? Something to chew on (By Dr Kerry Mack – August 2020 Issue)

A relaxed horse is a happy horse (By Dr Kerry Mack – July 2020 Issue)

The literate horse rider (By Dr Kerry Mack – June 2020 Issue)

Why horses love Ingrid Klimke (By Dr Kerry Mack – May 2020 Issue)