Stale Cheerios » Training Tips

Are You Having Fun Yet? (video)

Mary Hunter — Mon, 11 Jan 2010 18:38:55 +0000

Horse training (or dog training, parrot training, fish training, etc.) should be fun. Let me repeat, training an animal should be fun and enjoyable for both you and the animal! So, are you having fun yet? And is your horse or dog having the same amount of fun?

Many times we say an animal (or person) should do something because he ought to or because we think he should. The horse who disobeys and the dog who doesn’t listen are both being disrespectful and bad. Same goes for the child who won’t sit still in school.

But think about it this way, what’s in it for them to do as you say? If the answer is not a whole lot, you can bet the animal isn’t going to be so keen on listening or following directions.

Check out this video below. Many doctors and health experts are forever hounding on their patients to get more exercise. A simple way to get a bit more exercise is to take the stairs instead of walking. However, humans, like horses, are often masters of conservation of energy and prefer elevators and escalators. If we can make stairs fun, people are more likely to use them.

Happy people on piano stairs @ Yahoo! Video

So, here’s what I want to know

What are some simple ways you make your training fun? Do you have any special games or activities that your horse, dog or other pet just loves? What do you do when you have trouble motivating the animal?

Where are the magical musical stairs in your training?

I look forward to hearing your comments!

And a big thanks to Cheryl, of Painting Horse, who originally posted this video .

Teaching Your Dog (or Horse) to Think

Mary Hunter — Fri, 08 Jan 2010 13:28:05 +0000

I’ve been watching Bob Bailey’s Fundamentals of Animal Training DVD. I’ve watched the first disc, and am really enjoying the set so far. Bob Bailey is not only an excellent trainer; he has a gift for being able to communicate ideas clearly and concisely. I plan to write a proper review after I watch the whole set.

For now, here’s a quote that I really liked that is from the first disc.

“Now I see so often on the internet that you want to create the thinking animal. What have I already said? The animal is already thinking! The animal is not thinking about what you want it to think about–that’s the only problem. It’s up to you to communicate clearly to the animal what you want.”

I think there’s a lot of truth to this. I hear lots of clicker trainers (and I know I’ve been guilty of it before as well) talking about teaching animals to think. It’s common for clicker trainers and positive trainers to promote their methods by saying that they teach the animal to think. Instead, I think what we often mean is that good training is about teamwork and clarity of communication between animal and trainer. It’s also about giving the animal the opportunities to make choices and figure out solutions on her own.

What are your thoughts on teaching an animal to think?

Stale Cheerios: The Best of 2009

Mary Hunter — Thu, 31 Dec 2009 12:11:42 +0000

These are some of my favorite posts from 2009. I’ve picked a few of the ones I enjoyed writing the most, as well as some of your favorites–posts that got the most views and the most comments. Let me know if I’ve left any of your favorites off the list! If you are a newer reader and missed any of these posts, be sure to check some of them out. They are full of fun stories and great information about animal training and clicker training. (It was hard to pick just 10!)

1. Aggressive Training Methods and Aggressive Dogs
2. An Introduction to Reinforcement
3. Building Duration with 300 Peck Pigeons
4. Fish Training Progress
5. ORCA Great Minds Part 1: Steve Martin
6. In a Whisper or In a Shout? Training Under Time Constraints
7. Teaching the Parelli Friendly Game with Clicker Training
8. Waterhole Ritual 1: Sharing Territory
9. What is Stimulus Control?
10. What Makes Something Scary?

Mountain Coatis and Training Mechanics

Mary Hunter — Fri, 25 Sep 2009 22:49:46 +0000

I played just a bit this afternoon with the Mountain Coati at the Heard Museum. Coatis, which are close relatives of the raccoon, are found in South America and some parts of South Texas. This coati was a gregarious, agreeable sort of fellow, similar to most raccoons that I’ve met. We mainly worked on a bit of targeting with him, having him touch a ball on the end of a stick in exchange for a piece of apple or grape. He understood the targeting well, so it became a fun exercise of moving him up and down and around his cage as he eagerly followed the target stick. (Read more about target training.)

My main issues with working with the coati were mechanical in nature. I was working one-handed as I currently have my left hand in a cast. This meant as soon as he targeted the stick, I had to set it down so I could deliver food with the same hand. However, he was pretty forgiving of my lack of speed. A larger problem was actual food delivery. Coatis have really long noses. (They remind me almost of an anteater’s snout.) As well, the upper jaw extends quite further than the bottom. Since I was feeding him through the cage, I had to hold my tidbit of food just right so that he could grasp it with his teeth. I was a bit clumsy with this at the beginning and he kept dropping the food.

I was feeding cut up pieces of grapes and apples as reinforcers. After a few of each, it was clear that he much preferred the grapes as he would grab those much faster. After a few more pieces of apple the picky little fellow started refusing them entirely! This made me wonder a bit if the earlier dropped pieces of apple could have been somewhat on purpose…

The coati was super interested in me and willing to continue playing the targeting game even though my mechanics were less than stellar. However, with many animals, especially those that are shy, hesitant, newer to training or less eager to offer behavior, this big of a breakdown in mechanics could have been too much for the animal. A less eager animal might have become bored, distracted or disinterested by my slow food delivery or lack of food delivery (when pieces were dropped or when the coati refused the apples).

At it’s essence, training is a mechanical skill. When you train it’s always important to consider mechanics. If you are keeping a high rate of reinforcement and your animal still seems bored, disinterested or is leaving the session, consider if your mechanics could be at fault. Think about these questions:

Is the animal getting the reward every time?
Could the timing of your food delivery be improved?
Can the timing of your click or marker signal be improved?
Are you using a good enough reinforcer?
What are you doing that is preventing or interrupting the flow of the training session?
(Click here for suggestions to improve your mechanical skills.)

Bad mechanics often become apparent when the animal leaves or becomes frustrated. However, even if you think your training is going well, it still could benefit from improved mechanics. I find certain styles of clickers are easier to click than others. With some designs, I find the stiffness of the button means I click too slowly. (And then I potentially reinforce the wrong behavior!) I can improve my timing just by changing the type of clicker I use. With the coati, even though he was targeting well, I had plenty of room for improvement So, even when you think training is going well, it’s beneficial to pause every once in awhile and assess whether improving your mechanical skills could make a good training session into a great training session.

Are You a Splitter or a Lumper?

Mary Hunter — Wed, 23 Sep 2009 03:25:10 +0000

Horse clicker trainer Alexandra Kurland often speaks of splitters and lumpers. These are funny words, but they refer to an often serious training problem!

Most behaviors can be broken down into many smaller pieces and approximations. When we break our goal down into tiny chunks and build gradually to a target behavior, we’re being a splitter. When we lump everything together and try to teach the behavior in large chunks, we’re being lumpers. Training difficulties often arise when we do too much lumping and not enough splitting.

When we lump, we ask for too much too soon. The animal often becomes confused or frustrated and the trainer is liable to become frustrated if the animal does not respond correctly.

When we split the behavior into tiny chunks, we can build the behavior gradually, always setting the animal up for success. Since each new thing we ask is similar to other things we have taught, the animal is able to understand what we ask. Being a good splitter means you know how to use the principles of shaping to get to your target behavior. Here are two examples from the horse world. Can you think of ways you’ve been a splitter or a lumper recently?

Standing Tied for Grooming

Standing tied for grooming, hoof care, saddling and other procedures is important for any young horse. However, if we go to fast and teach too much at once, it’s easy to create a horse that pulls back and won’t stand tied. Don’t teach tying and grooming at the same time to a horse that’s skeptical about being touched, as then the horse has too many things to deal with. Break it down, get the horse use to human contact and being touched, then work on tying, the work on grooming untied. Once the horse is comfortable being groomed and comfortable being tied, it’s much easier to combine the two together!

Trailer Loading.

A trailer is a big dark box on wheels. The horse has to step up and walk across a strange surface that sounds and feels pretty weird, into a dark, confined space with tight walls and a low ceiling. Pretty scary for a prey animal! If we break this obstacle down into tiny components, it becomes easier for the horse to be comfortable with trailer.

First, we need a horse who will follow reliably where we lead and understands cues for moving forward, stopping and moving backwards. Then, we can begin simulating the various aspects of the trailer that the horse must be comfortable with such as:
Walking over strange surfaces: tarps, boards, bridges, rain jackets.
Going through narrow spaces: narrow gates, between 2 barrels, wash stalls and grooming stalls, horse stocks.
Steping up onto objects: pedestals, flatbed trailers, bridges, raised platforms, solid wooden pallets.

It’s also helpful practicing having the horse walk ahead of us in these simulations to help build confidence and practicing having the horse go backwards through these obstacles. Once the horse is comfortable, it’s easy to start combining some of these to more closely approximate the trailer. If we start with components of the behavior, we’ll have a lot easier time when we actually get to the trailer.

So, are you more of a splitter or a lumper? Think about this the next time you go to train!

Animals in captivity

Mary Hunter — Tue, 30 Jun 2009 19:17:57 +0000

Zoos, aquariums, wildlife parks, sea world, nature centers, personal pets, we encounter animals around us all the time. Many of the wild animals we encounter today (especially in zoos and aquariums) are bred in captivity, rather than collected from the wild.

Animals have a host of needs that must be met in captivity. If the needs aren’t met, animals often start exhibiting stereotypies and other abnormal behaviors. I’ve blogged before about what animals need for a good mental life. I think animals can be happy and healthy in captivity, especially when the owner, trainer or facility is willing to consider the animal’s needs before focusing on their profit, ego or reputation.

I recently ran across a great passage in Karen Pryor’s book, Lads Before the Wind.

Perhaps as important, though, the keeping and display of porpoises in oceanariums has helped to awakened the public to the value of these animals. Conservation begins with understanding, and understanding can begin with personal contact: a child in the audience catching a ball that a porpoise threw, a governor or senator stroking Makua’s ample belly. I feel sure that no “sportsman” who saw one of our shows ever went out again and put a rifle bullet through a porpoise at sea just for fun. National conservation efforts by an informed public have recently brought all cetaceans in U.S. waters under federal protection, so that today you need a permit and a good reason to go out and catch a porpoise. U.S. whaling has stopped, and the importation of whale products is now illegal, a first step to halting the slaughter on a worldwide basis.

This passage was written in the mid-70s, when oceanariums and marine animal parks were a relatively new phenomenon; however, the words still ring true today.

When we come into contact with wild animals, we observe their beauty, strength and intelligence. We are more likely to value the animals and support conservation efforts if we can begin to understand the species.

Using Shaping to Teach New Behaviors

Mary Hunter — Thu, 25 Jun 2009 18:13:03 +0000

Shaping is a powerful way to build behavior. Animals who understand shaping truly understand that they can control their environment and that their behavior earns them rewards.

This is different from training with coercive methods (punishment, negative reinforcement) where the animal learns to perform or offer behaviors only to avoid the correction or get rid of the pressure.

Many positive reinforcement trainers and click trainers become very dependent on targeting, luring and guiding the animal. While these are good methods for establishing behaviors, these methods do not create the same kind of operant animal behavior as pure shaping (often called “free shaping). Too much emphasis on targeting can create an animal who is overly dependent on the trainer for direction and guidance.

Check out this video for an awesome, short example of teaching a dog to close a cabinet door using shaping. She first rewards the dog for approaching the door, then touching the door, then bumping the door with it’s nose.

Stimulus Control and the Do Nots

Mary Hunter — Tue, 23 Jun 2009 18:08:42 +0000

As I’ve talked about before, cues are powerful only if we can get them under stimulus control. (What is stimulus control?) The animal must be able to distinguish between a variety of different cues, know which behavior goes with which cue and know not to perform the behaviors unless the cue is given.

I’ve been reading Karen Pryor’s book recounting her experiences training dolphins, which includes a long description of these concepts. The first cue she taught the park’s spinner dolphins (stenella attenuata) was a cue to spin. Then, she taught a second behavior (porpoising) and introduced a cue for it. But, whenever she gave the cue for porpoising, the dolphins would spin. It took careful training to teach the animals to associate each behavior with a separate cue.

The animals had learned to spin on cue, and not to spin of cue. Now they had to learn to porpoise on cue and not to porpoise off cue, and also not to spin on the porpoising cue and not to porpoise on the spin cue. Training the do nots was fully as important as training the do’s. I have seen many a dog and horse trainer overlook this fundamental fact. You do not have full control over the behavior just because you can order it up when you want it, you must also make sure it is no longer offered spontaneously when you did not ask for it. The army sergeant whose platoon will advance under fire when he tells them to is still in a bad way if they may sometimes advance when he didn’t tell them to. In fact a primary function of drill teams and marching exercises and so on is not only to bring specific behavior under stimulus control, but also to establish a pattern, a habit, a skill, even, of responding on cue and not responding in the absence of the cue.

Porpoises can generalize (so can many, many animals). By the time our group had learned to spin reliably on the spin cue and to porpoise on the porpoising cue, they had also learned that “cues mean do something,” “no cue means do nothing,” and “different cues mean do different things.”

It took days and days to teach the spinners the first cue, and many session to get the second cue down pat. The third cue, a sound for the lobtailing behavior, was pretty well established in a morning. The animals had become “sophisticated” about cues; from now on it became much easier to teach those individuals not only underwater sound cues but hand signals or any other kind of command.

A good trainer helps an animal learn behaviors and cues. A great trainer really helps an animal learn how to learn. Once animals learn the rules of the game for earning reinforcement, training can progress much more rapidly. However, this requires the trainer to know the animal, have a plan, and be able to set the animal up for success.

Some behaviors can be much harder to put on cue than others. Are there any particular behaviors that you’ve found have been hard to put under stimulus control when training an animal?

The Four Quadrants of Training

Mary Hunter — Thu, 18 Jun 2009 19:07:14 +0000

Have some reading time? Here’s a great and thorough look at the four quadrants of training (positive reinforcement, negative reinforcement, positive punishment, negative punishment) written by Katie Bartlett, an equine clicker trainer. Many of her examples deal with horses, but the theory and philosophy applies to much more than just horses.
The article can be found here.

If you’re new to clicker training and operant conditioning, you might want to start with this shorter article I wrote recently discussing the differences between positive and negative reinforcement.

Training Tips: Improving Cues with Limited Holds

Mary Hunter — Tue, 16 Jun 2009 18:18:01 +0000

How do we increase response time to our cues and requests? Say we ask our dog to sit. The dog sits, but she takes her time and waits several seconds before sitting. We can reduce the time interval between when the cue is given and when the animal responds by using what is called a limited hold. Basically, this means that once an animal knows a behavior, we shape the response time to be smaller by only reinforcing response times that are less than a certain value.

Karen Pryor explains this well in Lads Before the Wind:

Ron’s manual showed us one more refinement: the limited hold. This was a method for producing instant rather than dilatory response. At first, when we turned on the spin cue, some animals didn’t get around to spinning until the cue had been on for ten or fifteen seconds and the other animals had spun and were eating their fish already. So we started with the average time–about fifteen seconds–that it took for all the animals to spin on cue, and we turned the cue on for just that long. If Moki, say, was pokey, and didn’t spin in fifteen seconds, the cue went off and he was out of luck.

A stopwatch helped the trainer stay honest; it was always tempting to leave the cue on just a little longer when you saw that your dilatory animal was about to jump. If you succumbed to that temptation, however, you could end up with the animals training you to leave the cue on longer and longer.

Animals have a good sense of time. Soon every animal began to hustle a little. Some jumped before the fifteen seconds were up. Then we shortened the cue-on time to twelve seconds. Again, the laggards were out of luck. Again they had to learn to hustle. We found that we could tighten up this “limited hold” to the smallest length of time in which the animals could physically accomplish the behavior. When our shows opened, the spinning in Whaler’s Cove was on a three-second limited hold. The trainer punched the button, and bingo! six animals dove out of sight and then flew into the air all over the place. It was dramatic.

How long do your animals take to respond when you give cues? The behavior we get is the behavior we’ve trained. So, if our animals are slow and sluggish to respond, we are either not providing enough reinforcement to make the job worthwhile or we are being lax in our standards and reinforcing slow responses (or both). By tightening up the response time, training sessions move at a faster pace and have a better flow. The animal is more responsive and learns to perform the behavior promptly when cued. Even if you don’t want to pull out the stopwatch to measure response time, it’s good to consider how quickly we expect the animal to respond to known cues and whether or not this interval can be improved.

Read more here about how to teach and use cues.