Tuesday, November 22, 2011

Sit to Say "Please"

Sit seems such a basic behavior, that we often take it for granted. In reality, sit is the very beginning of self-control. Dogs that control their own impulses are a joy to live with, unlike those dogs that were never taught and it becomes the owner's responsibility to maintain control 24-7. Self-control of a dog's natural impulses doesn't come naturally.

When your dog wants a share of your lunch, its his natural inclination to jump on you and try to snatch a bite. When your dog wants your attention, its her natural impulse to jump up, trying to get to your face, probably putting dirty paws on you in the process. When your dog wants to go outside, he thinks its perfectly acceptable to scratch at the door and woodwork until that door opens. When your dog wants inside, he'll do the same thing again. And when you arrive at your destination, your dog sees no reason to wait before making a mad dash to get out of the car, leaving scratches on your leg as she goes. All of these are natural and normal behaviors for a dog; however, they are not much fun to live with.

You can teach your dog to sit on cue quite easily, using capturing or luring. More important is to make sit a "required" behavior to get whatever your dog wants. I don't cue the sit (either verbally or with a hand signal). Instead I stand there and wait for the dog to figure out each time how to get something, so she thinks its always her idea and thus is very motivated to sit.

When I picked up my puppy, we had a long flight ahead of us, so I ate a light lunch of crackers and cheese before the long drive to the airport. My puppy repeatedly stood up with her paws on my leg begging for her share. I pretended to ignore her, but in reality 90% of my attention was on her. After several minutes of begging, she happened to lose her balance and fall back into a sit. I instantly popped a piece of cheese in her mouth and she ran off to enjoy her treasure. Minutes later she was back, wanting more. Again, her paws came up on my leg and I waited. In less than 30 seconds she sat again, this time tentatively, but on purpose. Again cheese appeared in her mouth and thus was born, the default sit.

Before we left on the first flight of our trip, Gimme had discovered that sit worked to get her all kinds of things besides cheese. She learned to sit when she wanted attention, when she wanted to be picked up, when she wanted me to pet her, and when she wanted strangers to pet her. By sitting instead of doing what came naturally, Gimme was already learning self-control and was on the way to becoming a pleasure to live with.

In the first couple of weeks, she learned to offer a sit to go outside, come inside, get treats, get meals, get out of a crate, get out of the car, have a ball/toy thrown, initiate training - really everything. Over time, I adjusted the length of time that she had to remain in a sit, to increase her self control. Then when I thought she was ready to learn a formal stay, it was very easy. It only took only one gentle reminder for her to understand what I wanted, because she'd already been practicing an easier version in her self control exercises.

Gimme here: I have to tell you, humans have a real weakness for a dog that sits and looks pretty. I can get my person to do almost anything when I sit and cock my head to one side. I think it’s the best thing I ever taught her. Its been very useful.

I have to warn you that there are times to ask for that default sit and times not to. For instance, it would be foolish to wait for a default sit first thing in the morning when your puppy or young dog probably has an urgent need to potty.

See separate entries on how to teach your dog to sit on cue and about other impulse control exercises to teach your dog.

The Basic Sit

When we talk of "teaching the dog to sit" we really mean that we are teaching the dog to sit when we want them to sit. Every dog knows how to sit. There are two basic ways to teach a sit – capturing and luring. You can also free shape a sit, but I usually reserve that for special cases, such as retired racing greyhounds which have often been punished for sitting.

Capturing a Sit
This is very easy to do and my favorite way to teach sit; its simple and has a special side benefit. Get some treats and your clicker, sit yourself down and watch your dog. At some point the dog will sit on its own, you click, and immediately follow with a treat. Chances are the moment you click the dog will stand up, especially if she already knows what click means. That's okay, just hand over that treat and go back to waiting. Every time you see your dog sit, click and treat. Very soon your dog will realize that sitting pays. Then your dog will start sitting on purpose to make you click and treat.

          This is a very special moment. Your dog has just made the leap
          from casually happening to sit, to consciously choosing to sit.
          Your dog has just become "operant". An operant dog realizes
          they have the ability to make good things happen. Operant dogs
          are more confident than they would be otherwise and they
          participate in training as a willing teammate. Non-operant dogs
          (including all dogs trained with methods that feature punishment
          as standard fare) don't have the understanding that they can
          make good stuff happen.  Non-operant dogs  participate in
          training because training is something that is done "to" them
          and to avoid bad stuff.

When you are willing to bet $10 that your dog is going to sit right now, you can start adding the cue. See the section on adding a cue below.

Luring a Sit
Many people teach their dog to sit by luring and it has one distinct advantage – the luring motion to get the sit behavior, becomes the hand signal for the sit. The drawbacks are that you miss the opportunity for your dog to become operant, you risk becoming dependent on luring and you have to work through moving your hand away from the dog's nose.  Also, the things you have to do with your hands (to avoid becoming lure dependent), while not awkward, can be confusing.

The first step is to hold a treat in your left hand and the clicker in your right.  You can certainly reverse your hands (treat in right hand and clicker in left hand), but you'll have to reverse these directions.  Hold the treat just in front of your dog's nose, then gradually move the treat slightly up and to the back of the dog's head. Gradual is important, since you want the dog to follow the treat with his nose. As his nose tips up and back, his rump will go down into a sit or near sit. Click that sit and give the treat. Repeat this luring motion with treat in your hand and click/treat – for a total of five times.

Troubleshooting:  If your dog lifts its front feet up, you are raising your hand too much. If he doesn't follow the treat with his nose, you are probably going too fast.  If your dog seems especially resistant to sitting, you may have to break this into pieces and start by clicking for him lowering his rump and then ask for more and more in each repetition until the rump hits the ground.

Now its time to get the food out of your luring (left) hand. Hold the treat and clicker in your right hand and lure with the left hand that just had food half a minute before and still smells like food. When the dog sits, click. Then very quickly and smoothly use your right hand to deliver the treat into your left hand, without the left hand moving away from the dog, and then give the treat to the dog from the luring left hand. Repeat this process – for a total of five times.

Now you can repeat this again using an empty luring hand, holding clicker and treats in the right hand. Do the luring motion, as the dog sits, click and then use the left hand to take the treat from the right and give it to the dog. Repeat 5-10 times.

Troubleshooting:  If your dog tries to follow the luring hand as it moves away to get treats from the othe rhand, try moving more quickly.  If that doesn't work, your right hand may have to meet the left hand halfway to start with.

If your dog is reliably and smoothly sitting when you make the luring motion, its time to begin fading the proximity of the hand signal by moving it gradually away from the dog. Instead of having your fingers an inch from the dog's nose, try two inches, then four. When you can cue the sit with a hand signal from two feet away, its time to add a verbal cue. Hand signals are nice and they certainly look cool, but it does require your dog to be looking at you.

Adding a verbal cue
Adding verbal cues is a process that follows the same pattern every time. The more behaviors you add cues to, the quicker your dog will learn verbal cues each time.  To begin with, you will do each step 15 – 30 times, depending on how quickly your dog demonstrates understanding. Later you can do less repetitions per step. For an experienced dog, you may only do 5 repetitions per step. For a very experienced dog, it may be only 5 repetitions altogether. Remember to click/treat for every repetition.
  1. Say "sit" as your dog is sitting, 15 – 30 repetitions
  2. Say "sit" as your dog is starting to sit, 15 – 30 repetitions
  3. Say "sit" just before your dog starts to sit, 15 – 30 repetitions
  4. Test by saying "sit" and then wait to see if the dog sits. If he sits, the cue is attached, if he doesn't sit, repeat step 3 and then retest.
Adding a new verbal cue to a behavior on a hand signal
If you are using a hand signal, your dog already knows a cue for sit, so you are going to attach a new verbal cue to it. Remember to click/treat for every repetition.
  1. Say "sit" just before you give the hand signal – about a quarter of a second before – repeat 15 times.
  2. Say "sit" about one half second before giving the hand signal – repeat 15 times.
  3. Say "sit" about one second before giving the hand signal – repeat 15 times.
  4. Gradually increase the time between giving the new verbal cue and the old hand signal in one second increments, repeating 10 – 15 times each time.
  5. When your dog starts to sit after hearing the verbal cue and before you start the hand signal, then you can test his understanding. Say the verbal cue and wait. If the dog sits, click/treat. If the dog doesn't sit, continue the process following up with the hand signal for another 30 repetitions and then test again.
Gimme here: I was part of a demonstration to show how to attach a special cue to my go to mat behavior. When my person says "mat" I run to my bed and lay down.  The instructor wanted to show how to teach a dog to go to mat when the doorbell rings. So they made the doorbell sound, my person said "mat" and I got clicked and treated for doing what I was supposed to do. By the fifth time, when the doorbell rang, I just ran to my bed and laid down. I already knew what it meant. That's because I'm very smart. I'm just saying…

Friday, November 4, 2011

Reward Value

When you use positive reward based training, the quality and type of reward you use is very important. Several factors influence the value of the rewards you use, such as:
  1. How much does the dog like the reward you are offering? Is it desirable enough to overcome distraction? Is it motivating enough for harder tasks?
  2. How often does the dog get that reward?
  3. What schedule of reinforcement are you on, fixed or random?
  4. What size is the reward? How many does the dog get at one time?
  5. How are rewards delivered?
While we talk here about food rewards, you also want to consider using play, touch, and other physical rewards.

Gimme here: I love food, all kinds of food. My person has a veritable treasure trove of yummy things for me to eat. And sometimes, she surprises me and pulls a toy out to play with. I never know what I'm going to get for my efforts, but I do know its always something good. I have my person very well trained. I'm just saying…

How much does the dog like the reward you are offering? Is it desirable enough to overcome distraction? Is it motivating enough for harder tasks?

When you train in a new place, around distraction or if the training is harder than usual, you need better than average rewards. Maybe your dog is tired, while its not the best time to train, sometimes that is all you have. Make sure you have better than average rewards available.

When my boss called me and wanted me to leave my dinner to come to the office to solve his computer problem or help him find things – he knew to make it worth my while. My boss understood how to motivate me to a special effort. Our dogs are much the same, though fortunately for us, their wants are much simpler.

How often does the dog get that reward?

You need to know what your dog's opinions are about the treats you use. Even if you totally love chocolate, if you get it in several forms in each and every meal and for every snack and drink – after awhile even chocolate will lose its allure. For Gimme, the A-number-one-never-fail reward is peanut butter. I protect its value by saving it for those times when I want a special effort.

Quality of reward - On an ever improving scale of delectability there's kibl, commercial treats, hot dog pieces, French fries, pizza bones, homemade real dog cookies, popcorn, refrigerator jerky, garlic steak, braised chicken hearts, teriyaki turkey, peanut butter... you get the idea. Since I don't ever use commercial stuff, hot dog is the lowest value treat I use. Remember, the scale of delectability may be very different for your dog.

When I'm training I have a mix of great stuff and regular stuff, so my dog doesn't know what she'll get. Your dog should think that you can always pull another incredible delectable edible out of thin air... Your dog should be thinking all the time about what he can do to make you deliver those incredible delectable edibles.

What schedule of reinforcement are you on, fixed or random?

You always want to move to a variable reward schedule, since that creates the strongest behaviors. Varying rewards is multi-faceted. When we speak of "schedule" we mean how often the dog gets rewarded.

Think of it like this: A soda machine is a fixed schedule and you only put money in the machine when you are thirsty. You never put money in a trash can, because you never get anything out - its on a fixed schedule of zero.  Las Vegas slot machines are a variable schedule and the tycoons that run them know exactly how often a person has to get a payoff to keep them playing. People will play their paycheck into those machines month after month with an occasional small payoff, because they are addicted to the possibility of a huge payoff.

Does your dog get a reward every time they perform, every other or every third time, or on a random schedule. If its not every time, consider giving greater reward to keep the dog's interest up. Variable schedule rewarding creates strong persistent behaviors. That's what you want.

What size is the reward? How many does the dog get at one time?

Does your dog get a tiny treat for minimal effort? Does he get a bigger treat when he does especially well? Does he sometimes get a huge treat? How many treats does he get? One for average effort or more for greater effort? A jackpot for a stellar effort? All of these are ways to influence your dog's response to the reward you are using.

How are rewards delivered?

We usually just hand a treat to our dog. Maybe we toss the treat to the side to reset the dog for the next repetition. Other than that, we can be pretty boring about how we deliver the goods.

There are many ways to increase the value of rewards by using our delivery to make the reward seem more exciting.
  • If you are giving more than one treat or a jackpot, deliver them one at a time, counting out loud with a bit of flourish thrown in. Three treats delivered this way will have more impact than 6 to 10 treats gobbled out of your hand all together.
  • Run to where the treats are. Sometimes my dog does something great in the yard when I'm out of treats and running with me into the house to get goodies really makes her eyes shine with excitement. After a good agility run, my dog got a small reward right outside the ring and then we ran to the car for the big payoff.
  • Let your dog get the treats "gobble style". I often train with popcorn (use rice krispies tossed with jerky dust for tiny dogs), which I keep in a big bowl. I also have a paper bag that I put a couple handfuls of popcorn in. If my dog does especially well, she gets to stick her head in the bag and grab all she can get in one good gobble. That's a great motivator.
  • A variation on the gobble is to let my dog stick his face in the treat pouch, while I provide resistance pulling up on the strap (envision a horse's feedbag).
  • Another variation of the paper bag gobble is to put some treats in a paper bag (a recycling idea for all those used fast food sacks), wad it up and let your dog tear it up to get at the treats.
  • When I make refrigerator jerky, most of the treats are bite sized and I stop drying while they are still somewhat moist. But, I also make some that are long strips of meat, which I over-dry, making them really tough. My dogs get extra reward value by playing tug with me over these jerky sticks... tearing off pieces and coming back for more, until the piece is so short it pulls out of my hand.