I don't have a platform for marking heeling position, but I did find a scrap carpet, folded it in half and used that has a mat. I would stand by it and have Mickey in heel position using the mat. (This concept explained more in the previous blog entry.)
With my low vision, I think a platform will be much better as the contrast and distinction will be much easier to see. But since I have little time to find a platform, I'm winging it for now. I will try to hit the yard/garage sales to find maybe an aerobic platform, step exercise equipment that no one wants anymore. think it is large enough for Mickey. Oh well, no time today to look for one, make one or pay a lot of money for one, I'll just keep training. Figure it all out with what I have for now.
However, in the meantime, I'm still sh aping my way. I call what I'm doing as dabbling. Because I'm really not sure what my criteria should be sometimes. How do I shape for this or what do I need to shape? So I just start shaping to see where it takes me. Trial and Error through this maze of precision training.
Last night I saw that Mickey was understanding the correct place for heeling. I was careful not to feed from my body and have him lean over to get the treat. This swings his rump out. SO I'm learning that all treats in heel position must come straight down over his head. This is not only true for his sitting heel position, but for his fronts as well.
I had recently went to two trainers, both top competitor, one having been to the National Obedience Invitational and has gotten an OTCh (Obedience Title Champion) and the other having gotten an OTCh with scores of 200. Both instructed me to use a lure for heeling. Both slightly different.
I then thought I would try putting my left hand on my abdomen. Since I had been luring his heel, he was very focused on my hand. One second after I placed my hand on the abdomen while he was in heel position, he scooted up to make sure he could see my hand, waiting and anticipating when I would say "Yes!" My hand to Mickey is a target, it brings the life of a treat to him. He has followed that hand many times and knows to follow it.
I had to think about changing his perception since he cannot scoot forward to look at my hand in the obedience ring. I would loose points. This also tells me that Mickey's world is following that hand, not understanding his position in the heel. As so many trainers know, the dog follows the hand with the treat, not necessarily knowing that you are really treating for a position. This is where shaping does come in. Luring is an excellent tool, but we need to know when to get rid of it and when we can start shaping behaviors where the dog is not so focused on our hands when working so close to us. I use luring constantly in service dog work to place him in a tight area under a table, or chair, or under a seat on an bus, airplane etc. This direction has also taught him where to lie down when out in public as sometimes i need him on my left side, right side, behind me or in front of me. The hand is a target, lure, and cue for working Service Dog work.
The fascination I have with shaping is trying to set my criteria. Sometimes I don't know what that is until I dabble and shape to see what I need and want. Competition obedience is such precision training, and this higher level of precision is new to me, so I need to play around and dabble. Then I get set in what I need. So how do I keep Mickey in place in his heelp position, while I have my hand, that he has been highly conditioned to follow? This is the beauty of clicker training is that you can play around with luring or shaping until you get what you need. Again, I'm so fortunate to have a dog that can handle the changes as I go through this maze of figuring out what is the best way to shape obedience behaviors specifically for Mickey. Each dog is different.
I put Mickey in heel position at my left side. I then placed my hand on my abdomen. Clicked and then the key point, where I delivered the treat. If I treated him anywhere, that allowed him to get out of position, this is defeating my purpose. So just like my fronts, I took the treat, raised it above his head so he is still in perfect position, and treated. His body must be straight with mine. No rump sticking in or out. No legs sticking way forward. So many "little" things to think about and yes, when you are shaping, one criteria at a time. So his rump will be reasonably in place as his feet. I will later shape specifically for feet and rump to be perfect straight.
After a few repetitions of clicking with my left hand on my abdomen, I wanted to do the big test. Moving forward. This also was a challenge because Mickey wants to keep his eye on that hand. When I move forward to heel, he will adjust himself forward with rump out to the self side. Next placement to fix. He's out of heel position. I did it with one step. The old rule is break down the criteria if the dog is not doing the placement or the behavior you want. So I did a half step. EUREKA! PERFECT position YIPPIE! treat and of course TREAT in appropriate position.
This morning I was in my long t-shirt nightie, and bare feet. I was waiting for my coffee and thought I would do a 2 minute dabble. For the dog, I changed the environment. I was not in shoes, I was not in pants. How does the dog adjust themselves for a heel? By my shoes? Pants? but...they are gone right now. His heel wasnt' perfect, but a little more shaping and he got it. I then jumped around, which is usually a no-no in training because you don't want to confuse your dog, but I put him in the front and wanted to get him into the finished (heeling at my self side) position. When I put my left hand down, he went behind me and then pushed on my hand wise, WHERE'S THE TREAT! Hmmm, what have I been teaching Mickey? This can't happen in the ring, what do I do? How do I problem solve to change this? Again, that hand is so conditioned from luring and following that treat in the hand, he can't be focused on it in the ring. So shaping to the rescue. I left my hand dangle. I shaped him a few times in heel sitting position, but fed with my right hand. To keep his good position so he wouldn't start focusing on what my right hand was doing, I fed above his head, bringing my hand down and giving him the treat. I also succeed in getting the focus off the left hand. The old rule of treat for position, even after you marked or clicked. A few repetitions of this and he lost interest in my left hand. Now let's try the finish again. A few repetitions and he lost interest in the hand, allowing it to dangle. Beautiful! Exercise Finished!
In the obedience competition ring, I can put my hand on my abdomen during a walking heel. But when I do a come, front, finish exercise, my hand must be down at its side. All these rules to remember. Can I do it? Will I be able to do a 20 year dream of wanting to enter ti obedience ring and didn't because I despised corrections, my first method of learning obedience and then my perception of precision training was obscure. Now I see the beauty of the journey in precision training. Mickey is right there with me, waiting for the next thing to do. A willing dog. He's so willing that when I'm sitting at my computer, he literally will come up to me, bop my elbow like a true Belgian, and want to train. Am I anthropomorphizing? That he really wants to train? He had dinner already! Is this true, this boring obedience stuff, and he loves it? I'm starting to, after 20 years loving the journey of figuring out these micro behaviors. What a fun journey! I can't wait to get into the ring!
All this precision training is really allowing me to break down the behaviors carefully. Look at each aspect and environment in how the dog sees the world and shape for that tiny micro moment, position and behavior.
Saturday, April 16, 2011
Sunday, April 10, 2011
Heeling
Heeling is when a dog walks on your left side with their shoulder in line with your pant seam. The task is to keep in that position the whole time when heeling in the competition obedience ring. The dog will need to stay in this position when turning left, right, about turn, slowing down, walking faster, walking in a figure eight pattern, etc. How well the dog walks with you in position is rated and scored. This one exercise of a few done in the ring.
There are many ways to teach this task. Traditionally was the snap of a choke chain to get the dog to heel in perfect fashion with you. The science behind this method was the dog is avoiding the aversion of the snap of the choke chain to stay in line to heel with you. Some people today still train with this method.
Then there is luring. The dog follows a treat in your hand. The treat is periodically given, but the dog is basically following the treat in the hand while walking on your left side. Eventually the hand will have to be removed since this is not how you can do heel work in the competition obedience ring. I have been trying some luring and find that when I put my left hand on my abdomen, my dog wants to focus on my hand. This type of focus is moving him forward where is not in line of my pant seem. He is slightly forward from position. I find that my dog focus more on what my hand is doing and what it is telling him, he is cuing off the hand, rather than staying in place. This is mostly because I have in the past used my hands like a target to move him where I need him to go. For some dogs this luring can work well, but due to how I use my hands in luring for direction, I will need to use another method.
Some people have used target sticks to keep dogs in heel position. I have not experimented with this method, but there are some people who have had some have success with a target stick.
There is also the platform or mat where the dog is stationed on a mat or platform to shape position. Once this is solid, the trainer can remove the platforms and get the dog in position without the platform. This has been successfully done with shaping. Then you will go to the next criteria to shaping the heeling while moving forward.Clicking at one step, then increasing the criteria to two steps and depending how well the dog is succeeding with this, will determine the future increments of how many steps the handler takes before clicking and treating.
I'm going to experiment with the latter. I'm using a folded over piece of carpet. Been shaping Mickey to stand on it. Of course he is thinking his previous shaped behaviors using a mat, which is to lie down on it. He would also put his foot on the carpet and move from it fast and look from me, so I started to feed him on the mad. Treat for position is the old rule. Which helped him slow down and stay on the mat for awhile. I didn't want to lure so he wouldn't be focused on my hand and not think that he is actually standing on a carpet. I noticed when I did lure, it was following the hand, not paying attention to where his feet were being placed. SO back to shaping.
I'm too impatient and since Mickey knows a loose heeling, I started walking forward to see if he will stay in place. Then, I realized, ok, I'm really not ready for this. I'm too impatient and trying to jump ahead because I know he isn't a beginner, but he's not quiet doing what I need. Communication break down. Poor dog, what do you expect when all of a sudden you are trying a different approach.
I need to work on the mat placements a bit more so I can give this approach enough time do it right. With Mickey doing loose lead walking with me in public as a service dog for 5 years, it is almost like we are going backwards to try this different approach. But the foundation for this method. When I jumped my criteria too fast, I realized Mickey was not understanding what I wanted and I quickly reverted to other methods I have used. Talk about confusing! To do this method justice, I need to slow down, think of what I am doing and think it through to get that picture in my head of what is needed. Just like when I started out with the fronts, this will take a little time to dabble and find my comfort zone that works for me. Mickey is so patient with these different approaches. What a trooper! I'm so fortunate to have this dog that is willing to allow me to dabble and experiment.
My goal and focus will always be using positive reinforcement. I will not use any corrective means as that is not the purpose of why I am going into the obedience ring. It is to show it can be done by all positive reinforcement. It is a lot of dabbling, but since there are only a handful of people who do positive obedience, sometimes you are on your own. A pioneer that has to sweat figuring out the maze, but in the end, it will work out and before long, Mickey will be showing in the ring, happy and eager to see what is next in the exercise.
There are many ways to teach this task. Traditionally was the snap of a choke chain to get the dog to heel in perfect fashion with you. The science behind this method was the dog is avoiding the aversion of the snap of the choke chain to stay in line to heel with you. Some people today still train with this method.
Then there is luring. The dog follows a treat in your hand. The treat is periodically given, but the dog is basically following the treat in the hand while walking on your left side. Eventually the hand will have to be removed since this is not how you can do heel work in the competition obedience ring. I have been trying some luring and find that when I put my left hand on my abdomen, my dog wants to focus on my hand. This type of focus is moving him forward where is not in line of my pant seem. He is slightly forward from position. I find that my dog focus more on what my hand is doing and what it is telling him, he is cuing off the hand, rather than staying in place. This is mostly because I have in the past used my hands like a target to move him where I need him to go. For some dogs this luring can work well, but due to how I use my hands in luring for direction, I will need to use another method.
Some people have used target sticks to keep dogs in heel position. I have not experimented with this method, but there are some people who have had some have success with a target stick.
There is also the platform or mat where the dog is stationed on a mat or platform to shape position. Once this is solid, the trainer can remove the platforms and get the dog in position without the platform. This has been successfully done with shaping. Then you will go to the next criteria to shaping the heeling while moving forward.Clicking at one step, then increasing the criteria to two steps and depending how well the dog is succeeding with this, will determine the future increments of how many steps the handler takes before clicking and treating.
I'm going to experiment with the latter. I'm using a folded over piece of carpet. Been shaping Mickey to stand on it. Of course he is thinking his previous shaped behaviors using a mat, which is to lie down on it. He would also put his foot on the carpet and move from it fast and look from me, so I started to feed him on the mad. Treat for position is the old rule. Which helped him slow down and stay on the mat for awhile. I didn't want to lure so he wouldn't be focused on my hand and not think that he is actually standing on a carpet. I noticed when I did lure, it was following the hand, not paying attention to where his feet were being placed. SO back to shaping.
I'm too impatient and since Mickey knows a loose heeling, I started walking forward to see if he will stay in place. Then, I realized, ok, I'm really not ready for this. I'm too impatient and trying to jump ahead because I know he isn't a beginner, but he's not quiet doing what I need. Communication break down. Poor dog, what do you expect when all of a sudden you are trying a different approach.
I need to work on the mat placements a bit more so I can give this approach enough time do it right. With Mickey doing loose lead walking with me in public as a service dog for 5 years, it is almost like we are going backwards to try this different approach. But the foundation for this method. When I jumped my criteria too fast, I realized Mickey was not understanding what I wanted and I quickly reverted to other methods I have used. Talk about confusing! To do this method justice, I need to slow down, think of what I am doing and think it through to get that picture in my head of what is needed. Just like when I started out with the fronts, this will take a little time to dabble and find my comfort zone that works for me. Mickey is so patient with these different approaches. What a trooper! I'm so fortunate to have this dog that is willing to allow me to dabble and experiment.
My goal and focus will always be using positive reinforcement. I will not use any corrective means as that is not the purpose of why I am going into the obedience ring. It is to show it can be done by all positive reinforcement. It is a lot of dabbling, but since there are only a handful of people who do positive obedience, sometimes you are on your own. A pioneer that has to sweat figuring out the maze, but in the end, it will work out and before long, Mickey will be showing in the ring, happy and eager to see what is next in the exercise.
Sunday, April 3, 2011
Setting Criteria
The three principles of training I follow. 1. rate of reinforcements, 2. timing and 3. criteria. These principles were pounded in my head from Bob and Marian Bailey when I attended their workshops. These principles have never failed me. Knowing them well has also made me a much better trainer. Call me narrow minded or stuck on a particular way of training, but it has been the best I've ever used. To deviate from this, my training skills do get sloppy.
My rate of reinforcements and timing come with practice, but criteria takes a lot of energy of my brain. The thought process of training. What do I want my dog to do? What do I choose to shape and have my dog do to set them up for the next step to get where I want? You can't expect to have what you want and get there in one step. It may take 100 steps to get there. Building slowly. The talent comes in knowing what step to teach next to make a strong foundation. Like a good teacher in mathematics. You don't all of a sudden start doing complex equations in two days, it takes several classes to get to higher level of equations and problem solving.
Each step I teach is a criteria. The rule of thumb is get 80% proficiency before moving to the next increment step or criteria. Always have your criteria plan ready as your subject may go fast and if you are not ready, training stops and you loose that fabulous momentum. Take the time to write out and plan your criteria. Or if you are better at doing it in your head, you can do it that way as well, but writing it out makes for good training records. Then in a year or two, you can see what you did and the written plan can jog your memory what path you took. I have had countless experiences where five or ten years later I think, how did I teach my previous dog how to do this behavior?
Recently I have been doing fronts. This is when the dog comes straight in to you and must sit just in front of you, not touching, but be within reach of you. Their sit must be straight in front of you. No hip shifted on either side. So what would be the criteria when training this?
At this point, the recall is set. The next step is back chaining. You go backwards. So the last thing the dog does in the behavior chain is what you reinforce, thus what you write your criteria. This is the dog sitting in front of you. There are many ways to achieve this and much easier with puppies. Since Mickey is 6 years old and has a lot of practice coming up to me, my approach is different than starting out a puppy.
I find that making one criteria part of your day is the best way to get it solid. If I'm sitting on the couch and Mickey comes up to me, shape for him sitting straight in front of me. If I'm standing in the kitchen and I have a few pieces of chicken around, do some shaping there. He has learned to come straight in to me and if his rump is off to the side, no chicken. Then of course give the piece of chicken in the exact spot you want him sitting. Treat in position is extremely powerful. I had to learn how to give his treat. Which is straight down from my fact or front of my body. If I give the treat to the side, he will shift his body to be in position for the goodie. So after several days have past and doing about five reinforcements for a straight sit in the kitchen, I will do a few in the living room or outside in the yard. Sometimes I use the ball, sometimes I use a yummy treat. Variety, what's going to happen next and do not do routine. I can't have Mickey predicting a pattern other than you will get reinforced for a straight sit in front of me. That will be consistent, but everything else will not, like where we train, how often we train, what time frame we train etc.
I get bored very easy with repetition and so will my Belgian Malinois, so I do vary what I do. To do the old traditional training that is so regime is important to some dogs, but other dogs will just tune out. Why I think so many people have gotten out of competition obedience. They are still doing that dry formality of training. Do three minutes in the kitchen, train while you are watching TV like during commercials and of course Debi Davis' famous "Training in the 'Loo!" Break it up. Lack of training is only a lack of planning. Also, we don't have to set an hour aside for training obedience. You need to build duration and training time slowly so you don't get that zoned out look from your dog. With proper shaping and conditioning skills you don't need more than 15 minutes in a full training session. If a shaping session is more than three minutes you are going to fry your dog's brain. This leads to burn out and stress. Yes, you can burn out and stress your dog out with clicker training. Especially if your shaping skills are really fine tuned and good.
Observe your dog and know when to back off. Many people want to perfect their dogs and over train. They a re very well meaning people, not realizing they are creating behavioral stresses in their dogs. Having trained several assistance dogs, we want them to succeed. Our heart and soul goes into their training because if they do not succeed, they will be career changed to another job, perhaps a pet. This expectation of the dog can literally burn the dog out and you do see the behavioral issues. Sometimes trainers don't see the behavioral problems or brush it off as something else and can't look at themselves. There is a lot of stress with training an assistance dog and too many put a lot of stress on young dogs that are not even a year old. Expectations are high in a young pup. This is a set up for disaster. Back off, enjoy the day and condition in a fun way for the dog.
Points to remember:
1. Plan your training session
2. Write out your criteria
3. Keep your training session short
4. Keep your shaping session timed and in the beginning, start out with 30 seconds. Eventually build up over a few weeks to a minute. Realize this is a shaping session where you have probably 20-30 reinforcements in a minute. When you get 80% success, incrementally move to the next criteria. Observe carefully if this criteria is working or not. If you are not achieving 80%, then you need to break down the criteria for success.
5. Have any training issues, look at your three principles for problem solving.
6. Try not to over train to burn out your dog. They are very willing creatures and will work past stress level. Watch for signs of stress.
My rate of reinforcements and timing come with practice, but criteria takes a lot of energy of my brain. The thought process of training. What do I want my dog to do? What do I choose to shape and have my dog do to set them up for the next step to get where I want? You can't expect to have what you want and get there in one step. It may take 100 steps to get there. Building slowly. The talent comes in knowing what step to teach next to make a strong foundation. Like a good teacher in mathematics. You don't all of a sudden start doing complex equations in two days, it takes several classes to get to higher level of equations and problem solving.
Each step I teach is a criteria. The rule of thumb is get 80% proficiency before moving to the next increment step or criteria. Always have your criteria plan ready as your subject may go fast and if you are not ready, training stops and you loose that fabulous momentum. Take the time to write out and plan your criteria. Or if you are better at doing it in your head, you can do it that way as well, but writing it out makes for good training records. Then in a year or two, you can see what you did and the written plan can jog your memory what path you took. I have had countless experiences where five or ten years later I think, how did I teach my previous dog how to do this behavior?
Recently I have been doing fronts. This is when the dog comes straight in to you and must sit just in front of you, not touching, but be within reach of you. Their sit must be straight in front of you. No hip shifted on either side. So what would be the criteria when training this?
At this point, the recall is set. The next step is back chaining. You go backwards. So the last thing the dog does in the behavior chain is what you reinforce, thus what you write your criteria. This is the dog sitting in front of you. There are many ways to achieve this and much easier with puppies. Since Mickey is 6 years old and has a lot of practice coming up to me, my approach is different than starting out a puppy.
I find that making one criteria part of your day is the best way to get it solid. If I'm sitting on the couch and Mickey comes up to me, shape for him sitting straight in front of me. If I'm standing in the kitchen and I have a few pieces of chicken around, do some shaping there. He has learned to come straight in to me and if his rump is off to the side, no chicken. Then of course give the piece of chicken in the exact spot you want him sitting. Treat in position is extremely powerful. I had to learn how to give his treat. Which is straight down from my fact or front of my body. If I give the treat to the side, he will shift his body to be in position for the goodie. So after several days have past and doing about five reinforcements for a straight sit in the kitchen, I will do a few in the living room or outside in the yard. Sometimes I use the ball, sometimes I use a yummy treat. Variety, what's going to happen next and do not do routine. I can't have Mickey predicting a pattern other than you will get reinforced for a straight sit in front of me. That will be consistent, but everything else will not, like where we train, how often we train, what time frame we train etc.
I get bored very easy with repetition and so will my Belgian Malinois, so I do vary what I do. To do the old traditional training that is so regime is important to some dogs, but other dogs will just tune out. Why I think so many people have gotten out of competition obedience. They are still doing that dry formality of training. Do three minutes in the kitchen, train while you are watching TV like during commercials and of course Debi Davis' famous "Training in the 'Loo!" Break it up. Lack of training is only a lack of planning. Also, we don't have to set an hour aside for training obedience. You need to build duration and training time slowly so you don't get that zoned out look from your dog. With proper shaping and conditioning skills you don't need more than 15 minutes in a full training session. If a shaping session is more than three minutes you are going to fry your dog's brain. This leads to burn out and stress. Yes, you can burn out and stress your dog out with clicker training. Especially if your shaping skills are really fine tuned and good.
Observe your dog and know when to back off. Many people want to perfect their dogs and over train. They a re very well meaning people, not realizing they are creating behavioral stresses in their dogs. Having trained several assistance dogs, we want them to succeed. Our heart and soul goes into their training because if they do not succeed, they will be career changed to another job, perhaps a pet. This expectation of the dog can literally burn the dog out and you do see the behavioral issues. Sometimes trainers don't see the behavioral problems or brush it off as something else and can't look at themselves. There is a lot of stress with training an assistance dog and too many put a lot of stress on young dogs that are not even a year old. Expectations are high in a young pup. This is a set up for disaster. Back off, enjoy the day and condition in a fun way for the dog.
Points to remember:
1. Plan your training session
2. Write out your criteria
3. Keep your training session short
4. Keep your shaping session timed and in the beginning, start out with 30 seconds. Eventually build up over a few weeks to a minute. Realize this is a shaping session where you have probably 20-30 reinforcements in a minute. When you get 80% success, incrementally move to the next criteria. Observe carefully if this criteria is working or not. If you are not achieving 80%, then you need to break down the criteria for success.
5. Have any training issues, look at your three principles for problem solving.
6. Try not to over train to burn out your dog. They are very willing creatures and will work past stress level. Watch for signs of stress.
Subscribe to:
Posts (Atom)