Friday, March 20, 2009

Reward, reward, reward!

After spending a couple of days doing a lot of distance training with Susan Perry it really drives home some important things about reinforcement. There are several key points about reinforcement and training that distance handling really emphasizes and without a good understanding of them it is hard to develop good distance skills.

1. Motivation: dogs must be motivated to do agility in order to have enough impulsion to work away from the handler.

2. Independent obstacle performance: dogs must be able to do all obstacles with speed and accuracy with the handler at a distance away from the obstacle.

3. Directionals: dogs must have an understanding of a lateral distance cue ("out"), a straight away send ("go") and a turn away/lead change cue ("switch", "left", "right").

In order to really train distance with your dog you also really need to understand some key principles of learning and reinforcement. Dogs need a lot of self-confidence to work at a distance from their handler so it is important to know how to gradually build that confidence and to maintain it with proper use of rewards. If you have a dog that lacks confidence, is stressed during training, is tired, is lacking understanding of foundation concepts, is not motivated to do agility or is physically sore then you will see an even more dramatic decrease in agility performance when you ask the dog to work away from you. It will magnify any issues you may have in your training program with your dog.


1. Placement of reward is critical in distance training. This is why I use targets a lot in the beginning because it is easier to get the placement right. I use an empty target that the dog has been taught to drive to and when the dog touches the target then the handler goes out to the target and rewards. I remember when I first met Pam. She came to me for help with distance training because she and Jedi were struggling to get those USDAA Masters Gamblers legs. It was incredible how something as simple as changing where he was reinforced made the light bulb go on. Before she would reward Jedi for doing distance by having him come back to her for the treat. He had a hard time staying out because of it. Once she started tossing her bait bag out to where he was then suddenly he was much more willing to stay out and work! He was not about to leave that bait bag!


2. Frequency of reward is critical in dog training in general and in distance training specifically. I hope by participating in the seminar I was able to provide a good example of this. I felt I was trying to reward my dogs at key moments when they did something well. Whenever I train I am not interested in completing the entire course. I have sections in mind that I think may be a challenge for us as a team and when that goes well I want to stop and reward with toys or food from me. If things don't go well and I need to regroup then I will reward my dog with a "screw-up cookie". In the case of this seminar because it involved a lot of distance I tried to keep the momentum going by jazzing them up and starting again especially if I knew quickly what I needed to change. If I need more time to think about how to fix something then stopping and feeding or tugging with my dog while I sort it out is a good way to keep your dog having fun. Too often students get caught up in wanting to finish the entire course.


3. Reward what you like. If you find yourself walking a course or sequence and thinking that a particular section will be a challenge for you and your dog then that is the part you should plan on rewarding when it goes well. This will build confidence for both the dog and the handler. I often hear students talking about that in class and then when it goes very well they get caught up in the moment and continue with the rest of the course instead of stopping and rewarding that brilliant sequence. Then the moment is lost to reward that brilliant sequence or obstacle performance. If you really like something the dog did, or something that both of you did then you need to mark it and reward it at that very moment.

4. Hierarchy of rewards. It is important to know what your dog likes for rewards and to have a hierarchy of those rewards. For example I have one dog who likes toys as much as food however there are certain toys that are more well liked than other toys. When the task I'm asking my dog to do is more difficult or when we are in a more distracting the environment then I will use those rewards that are of highest value to the dog. I want to be highly valued out there. I ALWAYS want to be more valuable than any agility obstacle out there. I never use agility obstacles as a reward - there are many reasons for this. I recommend carrying a variety of treats and toys to any training situation. If a dog is giving me good effort I may praise them but not feed them in order to keep them from being stressed. Verbal praise is a low value reward. I want to use the higher value of reward when something has been done well.

5. Use of agility obstacles as reward - NOT!!! The main reason for this is that I want to be the center of attention for my dog and I want to be the source of all rewards. If my dog is more interested in doing agility obstacles than in working with me then we will have a lot of team work problems on the agility course. Dogs will seek out obstacles to use to reward themselves and this will cause a lot of off-courses. Second I have seen in backfire too many times where a supposed favorite obstacle was used as a reward and that obstacle was not performed correctly or some other behavior occurred instead and now you've marked a good behavior with a knocked bar or missed contact or missed weave pole.


If you follow these things you will also create a random reinforcement schedule for your dog so that they never know when and where they will be rewarded when doing agility. This is key to creating a dog with a lot of desire to want to play the game. If a dog only gets rewarded after the last obstacle of a course then for many dogs you will see them run slower in the first half of the course and speed up on the last half of the course. Those dogs have learned that the reward only comes at the end. I want a dog to run fast throughout the entire course. This is why running short courses at trials works well to build in the random reinforcement schedule into a trial setting. This is necessary for many dogs. If you incorporate training in the ring in the form of short courses at trials then you will enhance your performance at trials because you've instilled a random reinforcement schedule in a trial setting. Some dogs will become stressed on the start line of a trial because they know this is the one place where they will be asked to do 20 obstacles without a reward and the pressure of doing that can be overwhelming for a dog. The more frequently you trial with a sensitive dog then the more likely you will create a dog who starts to stress at the start line or part way through the course. I don't ever want my dogs to feel stressed at the start line because they are worried about having to perform. I don't trial frequently and I will do short courses as needed with any dog.

Annelise

2 comments:

  1. This was a very good summary of a lot of key points. Thanks for organizing this for us - it will make a great reference tool!

    - Becky

    ReplyDelete
  2. Wow! Viva and I had a great experience with this over the weekend. I tried handing from the bonus line on a Tunnelers course on Sunday. Viva went with great impulsion from the first tunnel to the second tunnel, then a little more slowly to the third tunnel, and checked in with me, which put the wrong end of the tunnel in play. When I said, "Go, tunnel," she took the wrong end of the tunnel. I praised her, and left the course with her to a jackpot. My mom, who'd come along to watch, was confused by the reward for a wrong course. I explained that I was really happy that Viva would follow my "Go" command eighty yards away, and that I wanted to reinforce that as a desired behavior. When I tried handling from the bonus line in the next Tunnelers course (it was a NADAC Games trial), Viva charged ahead through the first, second, and third tunnels. I was thrilled. With one well-timed reward, she'd gained the confidence to go farther on the very next run. She managed the fourth and fifth tunnels happily and correctly as well, and then took the wrong end of the sixth tunnel when I moved a bit too far. I praised and we left happily and I rewarded. So we've now done half a Tunnelers course with handling from the bonus line! This is huge progress for us. Thanks for the inspiration!

    ReplyDelete