Monday, December 21, 2009

Placement Of Rewards

One of the common discussions I have with my students is the proper placement of rewards. They are, of course, very used to hearing me encouraging them to reward in a certain place and a certain way and they do a great job applying the concept. However, when I observe trainers that have described a problem between their ring performance and their training sessions, very few of them are rewarding the dog with food (or toys) at the proper time and in the correct place to help communicate exactly what is wanted at the time.

This happens especially often during heeling or during fronts and finishes in obedience and in agility it happens with start line stays, contacts and weaves. The food or toy is giving totally out of the context and most of the time what is rewarded is the release of that position or something totally different than what is wanted. In obedience exercises, the trainer many times incorrectly rewards the end of the exercise - not the exercise itself. What happens is that the trainer gets so excited about the progress, that they throw a party with tons of praise and release the dog from that position while mindlessly giving the dog the reward for what happened moments ago. The dog never associates the previous behavior with that reward.

It's important for the dog to "feel" the desired behavior and to be rewarded in that position. If you are working on a position (down, sit, 2o2o on contacts, etc) let the dog be in that position for a short period of time while you reward them with food or with a game of tug for 2o2o on contacts. Don't let them release from position just because you are rewarding them (another training issue that confuses dogs and probably another future topic).

So, because many obedience exhibitors train with food (incorrectly) and trial without food and see that their ring performances are horribly different than their training sessions, they think that they need to eliminate the food completely so that the dog gets used to not having food. That will only continue to confuse and stress the dogs because there will no longer by any pay for training and there has not been enough association between the rewards and the required performance in order to get a consistent happy performance. Without rewards, performances become lack luster, slow, robotic, and unenthusiastic by the dog. Most people want a happy dog in the obedience and the agility ring. They don't want slow and unenthusiastic performances because those lead to lots of mistakes, and nonqualifying performances. Games, rewards, and breaking down the exercises into core/foundation issues are very important to success.  Many times once the foundation issue is identified and worked, rewarded, proofed, and polished many of the other problems go away.

Classical conditioning is a very powerful tool to use while teaching obedience exercises, like heeling, fronts, finishes as well as all agility obstacles. It associates a very positive experience with certain positions as well as certain behaviors. If the placement of the reward is not correct, then the dog never truly understands what the trainer is trying to communicate. The dog only knows that their trainer is sincerely happy and get a cookie or a toy away from what just happened. If your placement of reward is not correct, then you most definitely will not have the performance that you want in obedience or in agility. Your training sessions will always be different than your ring performances because you have not properly associated what you want in the various core/foundation behaviors with proper rewards.

Here is an analogy I have used with my students that seems to make the most sense for them. When you get a new puppy, they initially do not like their crate. How to you crate train a puppy? You use classical conditioning with lots and lots of rewards and positive associations with the crate. Puppy goes in the crate, you throw a treat in there.  Simple, right? You would not make any positive association with that crate if you got the puppy to go into the crate and then only gave them food when they came out. You wouldn't send the puppy in the crate, then have a party and let them come out for a treat. The puppy would not positively associate the crate with something good and they would quickly learn that if they go into the crate and then come out, they are rewarded. They would not want to stay in the crate.

Once dogs have a positive association with a position, behavior, or any execution of an obstacle, then it's time to try and proof it by SERIOUSLY trying to get them to fail. Making a game out of making mistakes and getting the dog to problem solve will do wonders for your ring performances in obedience and agility as well as your confidence in the dog. If you are too quick to make things easy when the dog makes a mistake or the dog reacts to failing you are not helping them.  In fact, they are actually training you to help them so they CAN be rewarded. Most trainers don't go through enough steps to let their dogs think and problem solve enough in order to be an team partner. What usually happens is that trainers try to hard to control the dog as well as handle the exercises/obstacles and that is WAY more work than you really need to do. If you teach the dog his/her responsibility by allowing them to make mistakes without constantly intervening to help them, they will eventually learn their responsibility and then all you have to do is handle.  Handling is SO much easier when you don't have to control the dog, too.

Next time you are training an exercise, and obstacle, or a sequence of obstacles, pay attention to what you are rewarding and when you are rewarding. It's fine to play with your dog between exercises or obstacles, but make sure that they have been rewarded many times in the proper place in order to ensure that they are associating it with what is wanted.


  1. As usual, well said Nancy! Your timing and placement of reward is awesome:) You can tell just by watching how you and your dogs work together :) Can't wait to see more of Miss Schema as well!!