Reinforcing a dog for a job well done is important whether you want them to learn a new behavior or maintain a behavior they already know. There are different methods for reinforcing a behavior. Where your dog, (or a person for that matter), is in learning a new behavior will determine what type of reinforcement schedule you will use.

When a dog is learning a new behavior, you want to reinforce (positive reinforcement) him every time he does the behavior. This is a Fixed ratio schedule of reinforcement. Fixed means that you reinforce on a schedule and ratio means that schedule is based on the number of times the behavior is performed. You could have a fixed ratio of 2 or 3 or 100, which would mean every second, third or hundredth time the dog does the behavior you reward him. Of course when the dog is first learning a behavior you want to reward him every time, (a fixed ratio of 1), so that he learns what he is supposed to do.

Reinforcement can also be given at Fixed Intervals. This is when you give the reward based on time. So if the dog is doing a stay exercise you might reward him every minute he stays.

The problem with Fixed Ratios and Intervals are that they are very easy for dogs, (and people), to figure out. For example a coke machine is usually a fixed ratio of every time you put the money in a can falls out. So you only use it when your thirsty. If your dog learns that when you say sit and he sits he always gets a treat he'll start using you as a coke machine. When he wants a treat he'll sit. If he doesn't want the treat right then he'll continue sniffing the tree to see if the cat is still in it. Or if you reward him every minute he stays, after a minute he'll get up because he knows it's time for his treat. There is no surprise involved.

Which is why it is best to switch to a Variable schedule of reinforcement as soon as you know that the dog understands the behavior. Just as in a fixed schedule of reinforcement, a Variable schedule of reinforcement can be a Variable Ratio or Variable Interval.

A variable schedule of reinforcement needs to be random but there is still an average number of times the dog receives reinforcement. Pretend your dog knows what sit means and it is time to put him on a Variable Ratio schedule of reinforcement. You would start off rewarding him an average of every third time he sits. He might get a treat the first time he sits, and then the fourth time, and then the third time, and then the fourth time, and the the second time....As he progresses you would increase the schedule of reinforcement to where the average reinforcement is every 10 times. Then he might get a treat the fifth time, the twelfth time, the eighth time, the seventh time, the fifteenth time....

This is the hardest part of training on a Variable Reinforcement schedule: to not reward a dog for doing something good feels wrong. But stop and think about the lottery. This past weekend the powerball lottery was at 180 million dollars. Your chances of winning were something like 1 in 80 million. Yet EVERYONE wanted a ticket. If the reward is important enough people, (and dogs), will do the behaviors again and again to try and win the reward. Any type of gambling is based on a Variable Reinforcement schedule. People will sit and play slot machines forever. You don't see that kind of behavior at a Coke machine and there you always get something for your money.

So when you start switching to a Variable reinforcement schedule, think of yourself as a slot machine for your dog. Don't pay him off every time, and as the payoffs get spaced out further and further, make them bigger and better!! You will start getting stronger and more reliable behaviors because the dog now sees you as a slot machine, not a Coke machine.

Go Home

Back to Training

This page was created by and is copyrighted to Beth MacLehose. If you notice any graphics I have used that are yours and you would prefer I not use them let me know and I will remove them. Also all pictures and text is copyrighted to me or the cited author. Please email the author for permission to use material you find here.