Do I have to keep using treats?

You probably don't like the idea that you have to give your dog a treat in order to get him to do something. You probably want your dog to obey you simply because he loves you. You probably want him to work for you for free.

But dogs aren't robots. They're living creatures who can make a choice. If there's nothing's in it for them, they won't bother doing it. In other words, they wont work for free. And why should they? After all, you wouldn't go to work for free, would you? If you didn't receive a paycheck at the end of the week, you'd quit your job, right? Dogs are no different. There has to be something in it for them, or they won't continue to perform the behavior for you.

Positive Reinforcement

Clicker training is based on the scientific principle of positive reinforcement. Put simply, the principle of positive reinforcement says that your dog will only continue to do something if the consequence (of doing it) is really good!

In other words, if he loves what happens to him after he does a behavior, he'll keep doing that behavior in the future, because he knows the consequence is good!

For most dogs, the consequence has to be a treat!

Some dogs will perform a behavior for you simply to receive your praise, but for most dogs, praise is simply not enough, they want a treat as well.

So the long and the short of it is this:

Most dogs will respond to a cue (the command you give) only if you continue to reward their response (the behavior) with a treat.

If you stop rewarding a certain behavior, the behavior will die off and when you give the cue in the future your dog will not respond.

But you don't have to give him a treat every time

The good news is that after your dog has learned a behavior and reliably performs the behavior every time you give the cue (i.e. after many clicker training sessions and after you have stopped using the clicker too), you can stop giving him a treat for every correct response. And you only need to give him a treat every three, or so, correct responses (not every third correct response exactly, but at random intervals so that on average you are treating him once every three correct responses).

This is called a "variable reinforcement schedule". With a variable reinforcement schedule, you don't reward every response with a treat, you reward every few responses, at random intervals, to keep him guessing when a treat is coming! For example, on average you might reward one out of every three correct responses. It might look like this: treat, no treat, treat, treat, no treat, no treat, no treat, treat, treat, treat, no treat, treat, etc.

On a variable reinforcement schedule, your dog will keep working for you because he knows a treat is coming, but he just doesn't know when!

Note: A lot of beginners make a mistake when they put their dog on a variable reinforcement schedule. They spread rewards so thin and the dog simply stops performing the behavior because not enough performances are rewarded! So be sure that if you switch to a variable reinforcement schedule, you don't spread the treats to thin, and you do actually treat him one in every three correct responses, on average! A treat ratio of 1:3 (one treat for every three correct responses) is ideal. But you must make this the average, and give the treats at random intervals so he doesn't know when it's coming. If you give a treat for exactly every third performance (e.g. no-treat, no-treat, treat, no-treat, no-treat, treat) your dog will only perform the behavior well on the third time when he knows a treat is coming.

Below is an example of a variable reinforcement schedule on a ratio of 1:3. Which means the dog gets one treat for every three correct responses on average but the treats are delivered at random intervals:

Dog's response to cue	Treat Given
Correct	No
Correct	Yes
Correct	Yes
Correct	No
Correct	No
Correct	Yes
Correct	No
Correct	Yes
Correct	No
Correct	No
Correct	No
Correct	No
Correct	Yes
Correct	No
Correct	No

Remember: you only switch from a 1:1 ratio (one treat for every correct response) to a variable reinforcement schedule (such as one treat for every three correct responses) after your dog has mastered a behavior and performs it reliably any time you give the cue.

Next: Clicker Training Problems and Solutions