Lure/Reward is exactly what it sounds like – luring the dog into doing something and then rewarding him for doing it. It is a very useful technique for many circumstances, especially basic obedience.
Here is a video of me putting sit on cue with Caffeine. (Actually, Caffeine already knew this, but it’s a decent representation.)
The lure is anything that your dog will follow such as a piece of kibble or a toy. The reward is almost always the lure itself.
Lure/reward training use a positive reward marker to “mark” the instant the dog performs the behavior being trained. For example, when training a sit, the moment the dog’s rear end hits the floor would be marked. A frequent marker is the word “yes.”
The process of teaching a behavior with lure/reward training can be explained in a few basic steps:
1) Lure the behavior and mark the instant your dog performs it. (This is step 1 in the video.)
2) When the behavior can be reliably lured, add the cue or “command” before the lure. (This is step 2 in the video.)
3) Get rid of the lure as fast as possible. Continue to mark and reward the behavior, but only produce the reward after the behavior is performed. (This is step 3 in the video.)
4) Once the behavior can be reliably acquired, stop marking (and rewarding) every instance. Reward at random.
There are a few important concepts embedded in these steps.
Don’t introduce the cue (or “command”) until the dog has mastered the behavior. Dog’s tend to respond to visual cues, such as a lure, better than audible signals like words.
Get rid of the lure right away. This is critical to avoiding “show me the money-itis.” A treat you have to show your dog before he does something is a bribe, not a lure.
Once the cue is taught and the lure is removed, it’s time to start rewarding at random. Behaviors that are rewarded at random are very strong and reliable. A behavior that is always rewarded falls apart the first time a reward is missed.
Steps 1 – 3 should usually occur in one session, unless the behavior is particularly complex or there is some exceptional difficulty.
Moving to step 4 can be tricky – too soon and the behavior may fall apart. Too late and you’ll get the same result for a different reason. For basic obedience the general rule is to shift to random rewards after the behavior is reliable is a few different places with some basic distractions, usually after about a week of regular practice.