What Is a Reinforcement Schedule in Dog Training?

A reinforcement schedule is simply how often you pay your dog for a behavior. Pay every single repetition while the behavior is new. Once it is reliable, thin the payments out gradually. Those two sentences cover ninety percent of what you need, and the mistakes people make are almost always about the timing of the switch rather than the concept.

The reason this matters is that a behavior paid every time is easy to build and easy to lose, while a behavior paid unpredictably is slower to build and much harder to lose. You want the first one early and the second one later.

Phase one, pay everything

While your dog is learning what the cue means, every correct repetition gets paid. No exceptions, no testing, no seeing if he will do it for free.

• How long. Until he is at eight or better out of ten first cue responses, in that location, across three consecutive sessions.
• How much. One small piece of food per rep. Pea sized. You should be able to do fifteen reps without denting his daily food.
• How fast. Food in his mouth within two seconds of the behavior. Longer than that and he is no longer sure what he is being paid for.
• What counts as a rep. First cue, within three seconds, no repeating yourself. If you said it twice, that rep does not count and does not get paid.

The most common error here is leaving too early. People get a behavior working on day three and immediately start testing whether the dog will do it without food. Day three is not reliable, it is new. Testing it that early is how you end up with a dog who works about half the time.

Phase two, thin it out

Once the behavior is genuinely solid, start reducing. The mistake is going from every rep straight to occasional, which collapses the behavior. Step down instead.

1. Move to roughly three out of four. Run it for three or four sessions. You should see no drop in speed or reliability.
2. Then two out of three. Another three or four sessions.
3. Then one in two. This is where most pet dogs can comfortably live long term for everyday cues.
4. Then one in three or four for very established behaviors. Sit and hand targets can hold here. I would not take recall this far, ever.

Make it genuinely unpredictable. If you pay every third rep in a fixed pattern, your dog will learn the pattern and get sloppy on reps one and two. Vary it. Pay two in a row, then skip two, then pay one.

Pay the good ones. As you thin out, spend your payments on the fastest and cleanest repetitions rather than paying at random. This is how you get a behavior that improves after you stop paying everything, rather than one that slowly degrades.

When to go back to paying everything

Thinning out is not permanent and going back is not failure. Reload to every rep whenever any of these is true.

• New location. First three sessions anywhere new, pay everything. The behavior is effectively new again in that context.
• After a break. A week away, a vacation, an illness. Three days back on full pay.
• After a hard experience. A vet visit, a dog fight, a scare. Full pay for a few days while his baseline recovers.
• Whenever reliability drops below eight out of ten. This is the number that tells you, and it removes the guesswork entirely.

The behaviors I never thin out

This is where I differ from a lot of training advice, so it is worth being direct.

Recall gets paid every single time, forever. Not sometimes, not variably, every time, for the life of the dog. Recall is the behavior that keeps your dog out of traffic, and building in any uncertainty about whether coming back pays is a bad trade for the small convenience of not carrying food. If your dog comes when called, he gets something good. That is a permanent contract.

Anything genuinely dangerous stays on full pay too. An emergency stop, a leave it around something toxic, an out from another dog's face. These are used rarely, which means they get little practice, which means they need the strongest possible payment history.

How to fix a fade you rushed

This is a common repair job and it is straightforward. If your dog used to sit reliably and now responds about half the time, you almost certainly cut the food too fast.

1. Go back to paying every rep for one full week. Fifteen reps a day, in the easiest location you have. This feels like a step backward and it is a week well spent.
2. Use better food than you used the first time. Real meat, not kibble. You are rebuilding value in a cue that lost some.
3. Score at the end of the week. Ten reps, count first cue responses. You want nine or ten.
4. Then thin out on the four step ladder above. Slower than last time. Three or four sessions per step, not one.

Most dogs are fully repaired in two to three weeks. The behavior is not gone, it just stopped being worth doing, and that is a fixable problem.

Reinforcement schedules are easiest to run inside a fixed short session, because the rep counts do the thinking for you. My free 10 Minute Foundation blueprint lays out those sessions with the counts already worked out. Download the 10 Minute Foundation blueprint

Ready to train your dog?

Every drill in this article comes from the WorldWide K9 Academy course library. If you want the full step by step system, start with the $5 intro course.