Obedience training
What is a variable reward schedule in dog training?
Last updated 23 September 2026 · How we write and check this
Short answer
A variable reward schedule means rewarding a dog unpredictably rather than every single time, so sometimes a correct sit earns a treat immediately and sometimes it takes two or three attempts before one arrives. This unpredictability, once a behavior is already well established, tends to make dogs try harder and respond more persistently than constant reward does, similar to why unpredictable payouts keep people pulling a slot machine lever. It should only be introduced after a behavior is solid, never while a dog is still learning something new.
Why unpredictability increases effort
When a dog cannot predict exactly which attempt will earn a reward, they tend to try harder and more consistently across every single repetition, since any one of them might be the one that pays off. This is the same underlying principle that makes unpredictable rewards so effective at maintaining behavior in a wide range of animals, humans included.
What it looks like in practice
A dog with a solid sit might be rewarded after the first attempt one session, then after the third attempt the next, and occasionally get a jackpot of several treats at once for an especially fast or crisp response. The unpredictability itself, not a fixed pattern like every other rep, is what makes this approach effective.
Why it only works on already solid behaviors
Introducing a variable schedule while a dog is still learning a new command removes the consistent feedback needed to understand what is being asked, which typically slows learning down or causes confusion. Waiting until a behavior is reliable across multiple sessions and settings before introducing unpredictability keeps the approach genuinely useful rather than counterproductive.
Key takeaways
- A variable schedule rewards a dog unpredictably rather than every single correct attempt.
- Unpredictability tends to increase effort and persistence once a behavior is already solid.
- Occasional jackpot rewards can be mixed into a variable schedule for extra effect.
- Only introduce it after a behavior is reliable, never while it is still being learned.