Obedience training
How do you transition a dog off treats once they know a command?
Last updated 23 September 2026 · How we write and check this
Short answer
Once a command works reliably across several locations, gradually switching from a treat every time to a treat every second or third time, then adding praise, play or simply continuing the walk as the reward the rest of the time, keeps the behavior strong without permanent dependence on food. Never cutting treats out abruptly is the key; a sudden switch from constant rewards to none at all is what causes commands to fall apart.
Waiting until the command is truly solid first
A command practiced successfully in at least three different locations, with a success rate around 90 percent, is a reasonable point to start reducing treats, whereas cutting food out of a command still being learned almost always causes it to weaken.
Moving to a variable, unpredictable pattern
Rewarding every response for a while, then shifting to roughly every second correct response, then an unpredictable mix somewhere between one in two and one in four, mirrors how a slot machine keeps people trying, and dogs respond to unpredictable reward in a similar way.
Replacing food with real life rewards
On the responses that do not earn a treat, offering enthusiastic praise, a quick game, or simply letting the dog continue toward whatever it wanted, like sniffing an interesting patch of grass, keeps the behavior worthwhile without food being the only currency involved.
Keeping proofing sessions on food a bit longer
Commands still being tested in new, harder locations or around genuine distractions should keep a higher rate of food reward even after everyday cues at home have moved mostly to praise, since reducing pay in a genuinely difficult new context too early is a common reason a previously solid command starts to slip again. Treating each new environment as its own mini reset for reward rate avoids this trap and keeps a command reliable even as circumstances keep changing around it.
Key takeaways
- Only start reducing treats once a command is reliable across several different locations.
- Shift gradually to rewarding roughly every second or third correct response, not every time.
- An unpredictable reward pattern keeps a dog trying harder than a fixed, predictable one.
- Replace food some of the time with praise, play or simply continuing the activity.