Proofing and Generalization

You don’t need to read all of these first. Start with any command page — it will link you here when you need a concept.

In One Sentence

Dogs don’t generalize automatically — a command learned in one place has to be taught again, deliberately, in every new context where you want it to work.

What It Is

Your dog has a perfect sit. Quiet kitchen, you standing still, treat in hand — eight out of ten times, no problem. You take them to the park and ask for a sit. Nothing. Just staring at the other dogs.

Your dog isn’t ignoring you. They genuinely don’t know the word works here. Dogs learn commands as part of a whole picture: this room, this smell, you standing like that, those sounds. Change enough of the picture and the command doesn’t register the same way. This is called context-specificity, and it’s normal.

Proofing is the process of systematically teaching your dog that the command works everywhere. You do this by changing one variable at a time across three dimensions:

• Distance — how far you are from your dog
• Duration — how long they hold the behavior
• Distraction — what’s happening around them while they do it

Change one at a time. If you’re working on duration and you increase distance at the same time, you’ve made two changes and you won’t know which one caused the failure. More importantly, you’ve made the leap too large and your dog is now guessing.

How to Use It

The 8/10 rule: advance when it’s easy. If your dog succeeds 8 out of 10 repetitions at the current level, add one small increment to one variable. Not two variables. One.

The 6/10 rule: retreat when it’s hard. If your dog drops below 6 out of 10, you moved too fast. Go back one step. This is not failure — it’s calibration. The goal is to keep your dog in the learning zone: challenged but not overwhelmed.

Proof in real environments. The living room is where you start. The driveway, the sidewalk, the parking lot, the park at low-traffic times — these are where the command becomes real. Each new location is a mild proof. Each new distraction level is a proof. Go in with realistic expectations: the first time in a new place, your dog is probably at 60% of their home performance. That’s fine. Work there until it’s 80%, then add the next variable.

The order matters. Duration before distance before distraction is the most common approach for stay-based behaviors. For commands like “come,” distraction is often the main variable to proof. Match the proofing order to what the command actually requires.

Common Mistakes

Jumping from the living room to the dog park. That’s not one step — it’s twenty. The dog park has other dogs, new smells, movement, noise, and a dozen other variables you’ve never trained around. Start with one mild distraction, not everything at once.

Changing multiple variables at the same time. Adding distance AND a new location AND a new distraction in the same session is too much. Your dog will fail, you’ll get frustrated, and neither of you will know what went wrong.

Interpreting distraction failure as defiance. If your dog can’t sit reliably when a squirrel runs by, the sit hasn’t been proofed against squirrels yet. That’s a training gap, not a behavior problem. Fill the gap instead of punishing the failure.