At seven months, your dog was a star. Recall across the park. A sit-stay while you opened the door. Loose leash past the neighbor's cat. You posted about it. You told friends. You privately thought: we did it.
Then he turned nine months old, and someone replaced your dog with a stranger.
The recall that was automatic now earns you a glance, at best. The door stay dissolves. The leash manners you built for four months vanish exactly when a dog appears at the worst moment. And the internet has a diagnosis ready: he's dominant. He's testing you. He's blowing you off. You got sloppy.
Here's the less dramatic, much more useful truth: your dog may be in an adolescent stretch of reduced reliability.
What the regression window actually is
During adolescence, many dogs go through a period when trained reliability can drop. The timing varies by dog. The brain is reorganizing for adulthood - pruning connections, rewiring responses, recalibrating what matters in the environment. During that rebuild, behaviors that looked solid become unreliable. Not gone. Unreliable.
This shows up in predictable places: recall, stays, leash manners, and social tolerance with other dogs. A pattern owners sometimes report is - "he knew this perfectly last month."
The key word is developmental. This is not a training failure, not a personality change, and not a verdict on you or your dog. It's closer to a renovation: the building looks worse while the scaffolding is up.
What it is not
It is not dominance. Your dog is not making a move for household leadership; he is an adolescent with a brain under construction.
It is not spite. Dogs do not refuse recalls to punish you for the vet visit.
And it is not proof that "training doesn't work." The training is still in there - which may show up again as your dog matures and you keep practicing.
The three rules of the window
-
Lower your criteria, keep sessions short. If the recall works at ten meters but not forty, train at ten for a few weeks. You are not going backwards; you are keeping the behavior alive at a difficulty where it can still win. Wins are the currency. Bank them.
-
Management replaces trust. This is not the season to test reliability you don't currently have. Long line instead of off-leash. Gate instead of door stay. Distance instead of hoping past another dog. Every time you "test" the behavior and it fails, the failure gets practiced. Management is not giving up - it's refusing to rehearse the wrong version.
-
Keep paying. Adolescence is when many owners quietly stop rewarding, because "he already knows this." That is exactly backwards. Reinforcement history is what carries a behavior through the window. The behaviors that survive are the ones with the deepest bank accounts.
What not to do
Do not punish the regression. Correcting a dog for a behavior his developing brain currently can't reliably produce teaches avoidance, not reliability - and it damages the cue itself. "Come" that ends in punishment is a recall you're uninstalling.
Do not drill harder. Fifty repetitions of a failing recall builds a failing recall.
And do not decide it's hopeless and stop. Owners who quit inside the window meet an adult dog with no habits at all. Owners who kept the habits warm - low criteria, good management, steady reinforcement - meet an adult dog whose training comes back like it never left.
The other side
For many dogs, reliability improves with maturity and continued practice, though the timeline varies. The dogs who come out best are not the ones who were pushed hardest during it. They're the ones whose owners understood what was happening, adjusted the demands, and kept the reinforcement flowing.
Your nine-month-old is not a stranger. He's the same dog, mid-renovation. Keep paying him, stop testing him, and wait for the scaffolding to come down.
(The Regression Window, the second book in the Calm on the Leash series, walks this phase week by week.)
Explore the books: find your starting point.