How do I know the training isn't hurting my dog?

Kodak is where the standard is stated, and the circumstances matter: he is a flight-risk dog, frightened of sounds and people, on his first outing, and the trainer is explaining to camera why he is not going to drill him. The argument runs in one movement. He could get obedience out of this dog today. It would look gross, disgusting, sad, depressed and forced. That is not the product, and therefore the obedience waits. Then, immediately, the guard rail: this is not a refusal to use negative reinforcement, boundaries or penalties, it is an order of operations, and there are other things that have to happen first. What those other things are, he says next - environmental exploration on twenty-six feet of line in a natural place, with the dog left to sniff and look. The same file also supplies the failure sign in its plainest form. He tells the viewer to look at the dog's face: the mouth has been open for a good part of the walk, and it is not the heat, it is stress, and the stress is the barrier. Weeks later, in an evening session, the positive sign arrives with no pressure involved at all - Kodak falls into heel, holds it, stops when the handler stops and then looks up to ask if that was right.

Bo's file is where the standard is turned into a claim about learning rather than about aesthetics. Reporting progress to an owner who is not there, the trainer lists two things and says the second is the bigger one: not that the dog is becoming proficient, but his attitude when negative reinforcement is used, because that attitude is what makes him able to learn at all. The evidence he offers is the kind anybody can check on the footage - the dog comes in, sits, is released and immediately gets hyped, and often gets the zoomies straight after being given a command. He is careful, in the same minute, not to over-read: the dog shakes his head and the trainer says out loud that he did not use the e-collar and that this was not a reaction to anything. A second Bo session applies the same standard as an acceptance criterion on a single position. He is teaching a down without food, and he rejects two reps that would have counted as successes - one where the dog sinks into it by suppression, one where he slides into a sit - because neither is the behaviour being trained.

Roxy's go-home lesson has the owner-facing version, and it is the most portable thing in this concept. The trainer reduces the whole question of whether a consequence worked to one factor and one question: does it look like it. If it does, it worked; if it does not, it did not. His example is deliberately uncomfortable - you are holding the dog by the collar telling her never to do that again, and she is thrilled, in which case nothing has happened at all. Then he turns it around and applies it to rewards, using himself as the example: back when he trained with food he would hand a dog a treat the dog did not care about, and that was not a reward either. The test is symmetrical, it is checkable in the second after the event, and it requires no training to use.

Rocko is the one case where the standard is shown driving a decision that has nothing obvious to do with obedience. The trainer spends a short session on tug, and when he explains why, the reason is not that play is good for the dog in itself. It is that improving the dog's drive changes how the finished obedience will look - he wants to eliminate a quality in it that he cannot quite name and trails off trying to. It is the clearest evidence in the corpus that the happy-looking dog is treated as an engineering target rather than a hope: work is scheduled upstream of the obedience specifically to change its appearance. Rocko's session is filed in the kodak notes folder; Kodak is not in it.

How do I know the training isn't breaking my dog's spirit?

You watch his attitude, and you treat it as the primary reading rather than as a nice extra. The standard the trainers here state is that a happy-looking dog is the proof the method is being applied properly, and they say it about a dog who is frightened of nearly everything, while explaining why they are not going to drill him. The signs to look for are small and immediate: does he bounce, shake off or get silly after a rep; does he stop when you stop and look up at you to check. The signs that it is going wrong are equally small: does he sink into positions instead of moving into them, does he collapse into some other position, does his mouth hang open on a cool day. If you are seeing the second set, the answer in this corpus is never to push harder - it is to go back to something earlier and easier and let the stress come down first.

Does 'we want a happy dog' mean you don't use corrections?

No, and the trainer who says the line pre-empts exactly that reading in the next sentence. Negative reinforcement is still used, the dog is still required to do things, boundaries are still created and things are still penalised. What the standard governs is the sequencing - his phrase is an order of operations, and on a frightened dog the things that come first are exploration, decompression and simply being somewhere without being asked for anything. The happy dog is not an alternative to pressure. It is the condition pressure is supposed to leave him in.

He does the command but he looks miserable. Is that still a win?

Not in this method. A body in the right shape arrived at the wrong way is graded as a failed rep, and you can watch a trainer refuse one: teaching a down without food, he rejects the version where the dog suppresses into it and the version where the dog slides into a sit, and holds out for a rep where he simply goes down smoothly. He then releases and pays that one. The practical consequence for you is that you should stop counting compliance and start counting clean reps, which will be a smaller number and a more honest one.

How do I know whether my correction actually worked?

Look at the dog and ask one question: does it look like it. That is the whole test as it is given to an owner on the day her dogs go home. If the dog looks penalised, the penalty happened; if she is delighted while you are holding her collar telling her off, it did not, whatever you intended. The same trainer immediately runs the test the other way for rewards, using his own past as the example - handing a dog a treat she does not care about is not a reward either. Note what the test does not tell you, which is what to do next. Nothing in this corpus says whether a failed correction should be repeated harder, delivered differently, or abandoned.

Why is my dog playing tug in the training videos? I asked for obedience.

Because on some dogs the play is what changes the look of the obedience later. On one dog in this corpus the trainer runs a short tug session and says plainly what it is for: improving the dog's drive will make the finished obedience less like something he does not want it to look like. He is not training play because play is nice. He is putting work upstream of the obedience so that the obedience comes out right, which is the same standard applied a month earlier rather than a rep later.

Grading the position and not the dog. A body in the right shape, arrived at by suppression, is scored here as a failed rep rather than a success.

Telling the dog off and assuming, because you meant it, that something happened. If she is delighted while you hold her collar, nothing happened at all.

Assuming a treat is a reward. A dog who does not care about the food was not rewarded, and the trainer says this used to happen to him.

Drilling obedience into a frightened dog because he is technically capable of it. It can be got, and it comes out looking sad and forced, and that is not the product.

Reading the open mouth as heat. On a mild day it is stress, and stress is the barrier that has to come down before anything else works.

Blaming the tool for something it did not do. A head shake right after a cue looks like a reaction to the e-collar; the trainer stops and says the collar was not used.

He does what you ask and then goes flat - no shake-off, no bounce, nothing. The behaviour is right and something about it feels wrong to you.

He sinks into the down rather than moving into it, or folds into a sit instead, as though he is trying to make something stop.

His mouth hangs open on a walk on a mild day and he is not sniffing anything. You have been reading it as heat or thirst.

The good version: he gets silly straight after being told what to do - a shake-off, a bounce, a lap of the garden - and you have been treating that as him not taking it seriously.

The other good version: he stops when you stop and looks up at you, as a question rather than as a stare, and waits.

You tell him off and he looks pleased with himself. You have concluded he is stubborn; the trainers conclude the telling-off did not happen.

He takes the treat politely and does not seem to want it. You have been calling him food-motivated because he eats.

He has no play in him at all - no tug, no interest in a toy - and you have written that off as personality rather than as something to build.