Field Notes · An evolving learning library from Angier K9

Pressure and release · The four behavioural levers

How do dogs actually learn?

There are only four ways to change what a dog does - add something he likes, take away pressure when he does it right, take away something he wants, or add a penalty - and each one is judged by what the dog does the next time, not right now.

The answer

On film the trainers answer this with a short model they teach at almost every handover: four levers. Two of them build a behaviour and two of them get rid of one. You build a behaviour by adding a benefit when the dog does it (food, play, a rub, freedom to go sniff) or by putting pressure on and taking it away the moment he does the right thing - a pull on the leash that stops when he sits, the collar vibration that stops when he downs. The trainers say the strongest way to build is both together: pressure, release, then a reward. You get rid of a behaviour by taking away something the dog wants - the ball goes away when he jumps - or by adding a penalty, which for dogs is usually social and physical at once: the owner's disappointed voice with a leash pop or a collar tap.

Two points turn the model from a vocabulary lesson into a working tool. First, every lever is judged by the next repetition, not by the moment: Kelton tells two families on consecutive days that training is never about getting the dog to do it now. Forcing a heel with a locked lead may put the dog in position and still make the next heel less likely. Second, you only find out which lever you actually pulled by watching the dog. The same pop can be pressure that teaches or a penalty that stops; a correction that leaves the dog more wound up was not a penalty at all, and praise can frighten a fearful dog. The most expensive mix-up the trainers describe is the recall: the owner chases the dog, the dog finally comes back, the owner is angry - and the one thing the owner wanted has just been penalised.

What is said on film: the four levers, the next-rep rule, the pressure-release-reward combination, the social-plus-physical penalty, and the recall example. What is inferred here: that the model is a standard go-home unit rather than an occasional explanation. That rests on its appearance in at least seven households between January and May 2026, including a PDF worksheet the families are emailed and slides at the handover.

What the trainers say

works a little better is if you think about the four behavioral levers there's four one you add

Kelton · Sputnik · Watch at 35:03 ↗

and then the last lever is you uh you penalize so um i like the word penalty a lot better now

Kelton · Sputnik · Watch at 35:57 ↗

So it's always the next rep.

Kelton · Sputnik · Watch at 37:14 ↗

simplified the way i explain it instead of using dog trainer lingo i figured out a different way

Kelton · Murphy · Watch at 1:12:15 ↗

to explain it so there's four behavioral levers in dog training and in people training you have

Kelton · Murphy · Watch at 1:12:20 ↗

it off that makes something more likely to happen again the next time penalties make something they

Kelton · Murphy · Watch at 1:19:56 ↗

the next thing but it could look exactly the same as an example a pop on the prong collar or a tap

Kelton · Murphy · Watch at 1:20:11 ↗

there's four levers and people just have more avenues that you can use them

the trainer (not named on film) · Bubba Gum · Watch at 36:47 ↗

the most powerful combo to create behaviors is you add pressure and then when they do the thing you

the trainer (not named on film) · Bubba Gum · Watch at 37:40 ↗

going to be a normally a combination of something social and something physical

the trainer (not named on film) · Bubba Gum · Watch at 40:04 ↗

time right so it's never training is never about getting them to do it now

Kelton (probable · Drago · Watch at 31:26 ↗

It's called the Four Levers.

the trainer (not named on film) · Apollo (the Sharkie) · Watch at 20:46 ↗

understand are that each of the levers affect the next repetition they do

the trainer (not named on film) · Milo · Watch at 23:44 ↗

And that's the other really, really big thing here is, is that all of this, you only know

the trainer (not named on film) · Milo · Watch at 25:21 ↗

Like the, I call them the four levers

Kelton · Max (Goldendoodle) · Watch at 54:45 ↗

what we're doing is there's competing motivators here so motivator number one

Bane · Watch at 29:08 ↗

because if that competing motivator is strong enough if you want to do that thing bad enough

Bane · Watch at 3:58 ↗

How it looks, dog by dog

Sputnik

The fullest plain-word version, given in January 2026 at a private lesson for a three-dog household where Sputnik, a high-drive cattle dog, was harassing the older dog. Kelton tells the family he has grown tired of the confusing trainer vocabulary and gives the four levers in order - add a benefit, remove access, remove pressure, penalise - noting that removing access is usually the weakest unless the dog is crazy about a ball or tug. He then separates pressure from penalty by purpose and closes on the rule that makes the model usable: 'it's always the next rep'. His worked example is the recall a frustrated owner accidentally penalises.

On film

works a little better is if you think about the four behavioral levers there's four one you add

Watch at 35:03 ↗

do that thing again because they lost the thing that love they lost access um that's not as

Watch at 35:22 ↗

and then the last lever is you uh you penalize so um i like the word penalty a lot better now

Watch at 35:57 ↗

So it's always the next rep.

Watch at 37:14 ↗

come back right um so that's just an interesting example of how you can accidentally apply the

Watch at 38:50 ↗

Murphy

At Murphy's go-home lesson the next day, Kelton gives the same model to an older, soft, cuddly cattle dog's family, saying he has since simplified it out of trainer lingo. Each lever is tied to what Murphy actually values: belly rubs and moderate interest in treats, set against a food-crazy Labrador that another trainer had. The explanation is then applied immediately to the family's real problem - Murphy's clashes with the younger household dog - which is to be handled with penalties, and Kelton spells out that a pressure-off and a penalty can look identical, a pop on the prong or a tap on the e-collar, and differ only in which way they push the next repetition.

On film

simplified the way i explain it instead of using dog trainer lingo i figured out a different way

Watch at 1:12:15 ↗

they're not like right like cuz like with Murphy it's not the strongest I

Watch at 1:14:11 ↗

it off that makes something more likely to happen again the next time penalties make something they

Watch at 1:19:56 ↗

but if he and this is the one you'll be using with that that's an example of a penalty

Watch at 1:20:37 ↗

Bubba Gum

The go-home version for a sensitive shop dog. The trainer frames the levers as how you create behaviours and how you get rid of them, and names the strongest way to build one: pressure, release when she does it, then a benefit - demonstrated live with a leash pull into a down followed by a party and ball. For removal, taking away what she wants is the weaker lever for most dogs; the penalty dogs understand is social plus physical, and the owner's displeasure weighs more than the trainer's because of the relationship. The measure of a penalty is her reaction: if she chills out, it worked.

On film

there's four levers and people just have more avenues that you can use them

Watch at 36:47 ↗

the most powerful combo to create behaviors is you add pressure and then when they do the thing you

Watch at 37:40 ↗

going to be a normally a combination of something social and something physical

Watch at 40:04 ↗

And for you, your social penalty is going to be more effective with bubblegum than mine, right?

Watch at 40:13 ↗

and she's gonna be like oh my god right he's mad that sucks right um and if you get that reaction

Watch at 41:12 ↗

Bane

The same logic appears in Bane's protection-dog home sessions under another name, competing motivators. A South African Mastiff who guards the house is being taught to stay in his crate when a stranger is inside. The trainer describes it as a contest between the defence motivation and what breaking the crate costs him, and makes the job to make the second outweigh the first. Two days later he adds the nuance that a dog who wants something badly enough may push through a level of pressure and simply escape it quickly, which is why proofing is done at a level the dog would rather avoid than escape.

On film

what we're doing is there's competing motivators here so motivator number one

Watch at 29:08 ↗

right several times right um but the competing motivator is the negative reinforcement right

Watch at 29:20 ↗

because if that competing motivator is strong enough if you want to do that thing bad enough

Watch at 3:58 ↗

different if they don't know a thing you can't teach them that way at avoidance levels but you

Watch at 4:36 ↗

Apollo (the Sharkie)

Not a map member; read because its go-home lesson uses the model by name. The family is emailed a PDF, 'the Four Levers of Behavioral Change', and walked through it in the textbook terms - positive and negative reinforcement, positive and negative punishment - with the seatbelt chime and the nagged teenager and the dishes as the negative-reinforcement examples, and a long list of rewards beyond treats: environment, freedom, play, affection, exercise, overcoming a challenge. The trainer adds that reinforcement and punishment can use the same tools and differ in goal and usually in amount.

On film

It's called the Four Levers.

Watch at 20:46 ↗

reinforcement a good analogy that I like is when you get in the car if you don't

Watch at 24:48 ↗

to is the goal and normally the amount of pressure you use so if you're punishing something you might

Watch at 30:06 ↗

Milo

Not a map member; read because the in-program owner lesson for Milo and Josi opens on the model. The trainer adds three things the shorter versions leave out: the levers have limits, because personality rooted in genetics cannot be trained away; each lever only affects the next repetition, and the two reinforcement levers are used together; and you only know which lever you pulled from the dog's reaction - an amped dog who gets more excited after a pop was not punished, and a dog that looks as if it could go straight back to the behaviour was not punished either.

On film

aspects of your personality because those things are rooted in genetics so I

Watch at 12:06 ↗

understand are that each of the levers affect the next repetition they do

Watch at 23:44 ↗

And that's the other really, really big thing here is, is that all of this, you only know

Watch at 25:21 ↗

dog look if they look like they could go right back to it at any moment they did not get punished

Watch at 26:30 ↗

Max (Goldendoodle)

Not a map member. At Max's go-home in May 2026 Kelton shows slides he calls the four levers for behavioural change, in quadrant terms, again with the seatbelt chime. The new emphasis is on removal: taking away what the dog wants is less effective for most of the dogs he works with but carries less risk of fallout, and fallout is what happens when someone goes far too hard on a dog with no relationship or cooperation established.

On film

Like the, I call them the four levers

Watch at 54:45 ↗

And then the last one is negative punishment, which I've found to be less effective for most dogs that I work with.

Watch at 1:04:59 ↗

fallout so fallout is when you go way way way way way too hard and they don't understand and then

Watch at 1:05:37 ↗

Drago

Not a map member. At Drago's go-home the next-rep rule is the whole point: locking the flexi and dragging the dog into heel puts him in position now and probably makes the next heel less likely. Kelton then gives the two building levers in 'plain English' - add a benefit, which for Drago was rubs rather than treats, and the release.

On film

you have four levers, right? And in those four levers, what you're trying to do is decrease or

Watch at 31:13 ↗

time right so it's never training is never about getting them to do it now

Watch at 31:26 ↗

that's correct now yeah right but that doesn't increase the likelihood that he will do it it

Watch at 32:28 ↗

Thumper

Not a map member. The model used as a staff teaching tool: a coaching trainer quizzes Thumper's handler mid-recall on which lever the collar taps were (pressure and release) and which the cheerleading was (a positive reinforcer to speed him up).

On film

the four behavioral levers the ways you change behavior

Watch at 9:53 ↗

is it positive reinforcement negative reinforcement right so it's pressure and release

Watch at 10:04 ↗

Why it works

As the trainers explain it, a lever works through what the dog learns to expect next time. A benefit makes the behaviour that earned it more likely; pressure that stops when the dog does something makes that something more likely, because he learns he can escape the pressure and later avoid it; losing access to a wanted thing, or meeting a penalty, makes the behaviour that caused it less likely. Kelton's distinction between pressure-off and penalty is about direction, not appearance: 'it could look exactly the same'. The strongest build combines pressure, release and reward. For removal, a penalty is social plus physical because that is how dogs correct each other, and the owner's voice counts for more than a stranger's because of the relationship. At Bane's sessions the trainer adds the threshold idea: a dog motivated enough to want something will pay a small cost to get it, so the cost has to outweigh the want. The trainers say the model applies to people as well as dogs; the only difference is that people can be talked to.

On film

the next thing but it could look exactly the same as an example a pop on the prong collar or a tap

Watch at 1:20:11 ↗

sick right so they avoid pressure or they escape pressure and they like that that's a good thing

Watch at 37:45 ↗

the most powerful combo to create behaviors is you add pressure and then when they do the thing you

Watch at 37:40 ↗

going to be a normally a combination of something social and something physical

Watch at 40:04 ↗

because if that competing motivator is strong enough if you want to do that thing bad enough

Watch at 3:58 ↗

the time all this stuff works on people yeah everything i tell you about dog training works

Watch at 1:19:14 ↗

Common owner mistakes

  • Penalising the recall. The owner chases the dog, the dog finally comes back, and the owner is visibly angry or drags him inside - so coming back is what got penalised.
    On film

    really frustrated is an easy mistake to make with a recall say you're teaching a dog a recall and

    Watch at 38:08 ↗

    come back right um so that's just an interesting example of how you can accidentally apply the

    Watch at 38:50 ↗
  • Forcing the position now instead of training the next time: locking the flexi and pulling the dog into heel.
    On film

    example um say that i want him to heal right a lot of times people's first instinct is to do this

    Watch at 31:51 ↗
  • Believing a correction happened because the owner did something. If the dog looks as if he could go straight back to it, or gets more amped, no penalty occurred.
    On film

    or very often that positive punishment thing people think they're punishing their dog and then

    Watch at 26:15 ↗

    and she's like oh and chills out you did it right if she's still like and it's all about like seeing

    Watch at 41:19 ↗

Questions owners ask

Isn't this just the four quadrants?

Yes, and the trainers say so: the handover worksheet lists positive and negative reinforcement and positive and negative punishment. Kelton's plain-word version renames them - add a benefit, pressure turning off, remove access, penalty - because, he says, the textbook words don't mean what people think they mean. What the trainers add to the textbook is the next-rep test and the instruction to read the dog to find out which lever you actually used.

On film

So you got positive reinforcement,

Watch at 21:10 ↗

they made it really confusing um because the words don't mean what people think they mean

Watch at 34:46 ↗

Isn't pressure the same as punishment?

No, though the action can look identical. Pressure is already on and turns off when the dog does the thing, so he does it sooner next time. A penalty is added after he did something, so he is less likely to do it next time. A pop on the prong or a tap on the e-collar can be either; the difference is purpose and, often, amount.

On film

it off that makes something more likely to happen again the next time penalties make something they

Watch at 1:19:56 ↗

to is the goal and normally the amount of pressure you use so if you're punishing something you might

Watch at 30:06 ↗

What is the best way to teach a new command?

The trainers name one combination as the strongest: put light pressure on, release it the moment the dog does it, then add a benefit he actually cares about. For one dog that is ball and tug, for another it is a rub.

On film

the most powerful combo to create behaviors is you add pressure and then when they do the thing you

Watch at 37:40 ↗

let's go let's go right so that right there is exactly how you create

Watch at 38:22 ↗

Why does my voice matter more than the trainer's?

Because a penalty dogs understand is social as well as physical, and the social half depends on the relationship. The trainer tells Bubba Gum's owner outright that his annoyance will land harder than the trainer's.

On film

And for you, your social penalty is going to be more effective with bubblegum than mine, right?

Watch at 40:13 ↗

The lesson

The goal

An owner can name which of the four levers they are using, predict its effect on the next repetition, and check the dog's reaction to confirm it.

Learn first

The drill

[object Object]

Watch out for

  • Penalising the dog who finally comes back.
  • Judging a lever by the moment instead of the next repetition.
  • Assuming a correction worked because it was delivered, without checking the dog's reaction.

Check yourself

You tap the collar and your dog comes to you; later you tap the collar when he jumps on a guest. Which lever is each, and how will you know if either worked?