How do we get our dogs so reliable in their training that it almost seems like they’re addicted to it?
And how do you stop using so many treats in training?
A common piece of advice says the answer to both questions can be found in… Vegas.
“Slot machine training” is a technique that’s floated around the dog training world for a long time. The idea usually goes like this:
“Once your dog understands a command, start giving them a food reward only some of the time. So if you recall your dog, you’d give them a treat the first time. The second time, just give them a pat and send them on their way. They’ll never know whether they’ll get a treat or not, so they’ll gamble and come every time you call, because there might be a treat in it for them.”
(The technical term for slot machine training is a variable schedule of reinforcement)
The logic is that by making rewards unpredictable, the behaviour becomes addictive. Like gambling for humans.
But does it actually work?
Yes, it really works if you do it right.
And it really, really doesn’t work if you get it wrong.
Why the slot machine analogy isn’t quite right
On the surface, this neat little comparison makes sense. But as usual, life is more complicated than that.
There’s a reason casinos are designed the way they are. Full of bright lights in colours that increase our attentiveness, fun buttons and handles that are satisfying to play with, no windows to remind you of the outside world, and drinks flowing endlessly.
None of this is by accident. These details tap deep into human psychology to keep us comfortable, stimulated, and fully absorbed. We can’t help but be ensnared.
Now, would the intrigue of a slot machine hold as much power over us if it was a simple, boring grey box placed in the middle of Disneyland?
Probably not.
So when we’re standing in a field full of exciting smells, or a busy dog park – holding a tiny pouch of kibble that may or may not be given – and expecting our dog to be addicted to training, we’re going to be disappointed.
The addiction to slot machines doesn’t come from the rate of reinforcement alone. It also relies on the level of possible reinforcement.
People don’t play slot machines for hours on end in the hopes of maybe winning $100. They play because they think they could walk away with a life-changing jackpot.
And how big that jackpot needs to be to get us hooked depends entirely on how desperate we are for it.
Basically: that kibble ain’t gonna cut it.
At least not by itself, and not at first.
The psychology behind it
Let’s take a look at the mechanism behind this training technique and see if it actually holds up:
Imagine someone says to you, “Give me a high five to earn a prize.”
Weird, but okay sure, you can do that. You give them a high five, and…
Nothing happens.
You try again. Still nothing.
The third time, they hand you $5.
You might keep playing for a while. $5 isn’t a great prize, but it adds up. However, if too many rounds go by with no reward, you’ll start wondering if it’s worth it.
Now imagine this game is keeping you from doing other things you actually want or need to do. Like your job where you earn real money, or seeing friends, or watching your favorite show. Suddenly, the rounds where you don’t win aren’t just “no reward” rounds. They’re losses. They cost you other opportunities.
Eventually, the occasional $5 isn’t enough. You either need to know that there’s something really impactful available to win, or that you’re guaranteed to get $5 EVERY time you play.
And if this person now follows you around everywhere asking for high fives? Pfft. You’re over it. And their constant presence is getting annoying. You just want them to leave you alone.
This is how dogs experience slot machine training when it’s done wrong.
If we mess this up, we don’t end up with a dog who’s addicted to working with us. There’s a good chance we’ll get the opposite: a dog who sees us as an unpredictable nuisance not worth paying attention to.
How not to use the slot machine method
You may have heard a version of slot machine training like “once the dog understands a cue, stop using rewards every time. Random rewards create the strongest behaviours.“
Sounds clever, but it’s not really true.
Here’s what the research (and a whole lot of frustrated dog owners) tells us:
The strongest, most reliable behaviours come from rewarding your dog every single time during the learning phase.
And once your dog truly knows the behaviour, then you can start mixing things up. But the variation should come in what your dog gets, not whether they get anything at all.
When we talk about “variable reinforcement,” we don’t mean “sometimes the dog gets nothing.” We mean the type or intensity of the reward varies. The difference might sound subtle, but it’s huge in practice. And it’s the difference between a dog who’s motivated to work with you and one who tunes you out.
Think about it this way:
When we want to weaken an annoying dog behaviour, like going nuts when you put their leash on, we’ll start holding desensitization sessions where we put the leash on and nothing happens.
So sometimes the leash means they go for a walk, but mostly it doesn’t mean anything.
Basically, you make the reward unpredictable.
And if the only-reward-sometimes slot machine method worked, then this would just create dogs who are MORE crazy at the sight of the leash. But in most cases, the leash excitement fades.
So when you apply that same logic to obedience commands, uncertainty doesn’t create reliable behaviors. It just creates doubt.
Let’s break down how to use variable reinforcement the right way:
How to use slot machine (variable reinforcement) training to create strong behaviours and get your dog to pay attention
1. Make sure the potential winnings are worth it
Low-value rewards like kibble or dry dog biscuits are fine sometimes. But they’re the consolation prize, not the jackpot.
To tap into your dog’s natural seek-and-reward system, they have to know the possible payout could be leagues better than anything else around them. If your dog is in an exciting environment, a low-value treat won’t compete.
Speaking of competing with exciting environments, the best approach is usually to STOP competing. Instead, put those distractions to work for you by using them as rewards.
See: Is there hope for your stubborn dog? Why some breeds are harder to train, and the secret to success
2. Vary the type of reward, not whether they get one at all
This is the key to success here. A correct attempt should never go without some kind of reward.
Even when the dog doesn’t nail it perfectly, but they showed that they understand the concept, they should be rewarded with something meaningful to them. A for effort, and all that.
But rewards come in various value-levels and forms. Instead of aiming to fade out reinforcement altogether, aim to incorporate different types of rewards into your training.
For example:
- Maybe this time it’s a liver treat
- Next time, a piece of real chicken
- Maybe you suddenly get goofy and run with them
- Maybe you pull their favorite toy out of your bag
- Maybe they get the world’s best belly rub
- Maybe their best dog friend shows up right after they come when called
- Maybe you unclip their leash and give them some freedom (where safe to do so, of course)
This keeps dogs curious and invested in the game. Curiosity may have killed the cat, but it makes one hell of a good dog.
(And hey, this also means you get to carry fewer treats, if that’s something you’re concerned about)
3. Gamble responsibly
Variable reinforcement isn’t for the learning stage, it’s for the proofing and strengthening stage.
If you start “gambling” too soon, you’ll end up with confusion and frustration on both ends of the leash. Wait until your dog has demonstrated that they really understand what you’re asking.
For example:
If Fido sits on cue reliably in the house and backyard but struggles with it at the dog park, that’s a good sign that they know what the word “sit” means – they’re just distracted. That’s when you can start experimenting with variable rewards in easier, familiar settings where Fido can sit reliably.
The big takeaway
Slot machine training is powerful when it’s used to build addiction-like feelings around behaviours your dog already knows and that you want to strengthen. It’s not for teaching new behaviours, or testing ones they’re not already confident in.
Reward every time at first to build reliability, then vary the type of reward (not its existence) to maintain motivation and enthusiasm. That’s how you get a dog who isn’t just obedient, but genuinely loves the game.
