The three second rule

I was wondering about this the other day while training Yogi to track a scent.

According to modern dog training theory, a dog forgets what it has done (for correction purposes) after something like 3 seconds. How is this possible? If a dog forgot what it had been doing 3 seconds ago then how would it ever be able to be trained?

Not looking for arguments on training methods, just a question that has been on my mind.

It isn't that it forgets what it has done, it is that it doesn't relate the reward or punishment to whatever it had done, once that time is up. So....you reward it within 3 seconds for sitting when you tell it to. If you reward it after 5 seconds, it may think you are rewarding it for looking up at the sky.

I thought that was goldfish
Why then do trainers tell you to extend the amount of time between achieving the behaviour and the reward, and also to only reward intermittantly?
I always take longer than 3 seconds to reward so how they've ever learnt anything then is a complete mystery.
I gave up on treats ages ago too much faffing for my liking, my dogs are trained solely with voice, hand signals and facial exopressions, least I can't forget to take them with us.

That's what I thought it was too A.

ut you probably reward with a look or your voice or a tickle Trouble...doesn't have to be a treat.

Yes I do as I said but not within 3 seconds so it's a bit of a miracle then that they're all so well trained. Total tosh in my opinion. FWIW

rewarding intermitently is called the variable reward schedule and the most effective and attractive way of rewarding. gambling is a good example of the variable reward schedule....people never get adicted to the change machines where you put a pound in and get ten ten pences back out again but often get addicted to one arm bandit machines? Why? It's the same hand movement, you are doing the same, putting money in and getting money out. With the one armed bandit the reward seems greater because you never know when it is going to come and whether it will be a jack pot or not.

behaviours that are rewarded on a variable reward schedule will be stronger than those that aren't.

The waiting until you reward thing is interesting, you do have to do it gradually or the dog will not make the conection, but if you teach sit well, and then increase the length of time unitl you reward gradually the dog should stil be sitting so knows that it is still the sit that got the reward. If they sat, stood up adn then got the reward, they would think it was for the standing up (obviously).

I can't remember why it works differently with negative connections, e.g why punishing after the fact doesn't work, maybe it is because to us it seems so obvious why they shouldn't have eaten the trifle on the table and what else could we be talking about, but to the dog there are so many things in the room that we could be refering to.

I bet when they're pups though you do it within 3 seconds...
The intermittent thing is the dog sitting at the table begging scenario, if an owner feeds them fromt he table twice, over a month...they will try and beg all the time, intermittent rewarding can reinforce a behaviour more than rewarding each time in the end...

I think it is also called living in hope!!!!!!!!!!!!!!!!!

I don't do punishment, full stop.
I do know why it works etc. but never mind that's not the point.