Positive reinforcement

"Positive Reinforcement" is a training method by which the dog receives a reward for every desired performed behavior. A reward is anything that the dog could consider rewarding at that particular moment. That could be a food treat, but also could be a game, running over to greet a doggie buddy or just sniffing a certain interesting spot or more. By giving the dog a reward, I'm communicating to him what I want him to do, that he's done it correctly. A rewarded behavior is more likely to be correctly performed in the future. Distractions are gradually added so that the probability that the dog correctly performs is always optimally high.

For example: the dog sits and receives a treat. Because of this, the dog will more likely perform the sit in the future. This is first taught at home with few distractions and these distraction are slowly added, such as moving to the garden and then perhaps when another dog is present, house guests come over, etc.

Undesired behaviors are often where possible ignored. And because in such cases no reward is given, these undesired behavior are then less often performed.

For example: The dog jumps around instead of sitting. The owner ignores the dog and this behavior until the dog finally does sit. THEN the owner will give the dog a reward (reinforcer). The dog doesn't jump around as often, because doing so does NOT get him his reward.

In case the owner can NOT ignore the undesired behavior - for example self-rewarding behaviors - the owner the trains an alternative behavior that makes the undesired behavior impossible.

For example: A dog jumps continually on house guests because he's so excited t see them. "Sit" is a behavior, that the dog can perform, that makes jumping up impossible for the dog. This means that the owner trains the dog to reliably perform a "sit" under highest distractions. then when guests come, the owner gives the cue "sit" when the guests enter the residence. And when the dog sits, he will be rewarded. If this is consistently trained, the dog will stop jumping on house guests.

Clicker Training is a special style of positive reinforcement training. We use a special mechanical devise that issues a click-clack sound as a marker signal, that a reward is on it's way.

Source: Hundeschule Holledau, Sonja Meiburg (please request permission for further distribution and use from the author)

the Marker Signal

One of the most important facets of dog training is the right timing.

I can always reward a dog sitting right in front of me. He gets his treat as soon as his rear starts towards the ground. He understands this and has no problem repeating this in the future.

More difficult are behaviors which are either done in the blink of an eye (such as quick sight-contact with the owner) or behaviors done at a distance (for example a "sit" on the other side of a sidewalk). In such a case, it can be difficult for the dog to understand exactly what is expected and if he's done it correctly or not. As a result, the training can take longer and can be more trying for both the owner and the dog.

In order to make this training easier for both of you, I recommend using a signal that bridges the time-gap between the perfectly executed behavior and the reward for this behavior. this bridge is called a marker signal, because it marks the exact time and behavior as it's executed, making it as plain as colored text in a book.

This marker signal says to the dog "That was super, what you just did. Hang on, your reward is on it's way." and this is always true, every time, like a contract between you and your dog. Every time the marker signal is given, your dog can be sure, that he will get something terrific. It's a good idea to vary the type of reward your dog gets in order to keep it interesting and to elicit an anticipatory reaction from the dog.

I also recommend you use a short, nonsense word that isn't otherwise used in every day life, for example "click", "tack", "gack". Please don't use normal praise-words, like "super" or "good" because these can be too emotionally charged and may be difficult to use in difficult situations, and therefore perhaps not clear to the dog.

Whenever your dog does something correctly, use this signal and give him his reward.

For example: If you call your dog he whips around and comes quickly to you. the very second he turns to come back, you call out "click" (or whatever) and have a piece of terrific cheese ready for him. You called out "click" for that turn-around after your recall cue. The very fact that this turn-around was marked means that he'll be sure to repeat this in future.

In the beginning, you have to see, that the reward comes very quickly after the marker signal so that he understands what is being marked and then rewarded. Later you can wait a couple of seconds.

The more you use this marker word, the happier your dog will react to it.

There is very little that can go wrong with this system. Sure, you might once signal incorrectly such that your dog might understand something incorrectly, but you can easily retrain it correctly, as opposed to behaviors which are learn via punishment.

The only thing you shouldn't do is to use the marker signal itself as a recall signal (for example because "well, the dog always comes back when I use it."). That which you mark, is that which you get. So if your were to mark your dog while he is sniff a deer trail because you want him to come back, what will you have marked and rewarded for? For inspecting the deer trail. And what happens with behaviors that we mark? do you see what I'm talking about?

So to recap: First call the dog by name and when he turns towards you, in order to run towards you, THEN mark THAT behavior with "click" and reward your dog generously. when he arrives.

It could be advantageous to mark your dog sniff deer trail in terms of anti-hunting training, but that's a much more advanced type of training. That is something best discussed personally with your trainer.

Have fun trying this out!

Source: Hundeschule Holledau, Sonja Meiburg (please request permission for further distribution and use from the author)

To the top