Errorless Learning – A good thing?

Errorless Learning – A good thing?

I see a lot of dog trainers promoting errorless learning, this is where we set up a training session where the problem-solving steps are so small that the dog can easily achieve each next level without too much effort and so they do not make any mistakes when learning a task.

On the surface this seems like a great idea and will keep frustration levels very low for the learner.

At the start of creating a training/learning relationship with a dog I think that this is a great ideal to strive for. Setting up the dog to have success easily and frequently helps the dog want to engage in training with their handler.

However, if we only train this way then I feel that we can create problems in the long term.

Dogs who never make mistakes, never learn how to deal with mistakes. Also, handlers who think that mistakes are a problem become very stressed and anxious when their dog makes a mistake.

This can lead to training becoming an unpleasant experience for both dog and handler, all because of the pressure to create errorless learning.

As with everything in life and dog training, one size does not fit all and one style of training isn’t right for educating everything that a dog needs to learn.

Do I use Errorless learning?

Probably not – I am not that sort of person. I use ‘As error free as possible under the circumstances’ training in the early stages of starting a dog on their way to understanding something new.

I absolutely agree that in order to at least give the dog a fighting chance of understanding what it is that we would like them to do it is only fair to set them up with as few error possibilities available.

I will use physical barriers to block incorrect choices, leads, luring and also shaping to start the process off and help the dog to understand what it is that I am trying to show them.

Do I worry about Errors?

Absolutely not 😊

I embrace errors as feedback. The dog is telling me that they do not have enough information to do the task yet. The errors inform both me and the dog about how the information is travelling between both of us.

I want the dog to learn that the worst that happens in a training session is nothing. OK they didn’t earn an anticipated reward, but nothing bad happens. We just reset and try again.

Training should be a conversation.

If a dog never makes a mistake, then they do not build their emotional resilience. They never learn how to deal with the emotions, sometimes big emotions, that come along with failure and frustration.

If a dog doesn’t learn how to communicate to their handler things like: –
• This is too hard
• I don’t understand
• I have run out of concentration
• I don’t want to do this
• I want to stop now
• I would prefer to solve this problem a different way
• This way is better for me
• I don’t want to train today
• This is too exciting

Then we are effectively removing choice from the dog. We can be coercive even when using reinforcement.
Learning to listen to our dogs and watch how they respond to our questions is very illuminating. If we are so hung up on errors and mistakes being an issue we can miss the bigger picture.

A classic that I see with many dog owners is that they have been brainwashed into thinking that unless a recall is instant and happens on the first cue then is an error/failure.

In reality there are so many ways that a recall can be a success even if the handler has to repeat the cue and it seems to take forever for the dog to respond.

A dog sees a distraction in the distance, their brain is focused on the distraction, their ears have gone into standby mode because their brain is using up a lot of energy processing the information coming to them from the distraction.
The handler calls the dog. Observing from the outside it can look like nothing has happened. The dog doesn’t turn towards the handler or seem to acknowledge the recall cue.

Internally though the brain has switched the ears back on, so that they can process the audible information coming from the handler.

If, at that point the handler decides that the dog has failed to respond then they tend towards punishing the failure. Shouting at the dog, potentially running over to them and putting them on the lead as a way to show the dog that they failed.

The dog only registers the negative response because they are only just starting to process the information once the handler has gone into – the dog is wrong mode.

However, if once the dog has switched their ears on again, the handler recues a recall, the chances are that the dog will at least reorientate to the handler in recognition that they are saying something. Depending upon the level of the distraction, the dog may start back to the handler or may only be able to stand and look at the handler.

Each of those responses give the handler feedback as to how much bandwidth the dog has available to process the information coming from the handler verses the distraction.

If the dog is able to look but not return, they are probably asking for more help, they need more input from the handler to finish the task.

Sometimes running away at that point can trigger movement from the dog, other times they need the handler to go to them to help them move away from the distraction.

Changing our response to errors from, something that needs to be eliminated at all costs; to my dog is giving me information about how much help they need right now, means that we are truly having a conversation. We can give the dogs the help they need rather than focusing on what went wrong.

There is nothing wrong with errors, they are just feedback.

Absolutely there are time when we need flawless performances from our dogs, especially when they are working or competing. However, how we get there effects the emotional state of the dog when they are asked to perform.
All I ever want from my dog and myself is to do our best on the day. Today’s best can look very different from tomorrow’s best and that is OK.

Dogs who are allowed to fail during learning become much more robust.

Dogs who are allowed to learn through failure alongside success do not become so over aroused by working. They seem to be more thoughtful and deliberate in their work because reinforcement is not guaranteed and their anticipation doesn’t overtake their ability to think.

They learn to listen to the handler rather than just guess. Busyness isn’t the answer, thoughtfulness is. They become much more efficient in their learning.

How much failure is too much?

This is very much determined by each dog. Some dogs need to have mostly success as they are not very confident learners. Some dogs need to fail more because they are overconfident in their learning.

This again goes back to handlers having a good relationship and understanding of their individual dogs. Tailoring the amount of success and failure to each individual in order to balance confidence and thoughtfulness.
As always, it is the Goldilocks principle – not too much, not too little, just enough.

The other thing to note is that it is OK for handlers to make errors too. Part of learning about a new dog and how best to help them, means that we will sometimes get things wrong, misjudge a situation with our dogs. That is OK, be kind to yourself. Pay attention to your mistake, learn from it and move on.

Embrace the mess, because none of us are perfect, and to be honest that is what makes life fun 😊

 

These are a couple of blogs that reference the scientific research on Errorless learning.

“Errorless” Learning

Errorless Learning II