Reducing reinforcement

[snowbunny]

Interesting podcast from Denise Fenzi on reducing reinforcement.

https://www.facebook.com/video.php?v=573544973072592

I really like the part about making yourself more than a pez dispenser - if you don't build value in yourself, then your dog will see you just as a food delivery machine, and when the food is no longer available, he will lose interest in you. I know for sure with my lot that things like their heelwork are massively reinforced by me looking at them, smiling and telling them they're amazing. Especially with Willow, who can be a bit stressy about wanting to make sure she's doing it right, it helps her relax and you can see her body loosen just by being told that she has it, without depending on the delivery of the food reinforcer.

 

[Emily_Babbelhund]

Thank you for sharing, Fiona. I will listen to this in bed tonight because - let's face it - to Carbon I AM a human Pez dispenser.

 

[HAH]

What @Emily_Babbelhund said - my bedtime listening too!

 

[VanillaBean]

Interesting and yes I agree. I think in the beginnings it is so easy to be just the treat dispenser as we 'C&T'.
I know OH thinks I am nuts as I talk to Vanilla all the time, but the interaction and bond is very strong. Vanilla will respond with facial changes.

Plenty of times I have walked Vanilla and to get Vanilla to heel I tall to her as get a far better response and neater walk then just dispensing treats. I had plenty of problems in the early days until I started talking to her. Of course when you get a person responding sometimes, you realise you are just slightly nuts. But who cares, I get the attention and engagement I want and Vanilla is happy

 

[Charlie]

I am confused as previously we were advised not to chat to our dogs endlessly and dispense treats now it's the opposite. I just can't keep up x

 

[VanillaBean]

Me personally think it is what ever suits the situation. I know for somethings Vanilla response far better with voice/sound, and others with signals and some just pure treats.
And yes agree it can get very confusing quickly...

 

[snowbunny]

I am confused as previously we were advised not to chat to our dogs endlessly and dispense treats now it's the opposite. I just can't keep up x

It’s not the opposite, it’s about finding a balance. You need to ensure you have a strong history of reinforcement and that means lots of treats until the behaviour is established. But Denise is saying that if all you do is deliver treats without any expression, it’s harder to fade the treats. You don’t want to “chatter endlessly” because you need clarity in your cues, and too much inane babble muddies that.
However, being expressive when your dog does something right helps to build a relationship where you are more than a treat dispenser and it allows you to fade the food when the behaviour is established.

Her focus is competitive sport dogs, so she is thinking specifically of dogs doing obedience, Rally and those sports that may not be naturally reinforcing to the dog. Sports like gundog work have wonderful built-in reinforcement once the dog understands “the rules” because retrieving and hunting are generally naturally reinforcing. Similarly with agility; most dogs love to run the course and find it rewarding in its own right. With IPO, the dogs usually live to bite.

The same often isn’t true for the obedience sports, so it’s harder to find a way to reduce reinforcement without losing the sparkle.

 

[Selina27]

I am confused as previously we were advised not to chat to our dogs endlessly and dispense treats now it's the opposite. I just can't keep up

To me, I think we do what suits us, what works for the partnership we have with our dogs, and what we require of our lives together.

But it's always very interesting to learn how others do things, and the current thinking thanks @snowbunny . For me, it is a reminder really to look at just how and when I deliver the reward, it's very easy to get sloppy. I like the advice to ask yourself "what have you just rewarded?".

 

[Boogie]

Mantrailing is the best - the sparkle is natural and all the reward at the end of the task

With Guide dogs and pups a lot of the work simply becomes habit. They guide round obstacles, they stop at kerbs, they refuse to move if there’s a car, they come to a crossing, they indicate the button etc - just as a matter of course. They are given lots food rewards when training. Once they are fully trained and working then every trip they get rewards, but not many and usually only kibble or carrot.
.

 

[Selina27]

Mantrailing is the best - the sparkle is natural and all the reward at the end of the task

 

[Charlie]

However, being expressive when your dog does something right helps to build a relationship where you are more than a treat dispenser and it allows you to fade the food when the behaviour is established.

I see. Oh I am expressive almost to an embarrassing level which Hattie thrives on. I still have great difficulty fading food. I am concerned if I do that the trained behaviour will diminish. If that happens do I just go back to treating for a while? I don't think I will ever fade Charlie's food rewards as it's taken so long to get where we are so it's not worth the risk to me. He also likes me to be expressive jumping around etc, unfortunately David is not xx

 

[snowbunny]

This video is really talking about fading reinforcers in a specific construct: somewhere where you are expecting a series of behaviours you want to cue (or extended duration on a single behaviour) and you can't reinforce it with food or toys. So, things like an obedience trial. I think it also applies for things like walking nicely on a loose leash over an extended distance, once you have established the behaviour with a shit-tonne of reinforcement.

Or, for example, if you are asking for a sit, once the dog knows the behaviour (in whatever context you're in) then there should be no need to reinforce it every time. If you're working towards that sit being different (straighter, tuck sit, quicker, lower latency etc) then you still need the reinforcer as you're still training the appropriate response. But once the dog consistently does what you are asking, then the reinforcer shouldn't be necessary. Technically, if you never reinforce this behaviour, the behaviour could degrade, but in the "pet dog" world, you use environmental reinforcers a lot, probably without even realising it, and you build loads of secondary reinforcers. For example, I can reinforce a sit by clipping the dog's collar on. That is a secondary reinforcer (something they have learned to associate with good stuff). Or I could reward a sit by giving the dog its dinner. Or by allowing it to pass through the door. Etc etc etc. We don't tend to ask for behaviours without there being some reinforcement available to the dog.
Dogs who are entered into obedience type trials often lose their behaviours in that context, because they learn there is no reinforcement available. The behaviours don't lead to environmental rewards. Dogs who are brilliant in training environments stop responding in trials. Some of this may be stress, may be the handler's clarity diminish in that environment, may be lots of things, but if the dog hasn't learnt to do without his primary reinforcers in the practice environments, then he will quickly learn that the trial environment gives him a bad deal. And if all you are to your dog in the training context is a food dispenser, then when that food dries up, there is no value left.