Clicker Training, A Good Idea?

[Aidan3]

Aidan can you explain why you feel a clicker is more effective than a verbal marker when conditioned the same way?

I have one client who has an absolutely consistent and untainted "yesss", but by and large I feel like I'm pushing uphill to get someone to be as consistent as they need to be :laugh: In my experience the clicker is faster and more precise. I aim for a rate of reinforcement of 30 clicks/minute and adjust criteria to maintain that, it's hard to keep up (and tiring) if you're using a verbal marker. When you're no longer shaping, your rate of reinforcement is lower so it's easy to use a verbal bridge.

It's plausible that a non-verbal sound can be more powerful because it's easily recognised and takes a quicker route in the dog's brain, it doesn't require cortical processing. There is a theoretical underpinning beneath that comment, but as far as I know no-one has set up an experiment to see if this translates to a practical benefit.

Similarly, we can respond faster by pressing a button than articulating a sound. That has been tested.

You'll always find some people who can't walk and chew gum at the same time and a verbal bridge, slower rate of reinforcement, lower criteria for the dog is better for those people.

[Aidan3]

Because nothing I could offer is even close to as rewarding as a gallop or a chase.

Studies of choice behaviour have consistently and repeatedly demonstrated that we can trump a big reward with lots of little rewards. So you don't necessarily need a better reward (or stronger negative reinforcer), but yes, overcoming a sighthound's drive to chase takes a lot more work!

[raineth]

Aidan can you explain why you feel a clicker is more effective than a verbal marker when conditioned the same way?

I have one client who has an absolutely consistent and untainted "yesss", but by and large I feel like I'm pushing uphill to get someone to be as consistent as they need to be :laugh: In my experience the clicker is faster and more precise. I aim for a rate of reinforcement of 30 clicks/minute and adjust criteria to maintain that, it's hard to keep up (and tiring) if you're using a verbal marker. When you're no longer shaping, your rate of reinforcement is lower so it's easy to use a verbal bridge.

It's plausible that a non-verbal sound can be more powerful because it's easily recognised and takes a quicker route in the dog's brain, it doesn't require cortical processing. There is a theoretical underpinning beneath that comment, but as far as I know no-one has set up an experiment to see if this translates to a practical benefit.

Similarly, we can respond faster by pressing a button than articulating a sound. That has been tested.

You'll always find some people who can't walk and chew gum at the same time and a verbal bridge, slower rate of reinforcement, lower criteria for the dog is better for those people.

haha, I think that last sentence describes me :laugh: oh boy, my poor dogs!

Actually what I am most unco at is not the timing but the speedy treat delivery, I need to improve on that for sure.

[raineth]

Maybe I'm having a really slow day but it only just clicked () that +R = Positive Reinforcement.

:laugh: I knew the +R but I can't think what NRM is referring to

[raineth]

Because nothing I could offer is even close to as rewarding as a gallop or a chase.

Studies of choice behaviour have consistently and repeatedly demonstrated that we can trump a big reward with lots of little rewards. So you don't necessarily need a better reward (or stronger negative reinforcer), but yes, overcoming a sighthound's drive to chase takes a lot more work!

I do find however that a Jack Pot sometimes gets through really well where smaller treats might not be quite doing it.

When training Digby not to eat a person with a camera, I used Jack pot reward (1 repetition per day, and 1 Jack pot) and it seemed to work exceptionally well.

oops three posts in a row

[Kavik]

Maybe I'm having a really slow day but it only just clicked () that +R = Positive Reinforcement.

:laugh: I knew the +R but I can't think what NRM is referring to

NRM = Non Reward Marker. Something unemotional like "Oops" which lets the dog know that it was not the action we were looking for and to try again. You ideally want the dog to remain upbeat, not to see it as a punisher. The only time I've really used a NRM was for weave pole training when working on entries, as I didn't want him to continue weaving if he chose the incorrect entry.

[raineth]

Maybe I'm having a really slow day but it only just clicked () that +R = Positive Reinforcement.

:laugh: I knew the +R but I can't think what NRM is referring to

NRM = Non Reward Marker. Something unemotional like "Oops" which lets the dog know that it was not the action we were looking for and to try again. You ideally want the dog to remain upbeat, not to see it as a punisher. The only time I've really used a NRM was for weave pole training when working on entries, as I didn't want him to continue weaving if he chose the incorrect entry.

Thanks Kavik

I have used them with Del before when she sometimes would revert to standing on the ottoman during free-shaping :laugh:

[Aidan3]

oops three posts in a row

Slow down! :laugh:

[Aidan3]

Because nothing I could offer is even close to as rewarding as a gallop or a chase.

Studies of choice behaviour have consistently and repeatedly demonstrated that we can trump a big reward with lots of little rewards. So you don't necessarily need a better reward (or stronger negative reinforcer), but yes, overcoming a sighthound's drive to chase takes a lot more work!

I do find however that a Jack Pot sometimes gets through really well where smaller treats might not be quite doing it.

When training Digby not to eat a person with a camera, I used Jack pot reward (1 repetition per day, and 1 Jack pot) and it seemed to work exceptionally well.

Classical conditioning?

[sheena]

I have trained classes of up to 10 people all training with clickers. The dogs were focussed on their handler and I do not remember any particular dog or occassion where the dog worked off other peoples clicks or was distracted by them or if they were it was brief and and easy to refocus them.

I use marker words and the clicker however I find the clicker far more precise than a word so that is why I use it, particularly when teaching something or refining something by degrees. It is also a consistant noise and not emotional such as the voice can be.

Times too....exactly. I use a clicker to teach something new but then when I have the behaviour I will mostly use a marker word, eg. "yes" for entering the weave poles correctly. But for teaching (shaping) a new behaviour, the clicker is a much more effective marker of a precise tiny movement, glance, whatever, than a word & unlike a word, the clicker tone never changes. The word "yes" can be said in so many different ways & tones.

[sheena]

Maybe I'm having a really slow day but it only just clicked () that +R = Positive Reinforcement.

:laugh: I knew the +R but I can't think what NRM is referring to

You don't need a speedy treat delivery. It is the clicker that is marking not the delivery of the treat. When I am using the clicker, generally, I will click & then throw the treat away, so that the dog has to run & get the treat & then think back to what it did to earn it & it is amazing to see how quick they pick this up & it really gets them thinking. That way it can't be said that I am luring, which is what happening when people click, then throw the reward closer to the goal.

[sheena]

You don't use NRM's when using the clicker, in fact, you say nothing at all, just have a happy smiley face. When I finally see the light bulb go off in my dog's head & he reaches his goal, I will click, at the same time break my silence with a "yes" & shower him with whatever treats I have left over....then end of game. :) For example, teaching my young guy to crawl under the coffee table. I will click & treat for glancing at it, then glancing under it, then moving a paw, then bending etc etc. all the time throwing the treat slightly away, so he has to think where he was up to when he got the click :laugh: then when he finally does the whole lot its party time After a couple of successes I might put a cue on it if it's something I want. (if that makes sense)

[raineth]

Maybe I'm having a really slow day but it only just clicked () that +R = Positive Reinforcement.

:laugh: I knew the +R but I can't think what NRM is referring to

You don't need a speedy treat delivery. It is the clicker that is marking not the delivery of the treat. When I am using the clicker, generally, I will click & then throw the treat away, so that the dog has to run & get the treat & then think back to what it did to earn it & it is amazing to see how quick they pick this up & it really gets them thinking. That way it can't be said that I am luring, which is what happening when people click, then throw the reward closer to the goal.

Hi Sheena :) I was referring to this quote from Aiden:

I have one client who has an absolutely consistent and untainted "yesss", but by and large I feel like I'm pushing uphill to get someone to be as consistent as they need to be :laugh: In my experience the clicker is faster and more precise. I aim for a rate of reinforcement of 30 clicks/minute and adjust criteria to maintain that, it's hard to keep up (and tiring) if you're using a verbal marker. When you're no longer shaping, your rate of reinforcement is lower so it's easy to use a verbal bridge.

I think in Behaviour Mod type situations a very high rate of reinforcement can really help with success :)

I also often deliver treats by tossing them. For my boy it is one of the things that helps him be more motivated as well as resetting him :)

[raineth]

You don't use NRM's when using the clicker, in fact, you say nothing at all, just have a happy smiley face. When I finally see the light bulb go off in my dog's head & he reaches his goal, I will click, at the same time break my silence with a "yes" & shower him with whatever treats I have left over....then end of game. :) For example, teaching my young guy to crawl under the coffee table. I will click & treat for glancing at it, then glancing under it, then moving a paw, then bending etc etc. all the time throwing the treat slightly away, so he has to think where he was up to when he got the click :laugh: then when he finally does the whole lot its party time After a couple of successes I might put a cue on it if it's something I want. (if that makes sense)

Yep that sounds good :)

I only use the NRM as I said for when I've forgotten to hide the ottoman during a training session. I don't want her on it as she is very big and could easily do herself and injury when she's trying to be a circus pony atop a narrow ottoman half a metre above the floor. A NRM encourages her to get off quickly and calmly rather than jumping off like loon :)

[Cosmolo]

I don't think the variation in a verbal markers sound is a big deal to be honest. But i recognise some of the factors you describe as being important Aidan, thanks. I find your rate of reinforcement really interesting- i thought mine was high but it's not that high!

I have used a NRM with a more experienced dog when shaping when they have looked like getting stuck on the wrong thing. Never de-motivating but i find it is directional- it encourages them to try something else.

[huski]

I actually think NRMs should be motivational to the dog (it is to mine) - it means they have an opportunity to try again. I don't think that means there is no stress involved as think there is stress involved in all learning, and I think there's a bit of frustration created in withholding (not just removing) a reward.

I use NRMs in my training and I find it very effective. As with anything if you've conditioned it properly it can be very useful.

[Podengo]

I find Elsie tries harder when I use NRMs, she throws herself into whatever we're doing with more enthusiasm. I think having that direction in training is good. I like it too for shaping, especially in the house, as I can stop her from trying destructive or silly things out (e.g. biting/grabbing things, or cocking her leg on stuff... Dumb trick I taught while drunken shaping :laugh: )

[huski]

I find Elsie tries harder when I use NRMs, she throws herself into whatever we're doing with more enthusiasm.

Yep that's my experience too :)

[ness]

If I use a NRM with my youngster she back chats and tells me exactly what she thinks but agree then tries extra hard :laugh:.

Edited October 1, 2013 by ness

[huski]

If I use a NRM with my youngster she back chats and tells me exactly what she thinks but agree then tries extra hard :laugh:.

Yep and if you're a Malinois sometimes you get a 'hurry up then' nip or teeth snapping :p

Edited October 1, 2013 by huski