Clicker Training

[NaturallyWild]

I partially disagree on this point. I don't believe you *always* have to reward in position although yes, sometimes it does help. But for some exercises, such as using a phone book for rear end awareness or 'finding' heel - throwing the food had been a great help. It sets the dog up for an opportunity for another reward. HAving said that - if the dog is offering other positions post click and you have consistently rewarded that, then yes, the dogs will think that part of the behaviour.

I didn't mean that you "have to" reward for position, but that it is an extremely big part of creating the behaviour. Can you not reward for position first, then throw a treat to get that break that you want. Do you find that when you are throwing food away from your dog that after the click there is then an immediate shift in behaviour to "wheres that food gonna be thrown". If you continually throw the food to one position in the room will your dog gravitate towards that area more, even pausing there to wait for more treats? Is this taking some focus away from the specific behaviour you are trying to teach? Sometimes it does, sometimes it doesn't.

[leopuppy04]

Can you not reward for position first, then throw a treat to get that break that you want.

You can - but I prefer to do one or the other - either reward in position or throw the food. Some exercises I won't even reward in position past the initial shaping phases (ie: look the food is on the phonebook... once they know what to do with THAT , the food will get thrown).

Do you find that when you are throwing food away from your dog that after the click there is then an immediate shift in behaviour to "wheres that food gonna be thrown".

Again - yes, I could see this happening if you were to throw food for ever behaviour and every time they do it. For my heelwork 90% of the time, the food is thrown. But if they hear the click they don't pop out of position and quickly try to run for the food or anything (well... unless it has already been thrown )... IMO it's a case (for me) of using both techniques together.

If you continually throw the food to one position in the room will your dog gravitate towards that area more, even pausing there to wait for more treats? Is this taking some focus away from the specific behaviour you are trying to teach? Sometimes it does, sometimes it doesn't

.

Perhaps - but i've never thrown food to the same area of the room really. Always changes. I find food throwing great for doing distance work though as the reward isn't connected to you.... much like throwing a toy for that.

I also find 'food throwing' increases their excitement a little (in my dogs) much like if a toy were involved.

[NaturallyWild]

You can - but I prefer to do one or the other - either reward in position or throw the food. Some exercises I won't even reward in position past the initial shaping phases (ie: look the food is on the phonebook... once they know what to do with THAT , the food will get thrown).

Ok, this makes things clearer for me. For you the throwing of food tends to be used after the behaviour has been learnt (and a cue/command added?), it is more to "fine tune" parts of the behaviour as well as practising and maybe challenging or proofing that behaviour - is this right? I will definitely utilise throwing food after i have it on cue, but it is not so much for "fine tuning" a behaviour. I think that maybe (and i need to do this more myself) the "fine tuning" should be done before adding the cue. Once the cue is added the clicker is far less important as you have already worked on the precision of the behaviour, its more about cementing that cue response through challenges.

[leopuppy04]

Ok, this makes things clearer for me. For you the throwing of food tends to be used after the behaviour has been learnt (and a cue/command added?), it is more to "fine tune" parts of the behaviour as well as practising and maybe challenging or proofing that behaviour - is this right? I will definitely utilise throwing food after i have it on cue, but it is not so much for "fine tuning" a behaviour. I think that maybe (and i need to do this more myself) the "fine tuning" should be done before adding the cue. Once the cue is added the clicker is far less important as you have already worked on the precision of the behaviour, its more about cementing that cue response through challenges.

Closer - but still not quite. You got me thinking though .

Sometimes the cue is still not present.... for example - the phone book - I get them to place their front feet on the phone book and once they are doing that reliably - I will throw the food so that they can come straight back onto it. Initially - if they look at the phonebook I will C&T and place the food on the phone book to build up the association etc. As soon as paws are required to be placed on the phone book - then I will start throwing... so that they can 'reset' themselves.. if that makes sense.

Then there are things like the 'quoit set' where I will throw food immediately - for the same reason that they can 'come back' and try again. They have never had a problem pairing the reward with the quoit set.

Then there are things such as sits and drops and positions while heeling where I will reward in position.

Things such as heeling - I may initially reward in position and later throw the food for the opportunity to come back and 'find' that position. Before throwing food in heeling, should they ever lose position, they weren't able to 'find' that spot again as well as they can now.... plus - their actual position itself is much more precise IMO.

So I guess (after reading back what i've just written)... for a position that I want to remain stationary, it is likely that I will reward in position until a cue is built. For many other things, where movement is required, I will throw food either to get them to 'find' the spot or have another chance immediately at 'finding' the reward again or to build up drive (eg: weaving in agility.... but this of course is already on cue).....

Does this make more sense?

I like these topics - they get me thinking

[PAX]

I meant to write in my reply that I do reward in position. Then after reading LP's post I realised that often I don't, like the phone book and other things but that would be after they have a grasp of what got the click.

I also don't think that the click ends the behavoiur. Sometimes I will click and delay the reward and my dog will repeat the behaviour so in my mind the click hasn't ended it.

LP you always start these thought provoking topics.

[PAX]

I always click and treat. However if I click by mistake I don't treat. I used to but after watching Bob Bailey DVD I don't anymore. If I click my mistake I go any bang my head on the wall

I don't remember that bit, I think you should reward and then bang your head on the wall.

Glad you liked the DVDs.

[helen]

I love clicker training. I use the every click earns a reward and end of behaviour as it has worked like a dream for me

[NaturallyWild]

Rewarding for position doesn't need to be a stationary thing. For example going over a jump you would reward by throwing out away from the jump because you want them to drive over the jump, ie that distance position is part of the behaviour that you want. I think that if you were to think about where you reward for each thing you are teaching it would speed up the learning and remove parts of the behaviour that are unwanted and need to be removed later. Yes there are many times where it doesn't matter so much with the position but it will always have some effect.

With the quoit set going in and back from the pole is part of the behaviour so throwing the food can be useful. I would possibly throw towards where the other quoits are - but this would be further along the teaching phase

[bridgie_cat]

LP- why do you think the clicker gets faster results, arouses more enthusiasm etc than any other marker? Is it timing? (i find timing easier with voice, you with a clicker- no right no wrong but i don't find the clicker teaches anything faster or gets the dogs more worked up) I struggle with the concept of the clicker vs voice and why the clicker would be better- except for people whose timing improves with the clicker.

Personally, the clicker has more set rules for ME - so it is a more consistent thing for Lana - I have a release word "OK", which has a consistent consequence so she understands this perfectly. Had I used "YES" like the clicker - she would (I assume) understand it also and it would work like the clicker...

The benefit of the clicker over a definite YES, in my mind, is that it sounds VERY different to anything we are going to say - and for someone who talks a lot in training, it is just easier for the dog to discern! Also, it is a shorter, sharper sound & perhaps easier to time for some (not that YES is hard!) - so just a slightly clearer maker... downside = another thing to carry!

[laffi]

I always click and treat. However if I click by mistake I don't treat. I used to but after watching Bob Bailey DVD I don't anymore. If I click my mistake I go any bang my head on the wall

I don't remember that bit, I think you should reward and then bang your head on the wall.

Glad you liked the DVDs.

Hmm I think it was on the 2nd or 3nd DVD. He was talking about the chicken workshops and how people would always reward the click even if they clicked by mistake. However he did say that for the dog training you should really take these mistakes seriously and therefore no reward for the dog and for yourself because you should really try not to make the mistakes. Mistakes slow down your dog, confuse him and you should be aviod them at all costs rather than just say "ooops" and treat (according to Bob).

[Sayreovi]

I use clickers, have for about 2-3 years now.

For my dogs the click means reward and end of behaviour. The click means 'yes thats exactly what i want' anything the dog does afterwards doesnt count (and if anything it would be reinforcing coming to me)

For example i am teaching my dogs box work for obedience, so when i initially started i clicked steps closer to the box, dogs immeditately returned to me for the reward and we started again. After awhile they got to the box and i stopped rewarding movement towards the box only them being in the box. Once i clicked they returned immediately.

I dont have any jumps at home yet so havent progressed to that but it will be in the same way, they return to me when i click and we start again.

Prefer to use clickers because one dog is a nut about food and i can get what i want by just clicking as i may not be fast enough saying yes. Plus i wont confuse the dogs if i may be late (which is a high possibility!)

First dog is a crossover boy and is an amazing worker now, i just have to control my nerves and my second dog was raised pretty much with hardly any corrections and with a clicker.

Neither dog throws behaviours at me, they will both wait until i say something then try their best to find out what results in a click!!

ETA: Ive just finished reading Morgan Spectors book so have all these new ideas that i want to try out. Especially my stays! Ive stopped using stays and just say sit or drop and i actually did an out of sight 'stay' with them both (Nova was tied up though best to be safe) while i ran to the car to get more things. Any other circumstances Darcy would have been up and flying after me!

Edited September 13, 2007 by tollersowned

[NaturallyWild]

For my dogs the click means reward and end of behaviour. The click means 'yes thats exactly what i want' anything the dog does afterwards doesnt count (and if anything it would be reinforcing coming to me)

According to who does it not count - thats the thing you need to be aware of. Like i mentioned with the example of the jump and throwing food forward what happens after the click can, and does effect there behaviour. If the food was given from you at the side of the jump you will get a dog that jumps a bit slower and at angles to turn once over to get to the reward.

The primary reinforcer (eg food) should be much more important than the secondary (marker), so there are going to be effects from how the primary is given.

[Arya]

It is my belief that it is kind of 'horses for courses'. Some dogs are more suited to clicker and some owners/handlers/trainers too. This is what I have found. Brilliant for some people and their dogs. Others, a very poor second to voice. For me, I don't use clicker at all for my dog. BUT (and don't laugh, this is not meant to be derogatory or anything) I do use it when training my cat. Works amazingly with the cat, who requires a quiet, still, less body language environment for training. The dog, however, totally needs the feedback to come from me. It is so much more than just the marker word. It's how it's said, it's how I hold my body and my face when I mark the moment, it's whether I am releasing her from the excercise or merely reinforcing a good behaviour in a chain of behaviours. So it might be 'yes', 'yes', 'YES' - and unconditioned reinforcer (hose or food) on the final 'yes'. My body and tone of voice gives away what's going on to the dog. Show me a clicker where you can change tone like the human voice. You can't. Show me a clicker that exhibits facial expression, which tends to sometimes lack in people who are not using clicker to its best ability by incorporating their body signals too. Not possible. So the refinements of the human voice to me are very important in communicating the conditioned reinforcer to the dog. It's too complex for a mechanical implement the way I use the marker. BUT, like I said, I have seen people use clicker and achieve results they just weren't able to achieve with their voice. Different for different trainers/dogs/other animals.

There you go, that's my long winded two cents worth

Edited September 13, 2007 by Arya

[Guest Steph & Bam]

I love my clicker! I found that the standard Tab Clicker was too harsh for Bam and he used to jump at the sound. I have switched to two kinds. One is an iClick used solely at home, the other is from the same people who make the iClick, but is volume adjustable. I use that when I'm at an obedience class. It's a unique sound and Bam responds brilliantly to it. I tried using my voice, but was a little off with my timing, so Bam was getting confused. When I started with the clicker (click = food/toy) he started giving me instant response, and started offering behaviour to figure out what I wanted from him. I am more than happy with the way he works and can't wait to teach him some more things. I'm joining a new obedience club soon, and can't wait to try something new

[leopuppy04]

Isn't it interesting that even within one training style, there are so many interpretations for it. Some of which I think are both correct.....

I think that if you were to think about where you reward for each thing you are teaching it would speed up the learning and remove parts of the behaviour that are unwanted and need to be removed later. Yes there are many times where it doesn't matter so much with the position but it will always have some effect.

My primary example for this would be 'heelwork' where position is vitally important. I've had to think a bit for this and yes, I think I did reward position first (ie: this is the spot for heeling) and later resorted to food throwing. Same for introduction of turns etc. So I guess the times that I throw food from the beginning would be - as you say - parts where the drive might be part of the behaviour. I was thinking about it all afternoon and I do think that in the initial phases they are rewarded in position first. But having said that - I still like the 'food throwing' and see it as a great tool for my training as the dog has the opportunity to come back into position for reward etc. The points I see where this does become an issue for example - would be in a situation where you aren't throwing the food (ie: dog is still rewarded immediately and is 'retrieving' the reward, similar to that or a thrown toy) but rather fumbling for the food in which the dog has already sat again or walked away or dropped.... so the food would be connected to *that* behaviour. In my food throwing instances... the dog has offered me nothing else to get confused other than to 'retrieve' the food the same as they would a toy.... does that make any sense?

With the quoit set going in and back from the pole is part of the behaviour so throwing the food can be useful. I would possibly throw towards where the other quoits are - but this would be further along the teaching phase

Well the way I taught it - the ring was initially on the pole and naturally the behaviour ends with the ring on the pole..... so in that instance they are running 'away' from the rings etc. When we are on multiple quoits they are still only getting the reward at the end.... but I do get what you are saying - throw the food and then they see the ring and pick it up 'on the way back'....

According to who does it not count - thats the thing you need to be aware of. Like i mentioned with the example of the jump and throwing food forward what happens after the click can, and does effect there behaviour. If the food was given from you at the side of the jump you will get a dog that jumps a bit slower and at angles to turn once over to get to the reward.

This makes sense - Think of the dogs that cut the 'broad jump' in obedience - simply because the reward comes from the handler.... so they cut it to get to the reward faster. Throw the food out and the dog will jump straight first. Another example would be my girl 'sitting' every time after she heard the click for 'stand'.... result was a 'stand' then sit immediately after the click.

It is my belief that it is kind of 'horses for courses'. Some dogs are more suited to clicker and some owners/handlers/trainers too. This is what I have found. Brilliant for some people and their dogs. Others, a very poor second to voice.

Totally agree with that Arya - what works for one - doesn't necessarily work for another

Edited September 13, 2007 by leopuppy04

[Sayreovi]

For my dogs the click means reward and end of behaviour. The click means 'yes thats exactly what i want' anything the dog does afterwards doesnt count (and if anything it would be reinforcing coming to me)

According to who does it not count - thats the thing you need to be aware of. Like i mentioned with the example of the jump and throwing food forward what happens after the click can, and does effect there behaviour. If the food was given from you at the side of the jump you will get a dog that jumps a bit slower and at angles to turn once over to get to the reward.

The primary reinforcer (eg food) should be much more important than the secondary (marker), so there are going to be effects from how the primary is given.

The dog is working for the click, the click indicates the correct behaviour and means reward come and get it.

If the dog did a detour on the way like did a wee before getting his treat it certainly doesnt mean i rewarded the wee otherwise wouldnt it happen all the time unlike what i actually clicked for?

For my dogs the click ends the behaviour. They break out of whatever position once the click has been heard because they know the behaviour is finished, what they did was correct and so they get a treat.

I dont throw food for them because in a way getting the food of me is also secondary indication that its finished and we can start again.

Sorry i dont quite understand some of your examples.....are you talking about obedience or agility in regards to the jump?

[Sayreovi]

According to who does it not count - thats the thing you need to be aware of. Like i mentioned with the example of the jump and throwing food forward what happens after the click can, and does effect there behaviour. If the food was given from you at the side of the jump you will get a dog that jumps a bit slower and at angles to turn once over to get to the reward.

This makes sense - Think of the dogs that cut the 'broad jump' in obedience - simply because the reward comes from the handler.... so they cut it to get to the reward faster. Throw the food out and the dog will jump straight first. Another example would be my girl 'sitting' every time after she heard the click for 'stand'.... result was a 'stand' then sit immediately after the click.

Wouldnt that be a training problem rather then a rewarding problem?

Dog wasnt trained to go out far enough or clear the jump properly. If the dog ended up cutting the corner the front wouldnt be straight thus wouldnt get rewarded because it wasnt the correct way.

[bridgie_cat]

According to who does it not count - thats the thing you need to be aware of.

must agree here - not so much for a random behaviour after the click (like you peeing eg) - but if the click ends the behaviour and in coming to get the treat they are constantly doing something (like sitting or a particular way of leaving the obstacle) - whilst the click is saying YES that is right - the treat is the main reinforcer and could be saying YES that bit was also right! ... depends if it becomes a habit - I am not so proficient with the system/havent been going long enough that I have experienced this - but understand the theory... more so if it is a habit formed during the training...

the skewed jumping I would not think is a learned trick from the clicker - more to do with the direction the dog is coming from and where it thinks it is going - OR if it is watching the handler - thus jumps a bit skewed...

[Sayreovi]

According to who does it not count - thats the thing you need to be aware of.

must agree here - not so much for a random behaviour after the click (like you peeing eg) - but if the click ends the behaviour and in coming to get the treat they are constantly doing something (like sitting or a particular way of leaving the obstacle) - whilst the click is saying YES that is right - the treat is the main reinforcer and could be saying YES that bit was also right! ... depends if it becomes a habit - I am not so proficient with the system/havent been going long enough that I have experienced this - but understand the theory... more so if it is a habit formed during the training...

.....and if the way they are sitting/leaving the area is correct then i have nothing to worry about.

[leopuppy04]

Wouldn't that be a training problem rather then a rewarding problem?

Dog wasnt trained to go out far enough or clear the jump properly. If the dog ended up cutting the corner the front wouldnt be straight thus wouldnt get rewarded because it wasnt the correct way.

Not always - sometimes it is the 'anticipation' of the reward, and if that reward is consistent enough - they are 'reinforced' for the 'cutting' of the broad or the constant sitting after a stand (because they have moved from the initial position). Keep in mind, this isn't solely for clicker training only - it happens with verbal markers too