Click, Click, Click ... Claaaaaaaaaaaaack.

[d112358]

This is the best thread.

 

I love it when smart people disagree politely Best learning environment that exists.

 

I think we mostly all use "clickers" - some of us just use the mechanical kind, others use the verbal kind.

 

In that regard, yes, they act like any other tool.

 

And in all things where rote learning is acceptable the clicker works as well as any other kind of marker. MARKERS are not used so much in advanced "vocational" training because you can't mark a decision or the application of an instinct/sense in the same way. Much like they can't be used, as noted, in advanced agility training.

 

I think that will be the conclusion I take away.

 

Great discussion.

 

I think it depends on the application and exactly what you would classify as 'rote'. Shaping is all about marking an independent action the dog gives you that approximates what you want, so even though there is repetition involved in the activity, the dog is using a lot of mental energy trying behaviors to puzzle it out. My shelter volunteer training involved clicker training a partner in 'stupid human tricks', like rubbing your belly or patting your head, just by freeshaping. It's an interesting experience to be on the other end of the partnership; you have to keep offering behaviors and your comprehension of what the behavior entails slowly coalesces into the actual behavior - it doesn't feel like memorization.

 

In terms of agility, I use the clicker not only on contacts but also for weave entries or combating 10th-pole-popout. A well-timed click can be more informative and produce better results than 5 reps using a verbal marker.

 

I start with clickers for a few sessions by just charging them - click then treat. One of the biggest new handler mistakes I see is that they want to move on before the dog has that association down cold, and then the clicker becomes less useful as a tool. I actually think it makes treating easier timing-wise to use a clicker, because once the clicker is charged you can click then reach for a treat and not have to deliver it simultaneously, just within a few seconds. It's been easier for my dogs to focus when the treats aren't out in front of their nose waiting for perfectly-timed delivery.

[alligande]

To clarify I said the clickers in agility had limited use once you moved on from basic techniques, I did not specify any detail as I did not want to write an essay, but I fully agree clickers or a verbal marker have a role to play in establishing and if nessacery reiterating fundamental skills such as contacts, weaves and good jumping form. Where they become hard to manage is when you are working sequences together.

I am sure my next dog I will use a clicker rather than a verbal marker as I no longer have a sound sensitive dog, who thought the noice in all it's forms terrifying.

[d112358]

To clarify I said the clickers in agility had limited use once you moved on from basic techniques, I did not specify any detail as I did not want to write an essay, but I fully agree clickers or a verbal marker have a role to play in establishing and if nessacery reiterating fundamental skills such as contacts, weaves and good jumping form. Where they become hard to manage is when you are working sequences together.

I am sure my next dog I will use a clicker rather than a verbal marker as I no longer have a sound sensitive dog, who thought the noice in all it's forms terrifying.

 

I used it when I was proofing the Ketschker turn (in sequence), so it's not only for stuff that's super-basic.

[Stanzi_and_Molly]

He is very noise sensitive. I've never seen a dog who hated noise as much as this dog. He is as noise sensitive as I am. We have low flying jets from Tidewater which bother him as much as the chimes from the college-on-the-hill, any siren he might hear, or the Great Dane across the street. The week the gas company tore up the street, to replace a line really messed with him. My vet who has BCs said she thought the stress brought on some of his symptoms while I had him on the first deworming regime.

 

He is food motivated. I think I will try tongue clicking with some cheese. I could give him Mr. Barky treats. I don't want to use anything with preservatives. It is simpler for me to tongue click as opposed to having a wrist clicker. I have two clickers, but he didn't like either one, not at all.

What Mum24dog says is correct. The clicker without a reward is meaningless to the dog, and to a noise sensitive dog could be torture.

 

You have to pair the clicker to a reward. Food is what's used most often, but it doesn't have to be that if the dog isn't food motivated. The dog first has to be taught what the clicker means.

 

This is why I suggested in another thread that you seek out a good positive reinforcement trainer so that you can learn the proper techniques.

 

I also agree with Root Beer that you may want to try a quieter marker such as Yes! or a soft tongue click for this dog at first so you're not stressing him with a sound he's afraid of. I once had a dog who was terrified of clickers, and I had to use another marker . . . but I charged the verbal marker like people charge clickers first, meaning that I just sat with her and said Yes! and gave her a very tasty treat over and over again until she knew that my saying Yes! meant she was going to get a treat. Once the dog knows that, you mark the desired behavior and then can follow up with the treat a moment later if it's not already ready in your hand.

 

Good luck.

[alligande]

 

I used it when I was proofing the Ketschker turn (in sequence), so it's not only for stuff that's super-basic.

This is an enquiring minds question, at what point did you click, when using a Ketschker or for that matter most turns I find it is 98% my movement and positioning that affects the quality of the turn, so when using a clicker where in the turn are you clicking? I would not have a clue how to use a clicker for something like this, unless my trainer had it for clicking me, as Kristine described earlier, as I make far more mistakes than my dog, his timing is much better.

[d112358]

This is an enquiring minds question, at what point did you click, when using a Ketschker or for that matter most turns I find it is 98% my movement and positioning that affects the quality of the turn, so when using a clicker where in the turn are you clicking? I would not have a clue how to use a clicker for something like this, unless my trainer had it for clicking me, as Kristine described earlier, as I make far more mistakes than my dog, his timing is much better.

 

It depends on the dog. If they're having trouble committing to the jump, reading too much pressure (with me in the correct position), I'd click for driving to the jump and toss a treat. Starting this turn with a dog that's not used to blind crosses, I set up at a very forgiving angle and get a second person to click a moment after the dog lands for a correct direction head turn - this sets them up to drive into the new hand and avoid backjumping. Next step is delaying the click until the dog clears the wing and proofing at crazy angles and in sequence.

[juliepoudrier]

But the above description seems to me, a non clicker trainer, as clicking specific parts of the sequence and not the whole sequence, which I think the difference in semantics that is catching people up, You're still not training a sequence with one single click to mark the whole sequence, which is what I think Alligande was referring to (correct me if I'm wrong).

 

Never thought about tongue clicking. I guess that's what I do when I "cluck" to a horse.

 

J.

[alligande]

But the above description seems to me, a non clicker trainer, as clicking specific parts of the sequence and not the whole sequence, which I think the difference in semantics that is catching people up, You're still not training a sequence with one single click to mark the whole sequence, which is what I think Alligande was referring to (correct me if I'm wrong).

 

Never thought about tongue clicking. I guess that's what I do when I "cluck" to a horse.

 

J.

Julie you have nailed my thoughts...

To be honest I would not approach training agility as you do, before I attempt the more esoteric turns in agility I would already expect my dog to understand his job, which is to drive over jumps and read my body language, which is all foundation work that I do agree is compatible with clicker training. I have spent many hours watching different levels of competitors running agility while ring working and in my opinion most mistakes are made by the humans, bad timing, bad choices on the path they chose, most of the dogs understood their jobs just got the wrong signals which is why I feel the emphasis at a certain point in agility needs to switch from the dog to the human.

Julie the tongue click is exactly what I used to do when riding, I never managed to use it for clicker training but I do use it for recall/attention I can't whistle and I have found it a great way to communicate without having to call his name.

[Root Beer]

But the above description seems to me, a non clicker trainer, as clicking specific parts of the sequence and not the whole sequence, which I think the difference in semantics that is catching people up, You're still not training a sequence with one single click to mark the whole sequence, which is what I think Alligande was referring to (correct me if I'm wrong).

 

Weeelll . . . you can. I do this from time to time.

 

Take backing, for example. At first I am going to click the dog just for offering a single foot movement backward. I will progress to clicking a full step. Then two. I don't click and treat every single step backward - duration of movement is gradually built. The dog learns to back (which is a sequence of steps) until the click is heard.

 

Eventually I will ask the dog to "wait" instead of clicking and then click the wait and build duration onto the wait. Now the sequence is baaaaaaaaaaack - wait - click treat. The click/treat covers the whole sequence, not just the duration on the backing.

 

Now I might add a distance move to the sequence. Baaaaaaack - wait - spin - click/treat. Again, the click/treat is for the sequence, not just the spin at the end. In practice, the dog understands this. Eventually I will drop the click from the whole sequence.

 

If I were using a clicker on a Ketschker in sequence (I probably wouldn't, but if I were going to), I might start with a jump with no bar and click when the dog commits to going through the stantions and then toss a treat or toy as the dog lands. Once I add the bar back in, I would click when the dog was in the air to include the jump itself, as well as the landing and getting to the correct path afterward on the latent side and reinforce with a tossed treat or toy.

 

Some say the clicker is like a "scalpel" that is only for marking very specific behaviors. I don't always use it that way. I will use it to mark entire chains or conglomerates of behaviors.

[CMP]

That makes sense.

 

Once the animal is conditioned to accept the clicker as a confirmation of success / promise of reward it does not matter a lot HOW you employ it, so long as it means the same thing all the time.

 

It's no different from the "good dog" we give to itty puppies when they pee anywhere outside. Eventually the "good dog" only comes when they pee in the right spot and make their way back to the door. It's all shaping of one kind or another.

 

All training is shaping. Both the human and canine kind.

 

A clicker simply allows a greater precision and possibly control when shaping behaviours. It makes sense that it would work more quickly than voice - it is more consistent and can be applied more prcisely.

 

Reading every one of these posts (where my respect for people who train agility has increased a million-fold) only confirms for me that clicker training is not a new thing, it is simply an updated version of a very very old thing.

 

People who find the method objectionable (on the grounds that it is somehow less "personal") are the same sort of people who resisted email on the same basis

 

*hangs head in shame*

[juliepoudrier]

I still think this is probably just semantics, Kristine, but in my mind, from reading your description, you are still marking individual parts of the final sequence (one step back, two steps back). By the time you are clicking once for the whole sequence, you have already trained the individual parts by using the clicker. That is, you didn't go from standing to a full sequence of backing up in one fell swoop with the clicker marking the entire sequence (backing all the way up). That's what I believe Alligande meant when first questioning the assertion that the clicker could be used to mark a more difficult sequence in agility. You progress to the sequence by first training the parts. You don't start with the full blown sequence.

 

So when D112358 said that he? she? trained the sequence using a clicker, Alligande (and me) understood that to mean that the entire sequence was trained at once, which was not the case if I understood D112358's later explanation, which seems to describe actually marking parts of the sequence rather than the entire sequence.

 

It's all academic for me since I don't train for agility, nor do I clicker train, but I think this part of the discussion does illustrate beautifully how several people can read the same words and interpret them differently! (And I may try clicker training with my fearful foster, though I would likely choose to use a mouth click rather than a mechanical one.)

 

J.

[Donald McCaig]

Dear Doggers,

 

Sheepdogger dogma is that whistles are less emotional than voice commands. I'm not at all sure the sheepdog, who is getting signals from the sheep, the handler's whistles and - even at a great distance - the handler's stance would agree. It's all information though certainly most handlers use voice in the early stages and when things are going badly with experienced dogs.

 

I don't know enough about clicker training to know whether a mechanical click (unless produced by a mechanical robot) delivers more pristine real information than voice to the dog who is desperate for information and gets it by many means

 

Donald McCaig

[Root Beer]

I still think this is probably just semantics, Kristine, but in my mind, from reading your description, you are still marking individual parts of the final sequence (one step back, two steps back). By the time you are clicking once for the whole sequence, you have already trained the individual parts by using the clicker. That is, you didn't go from standing to a full sequence of backing up in one fell swoop with the clicker marking the entire sequence (backing all the way up). That's what I believe Alligande meant when first questioning the assertion that the clicker could be used to mark a more difficult sequence in agility. You progress to the sequence by first training the parts. You don't start with the full blown sequence.

 

I guess it just depends. I know that I don't really use the clicker exclusively in any one way. While I can't think, off the top of my head, of a skill where I would start with the whole instead of the parts, I would seriously hesitate to say I never do it.

 

I use the clicker - here and there, where it is beneficial - all the way from initial introduction of behaviors (not all, but most) to fluency and generalization. I definitely use it the most in behavior modification, followed by Freestyle and trick training. I probably do use it the least in Agility. But, again, there is a very limited pool of distinct behaviors that a dog needs to know for Agility. When it gets to the point where it is more about learning proper handling I'll use the clicker a lot less because if I am handling correctly, the dog is going to be on the right path. I will typically only use the clicker with a more advanced Agility skill if something is deteriorating or needs to be sharpened or refreshed.

 

But I know I use the clicker in a lot of ways that people typically don't. I think that part of the beauty of the clicker is that it is far more versatile than most people realize. I've broken almost every clicker "rule" (other than always feed after a click even if the click was a mistake) and found that doing so can sometimes be more effective than following the "rules"!!!

 

So, if someone says "I never use a clicker for . . ." I will probably be thinking, "how do (or could) I use the clicker for exactly that!?"

[Laurelin]

I wouldn't say all training is shaping. I think what you meant is that the click is the same thing as marking a behavior with a 'good dog'. That is very true- clicker training is just using a marker. People have been doing that for ages.

 

Shaping is a common method clicker trainers use (and I'm sure others do to). But you can use a clicker with other training methods (like luring) as well. Free shaping is what I see shaping referring to most often. I can see how free shaping can appear very impersonal since the handler is generally giving little to no input other than a click. I like free shaping a lot though, I find it builds confidence and enthusiasm in a dog.

[Root Beer]

 

I don't know enough about clicker training to know whether a mechanical click (unless produced by a mechanical robot) delivers more pristine real information than voice to the dog who is desperate for information and gets it by many means

 

Donald McCaig

 

I don't know that I would say it is more "pristine real information", but it does function in a way that I have not seen happen in response to any word in certain circumstances with certain dogs.

 

I get students in my classes at times who try to teach the LAT game with a marker word. Yes, it technically can be done. But I have actually never seen it work with the dogs who need the game the most. The click has the ability to "cut through" the overexcitement, the fear, the distraction, or whatever. A lot of times I doubt these dogs even hear the marker words that are being used. If the click is properly conditioned, they hear it.

 

Once the dog knows the game, the handler can typically switch to a marker word, and eventually just verbal praise. But without the clicker step, it just doesn't seem to work out.

 

The click is distinct. It does not sound like any word we say. The click is to-the-second precise. The click can evoke a very strong positive emotional response (when properly conditioned). I doubt I would use the term "pristine real information", but I definitely call it a very, very good thing.

[CMP]

I wouldn't say all training is shaping. I think what you meant is that the click is the same thing as marking a behavior with a 'good dog'. That is very true- clicker training is just using a marker. People have been doing that for ages.

 

Shaping is a common method clicker trainers use (and I'm sure others do to). But you can use a clicker with other training methods (like luring) as well. Free shaping is what I see shaping referring to most often. I can see how free shaping can appear very impersonal since the handler is generally giving little to no input other than a click. I like free shaping a lot though, I find it builds confidence and enthusiasm in a dog.

I am not up to date on the terminology - but I meant what I typed literally. All training is done by shaping - whether it is the kind children get from adults/the world or the kind dogs get. Understanding is arrived at by way of learning the small component steps.

 

Children learn to walk by crawling, then pulling themselves up, then skittering along by holding something, then walking on their own. They learn to walk by shaping the behaviour from its discrete components.

 

Now I probably used the word incorrectly in the context of this discussion but the concept remains valid I believe.

[GentleLake]

I might call what you describe as a progression.

 

Shaping does have a very specific definition in terms of training vocabulary, and this isn't quite it.

[d112358]

Julie you have nailed my thoughts...

To be honest I would not approach training agility as you do, before I attempt the more esoteric turns in agility I would already expect my dog to understand his job, which is to drive over jumps and read my body language, which is all foundation work that I do agree is compatible with clicker training. I have spent many hours watching different levels of competitors running agility while ring working and in my opinion most mistakes are made by the humans, bad timing, bad choices on the path they chose, most of the dogs understood their jobs just got the wrong signals which is why I feel the emphasis at a certain point in agility needs to switch from the dog to the human.

Julie the tongue click is exactly what I used to do when riding, I never managed to use it for clicker training but I do use it for recall/attention I can't whistle and I have found it a great way to communicate without having to call his name.

 

I would agree that the handler giving wrong or late information is cause for errors about 90% of the time. The other 10% is on the dog. I like to intermittently reinforce even when the dog's running at a pretty high level, and find that the precision of the clicker comes in handy every once in a while (particularly when the dog has tried the wrong thing a couple of times in a row). This wouldn't be a method I'd use in sequence with a dog that thinks running is self-rewarding, but it works well for my terriers.

[Maralynn]

In dog training lingo shaping is a specific method even though. Other methods of training include luring, capturing and complusion. Examples of the three in dog training scenarios - Compulsion, dog is on a leash and given a "pop" when they're not in the heel position. Luring, handler has food in their hand and holds it in a way that the dog comes into/walks in heel position. Capturing, dog walks over to heel position while you're just standing there and you reward/praise. Shaping, handler marks/rewards the dog for moving towards and eventually getting into the heel position.

[alligande]

 

I definitely use it the most in behavior modification, followed by Freestyle and trick training. I probably do use it the least in Agility. But, again, there is a very limited pool of distinct behaviors that a dog needs to know for Agility. When it gets to the point where it is more about learning proper handling I'll use the clicker a lot less because if I am handling correctly, the dog is going to be on the right path. I will typically only use the clicker with a more advanced Agility skill if something is deteriorating or needs to be sharpened or refreshed.

"

^^^^ this is what I meant, phrased differently. I have been sharpening up entries to poles recently and if I used an actual clicker I would be using it for this as that is my partners job with clearly defined requirements, plus it should not be effected by my handling.