Stalking/herding Behaviour In Kelpie

[sheena]

Don't lure him. If you find you are luring him then you are too close & have taken things to fast. Don't set him up to fail, which is what you are doing if the environment is too stimulating for him. It really helps if you have already taught him a few tricks using the clicker. You want to "shape" his behaviour, if you know what that means. You could start off with LAT in your own backyard. Put a teddy bear or just a piece of paper on a stick or anything. Be ready with your clicker & treats & as soon as he looks at the object, click & he will turn to you for the treat. It takes time & wont happen over night & even when you have him going pretty good it helps to go back to playing the game every now & then. It's important that he really understands the value of the clicker before you start & use really yummy homemade treats that you can break into very small pieces. If I want to jackpot my girl, I will take her favourite treat & slowly break it into pieces & feed it to her a tiny bit at a time...it's called silver service treating...you make the treat last about 30 seconds....they don't forget. My girls favourite treat is VIP Lamb Chunkers fried in garlic oil. They are easy to handle, soft & easily broken up into tiny pieces.

Twelve months ago, I would have had trouble keeping my girl on the agility field...she would be off chasing the dog on the next field. Now she is up into Masters. I still practice LAT with her a lot, but usually when she knows I have the clicker, she melts like putty in my hands :laugh: We have a dog in my agility class who was quite agressive & was banned from taking part in class. I introduced his owner to the clicker & LAT & a few months later, he is a different dog...but still a work in progress. She brings him to class & starts off with him at a fair distance & slowly brings him in where he gets to train with the other dogs. If he begins to lose focus, she takes him away again.

[Nekhbet]

And what happens if the owner is stuck without the clicker and treats. You're using a tool when you should be concentrating on creating a bond between the owner and the dog. Yes be at a distance, but the dog should want to look at the owner, and if you're constantly interacting when it does you dont need to have a pocket full of stuff with you. Treats are good for jackpots, to sweeten the deal but why should you constantly click, treat.

If the dog looks at the target, say nothing, if he makes motion to look at you talk to him, stroke his head gently, interact. The dog will be throwing itself at you for attention, and if the treat for some reason is not there all the time it doesnt matter.

When you click you ALWAYS treat, so the clicker ALWAYS has value to the dog..even if you click by mistake you should follow it with a treat. Using the word "yes" in place of the clicker, means for it to have the same value as the clicker, you must ALWAYS treat when you say "yes" otherwise the word will lose it's value.

See I don't have this problem. When the dog learns it's a marker they're happy to get what they get or wait to be released because I insist on more time spent on bonding between handler and dog instead of loading equipment. A marker work can easily be taught quickly, relationship building takes longer so we focus on that. You can have all the treats and clicker you want but if the dog doesnt learn to willingly look to you for guidance it's just another trick.

Edited December 26, 2012 by Nekhbet

[piper]

And what happens if the owner is stuck without the clicker and treats. You're using a tool when you should be concentrating on creating a bond between the owner and the dog. Yes be at a distance, but the dog should want to look at the owner, and if you're constantly interacting when it does you dont need to have a pocket full of stuff with you. Treats are good for jackpots, to sweeten the deal but why should you constantly click, treat.

I haven't used the look at that game but would if I had a dog that I felt it would benefit.

My understanding though, is at the moment they are discussing teaching the dog the desired behaviour. Why would you not reward highly when teaching an alternate behaviour that in the dog's mind is lower value than the activity it wants to engage in? After all you are trying to increase the value of looking at you and there has to be sufficient reason for the dog to choose this. Especially when discussing a dog that has already learnt the fun of engaging in sticky eyed behaviour.

At this point of the process it is no different to teaching any other behaviour - reward highly when the dog gives you what you want and then move to a random reward schedule and vary the types of rewards given.

[corvus]

And what happens if the owner is stuck without the clicker and treats. You're using a tool when you should be concentrating on creating a bond between the owner and the dog. Yes be at a distance, but the dog should want to look at the owner, and if you're constantly interacting when it does you dont need to have a pocket full of stuff with you. Treats are good for jackpots, to sweeten the deal but why should you constantly click, treat.

Honestly, I think games like LAT are profoundly good at building trusting relationships between dogs and their handlers. If you ever watch a pro at it, it's far from tricks and clicker and treat exchange. It's a lot of cooperating and two-way communication with your dog. The benefit of LAT over straight focus on you is LAT gives the dog permission to gather information they have a strong compulsion to gather. It's kind of like acknowledging that there are things in the world that may demand your dog's attention other than you. LAT gives the dog the opportunity to attend to those things without disengaging from you. As a result, the dog stays attentive to you and also feels calmer because they can see what's going on. You're facing it as a team rather than asking them to ignore something that they are hard-wired not to ignore.

If the dog looks at the target, say nothing, if he makes motion to look at you talk to him, stroke his head gently, interact. The dog will be throwing itself at you for attention, and if the treat for some reason is not there all the time it doesnt matter.

The reason why we don't do it this way if we follow Control Unleashed is because we're trying to stop the dog from starting an unwanted behavioural sequence. This is especially important with sticky herding behaviours and the likes because once they get on the way it can be a challenge to interrupt them. The more a dog practices them, the harder it is to change their habits. I would guess at least 95% of the time, focusing on the target is the precursor to the unwanted behavioural sequence. If you interrupt then, you have a high probability of preventing the sequence from occurring at all. THIS IS VERY GOOD, particularly with self-reinforcing behaviours. In LAT, the clicker interrupts the behavioural sequence and the reward comes when the dog looks back to the handler. So it should go: dog orients to target -> click -> dog orients to handler -> treat. If the dog doesn't orient to the handler after the click, you know they are probably not in an operant state and you need more distance. You're substituting an unwanted behavioural sequence with a more appropriate and desirable one and hooking it onto whatever triggers the unwanted sequence, so in the end the trigger itself cues the desired behavioural sequence (LAT). It's like any behaviour in that you can fade out the rewards later if you want, but people have a tendency to use LAT with reinforcement as a bit of a management crutch. Probably because it's so effective! Depending on the nature of the problem you're trying to treat, you can completely get rid of it with LAT and stop cueing it at all, or you will eventually plateau, which suggests you probably have some bigger emotional or arousal problems to deal with if you really want to put it to bed.

It's not like it's the only way to tackle the problem and I'm not dissing other ways, but it is tried and true with all sorts of problematic dogs all over the world. To really understand why, you have to delve a bit deeper into the CU program. The subtleties that come out in discussions amongst CU practitioners are eye opening. It's a hell of a lot more than a few tricks. Relationship is a massive part of it.

[Weasels]

Excellent post Corvus.

Robert Kaleski, who wrote the first kelpie breed standard, described a dog that "would rather work than eat" so I think it's understandable that people struggle with such a 'down-to-the-bone' instinct in a modern environment.

Counter-conditioning and later LAT/an environmentally-cued heel have been magic for my sticky dog. Also just redirecting his herding drive to chasing a ball (he will follow it until it stops, then stare at it until I go over and get it :laugh:) means I can walk among all manner of distractions, from kids yelling/running/playing soccer, people riding boogie boards on the beach, to dogs actually jumping on top of him to get his attention and he remains locked on that ball like... well like a kelpie doing a job!

Edit - re. luring, if it's a choice between having the dog react/start being predatory and luring them away, I will choose the luring every time. Not letting my dog practice the unwanted behaviour is a priority in our training. But as others said, use it as a piece of information on where his threshold is and try to work further out next time to set up for success :) And if we were caught without clicker and treats/toy in the early stage of training I would move away from the stimulus and leave the training session for another day. My boy doesn't have high value for pats and praise (we've worked up to "moderately reinforcing" from "moderately aversive") so I can't see that would've been a better offer than the reinforcement of engaging his instincts prior to the last year of work we've done.

Edited December 26, 2012 by TheLBD

[Nekhbet]

If you ever watch a pro at it, it's far from tricks and clicker and treat exchange.

The problem is we're talking to the average joe (no disrespect towards the OP) so taking one more thing to potentially go wrong out of the equation makes it easier. Who says the dog likes food anyway?

My point is that you can build relationship without gadgets that take more learning on top - and why are you using a clicker as an interrupter? I didn't pull my theories from the sky, and they work extremely successfully no matter how crazy a dog is. If I tried to add a clicker to my clients hands it's one more thing on top to go wrong or frustrate them. They have a voice the dog understands, and they have their hands. Treats and toys are the sweetener while they build that bond of understanding. I understand a kelpie wants to work, I think we miss the dog wants to work FOR YOU. You want to be part of the drive, not skirt around it.

[Aidan3]

The clicker isn't being used as an interrupter in LAT. what do you think would happen if the owner were 'caught without a clicker'?

[Are You Serious Jo]

My coolie is a real dickhead around other dogs so I'm thinking of trying LAT (hadn't heard of it before). He recalls fine when working but is really reactive with dogs. Biggest problem is finding a distractor at the right distance away.

[Aidan3]

No distractors or interruptors; reinforcers only. Mark for controlled engagement with stimulus, then shape the behaviour that you want.

The minute you start thinking 'distract' or 'interrupt', you're headed for trouble.

[megan_]

You can click or say yes - dogs are smart enough to know both of you condition them. LAT has literally changed Lucy and my life. I don't mark anymore - she understands the game well enough and offers the behaviour without prmolting. I still say "who's that". I wait for her to look rather than pointing things out to her - works for us.

[Aidan3]

You can do it without a clicker. If you're not in a hurry and don't care what you mark.

[Are You Serious Jo]

Sorry, meant to use stimuli rather than distractor. I'm such a dummy these days, I should come with an interpretation manual :laugh: I live in the country and can only find other dogs by driving to some random strange dogs and have to hope they stay far enough away. If I'm too close he won't ever look at me so no chance to catch that behaviour. I have a clicker so that's no problem. Just need to practise it because it is counter-intuitive and will take a bit of getting used to.

[corvus]

If you ever watch a pro at it, it's far from tricks and clicker and treat exchange.

The problem is we're talking to the average joe (no disrespect towards the OP) so taking one more thing to potentially go wrong out of the equation makes it easier. Who says the dog likes food anyway?

Uh huh, but I think one of the reasons why there are CU fans all over the place in ever-growing numbers is because things like LAT are pretty easy to learn and implement, results come quickly, and it works for a variety of problems. Even if someone just does LAT and nothing else from CU and even if they don't even really do it well or properly, I would be surprised if they managed to address their problem without also getting some positive knock-on effects on their relationship with their dog. And LAT does tend to work. Most people seem to have success with it.

My point is that you can build relationship without gadgets that take more learning on top - and why are you using a clicker as an interrupter? I didn't pull my theories from the sky, and they work extremely successfully no matter how crazy a dog is.

Yep, and my point was that LAT does address relationships and gives you a really valuable management tool. If you want it. If you don't, then don't use it. OP seems to want it, though.

You try clicking a clicker-savvy dog and tell me the click doesn't interrupt behaviour. :p It's a side-effect, but a potentially useful one. I certainly exploit it in training. In fact, the dog doesn't even have to be clicker savvy. It just helps. I am pretty confident Leslie McDevitt does not pull theories from the sky, either, so you're in good company.

If I tried to add a clicker to my clients hands it's one more thing on top to go wrong or frustrate them. They have a voice the dog understands, and they have their hands. Treats and toys are the sweetener while they build that bond of understanding. I understand a kelpie wants to work, I think we miss the dog wants to work FOR YOU. You want to be part of the drive, not skirt around it.

So use a verbal marker. Who cares? Clicker will work better, though.

I hate the notion that a dog wants to work for their human. It just brings in all this emotional mess and expectations that are as likely to make things worse as better. I hate seeing people upset because they just can't figure out how to make their dog care about them. Some dogs are more independent than others, even if they are kelpies. I don't think we need to put this expectation on owners that their dog should want their attention and want to do what they tell them to. Maybe they do, maybe they don't. It doesn't really matter. If you build up a nice reinforcement history with them they will get that anyway, it will be a moot point and no one will feel miserable because their dog didn't read the working dog manual or some such. Really, it's just behaviour. There's nothing wrong with rewarding the heck out of it if you like it. That's how we get more of it! You use rewards that increase the behaviour they follow, whatever those are. I have plenty of reasons why I think it's worthwhile getting a dog to work for food, but that's another story.

[corvus]

I should make it clear that Leslie doesn't often say the clicker acts as an interruptor, probably because people misunderstand. You're not clicking TO interrupt a behaviour. You're clicking to mark a quiet look before your dog ramps up. You're clicking to say "Good dog, you noticed that thing and you just looked. That's your job." But a side-effect of that is that the dog looks back to you and so they are looking away from the thing before they can get fixated. Doesn't mean they won't look back and get fixated, though. You have to make sure you're clicking looking, not staring. It's much easier to click the right thing if you are thinking about what you want your dog to do rather than what you don't want them to do.

[sheena]

And what happens if the owner is stuck without the clicker and treats. You're using a tool when you should be concentrating on creating a bond between the owner and the dog. Yes be at a distance, but the dog should want to look at the owner, and if you're constantly interacting when it does you dont need to have a pocket full of stuff with you. Treats are good for jackpots, to sweeten the deal but why should you constantly click, treat.

Honestly, I think games like LAT are profoundly good at building trusting relationships between dogs and their handlers. If you ever watch a pro at it, it's far from tricks and clicker and treat exchange. It's a lot of cooperating and two-way communication with your dog. The benefit of LAT over straight focus on you is LAT gives the dog permission to gather information they have a strong compulsion to gather. It's kind of like acknowledging that there are things in the world that may demand your dog's attention other than you. LAT gives the dog the opportunity to attend to those things without disengaging from you. As a result, the dog stays attentive to you and also feels calmer because they can see what's going on. You're facing it as a team rather than asking them to ignore something that they are hard-wired not to ignore.

If the dog looks at the target, say nothing, if he makes motion to look at you talk to him, stroke his head gently, interact. The dog will be throwing itself at you for attention, and if the treat for some reason is not there all the time it doesnt matter.

The reason why we don't do it this way if we follow Control Unleashed is because we're trying to stop the dog from starting an unwanted behavioural sequence. This is especially important with sticky herding behaviours and the likes because once they get on the way it can be a challenge to interrupt them. The more a dog practices them, the harder it is to change their habits. I would guess at least 95% of the time, focusing on the target is the precursor to the unwanted behavioural sequence. If you interrupt then, you have a high probability of preventing the sequence from occurring at all. THIS IS VERY GOOD, particularly with self-reinforcing behaviours. In LAT, the clicker interrupts the behavioural sequence and the reward comes when the dog looks back to the handler. So it should go: dog orients to target -> click -> dog orients to handler -> treat. If the dog doesn't orient to the handler after the click, you know they are probably not in an operant state and you need more distance. You're substituting an unwanted behavioural sequence with a more appropriate and desirable one and hooking it onto whatever triggers the unwanted sequence, so in the end the trigger itself cues the desired behavioural sequence (LAT). It's like any behaviour in that you can fade out the rewards later if you want, but people have a tendency to use LAT with reinforcement as a bit of a management crutch. Probably because it's so effective! Depending on the nature of the problem you're trying to treat, you can completely get rid of it with LAT and stop cueing it at all, or you will eventually plateau, which suggests you probably have some bigger emotional or arousal problems to deal with if you really want to put it to bed.

It's not like it's the only way to tackle the problem and I'm not dissing other ways, but it is tried and true with all sorts of problematic dogs all over the world. To really understand why, you have to delve a bit deeper into the CU program. The subtleties that come out in discussions amongst CU practitioners are eye opening. It's a hell of a lot more than a few tricks. Relationship is a massive part of it.

Corvus, you have explained this far better than I ever could. I know what I am talking about but have trouble trying to explain it to someone who knows nothing about the benefits of clicker training with your dog. I doubt whether Nekhbet fully understands the principals behind shaping your dog using a clicker. Some people (& I lost a few friends over this), cannot understand why you mark the looking away behaviour, when all they have ever known is the "Look at Me". When I first started LAT with my girl, I was jeered at & made fun of, from people who didn't even want to understand. Now whos laughing :laugh: ME as I can honestly say I have the best dog in the club & now students are asking me for help with their own dogs which I am only too happy to give.

[sheena]

Corvus...you are so "spot on"

[superminty]

Corvus - great posts. I sometimes struggle to get my head around some of your ideas :laugh: , but you've hit the nail on the head in this thread.

I doubt whether Nekhbet fully understands the principals behind shaping your dog using a clicker.

:laugh: I think you'll find Nekhbet understands the principles pretty well.

Edited December 26, 2012 by superminty

[Weasels]

I hate the notion that a dog wants to work for their human. It just brings in all this emotional mess and expectations that are as likely to make things worse as better. I hate seeing people upset because they just can't figure out how to make their dog care about them. Some dogs are more independent than others, even if they are kelpies. I don't think we need to put this expectation on owners that their dog should want their attention and want to do what they tell them to. Maybe they do, maybe they don't. It doesn't really matter. If you build up a nice reinforcement history with them they will get that anyway, it will be a moot point and no one will feel miserable because their dog didn't read the working dog manual or some such. Really, it's just behaviour. There's nothing wrong with rewarding the heck out of it if you like it. That's how we get more of it! You use rewards that increase the behaviour they follow, whatever those are. I have plenty of reasons why I think it's worthwhile getting a dog to work for food, but that's another story.

I have a dog (my non-sticky one) who will just work for the love of working and for the opportunity to engage with one of her owners, but I still use clicker training with her because the communication of what I want is just so much clearer, so she picks things up in a few minutes. The main difference between training her and my other 'stuff-you-human' dog something new is that I can fade the reinforcement quicker. It's nifty :)

Another pro-clicker point: in things like LAT when you are reacting to your environment quickly I find I get far more tongue-tied trying to cue-marker-next cue using my voice (I usually add in some praise, I probably shouldn't but it just happens), so adding a clicker to mark actually requires less coordination. However I don't have the best connection between my brain and my voicebox at the best of times so YMMV :laugh:

Edited December 26, 2012 by TheLBD

[Cosmolo]

Just to add to what Corvus has already said. One of the problems with traditional focus/ look at me type training is that it depends on the dog remaining looking at the handler. Most handlers (and me as a trainer actually) WANT the dog to be able to look around and to look at things like dogs and people WITHOUT going down a reactive/ predatory/ aggressive/ panic type response. And when you think about it- for many dogs focusing on the handler instead of the stimulus is a form of avoidance. I don't want the dogs to feel as though they need to avoid the stimulus- the more LAT you do, the more the dog has the ability to cope with the stimulus instead of avoid it. This type of reward history then really pays off when critical distance is badly breached- it raises the dogs threshold/ tolerance.

I do use the word interupt when i talk about LAT- simply because the process of LAT interupts the normal 'bad' behavioural sequence (simple eg- look- lunge- growl- snap)before it becomes inappropriate.

To answer the question what happens when you are without your clicker/ marker? The dog offers the behaviour themselves- look at stimulus, look at you and you reward them any way you can! But with very reactive dogs i strongly encourage owners to take rewards often if not always and actively look for opportunities to reward. In my experience this maintains the appropriate behaviour and hand in hand emotional response much better than traditional focus and/ or compulsion alone.

[Aidan3]

I'm my reactive dog classes I don't have the dogs look back to the handler. They do that anyway, when you click, but we never click for looking at the handler.

It's really very simple: Antecedent Behaviour Consequence

The Antecedent stays the same, but we change the Behaviour and (and I think people fail to appreciate this aspect) the Consequence.

The behaviour changes from stalking/reacting/whatever to something more useful, more calm, more controlled.

The consequence changes from chasing/distance/control to food or toy - something we control (later it might be herding, retrieving or whatever the dog will ultimately be doing).