Variable Rewards And Conditioned Reinforcers

[gila]

How do you train recall Vickie?

[The Spotted Devil]

Just out of curiosity...why do you (or anyone else) want a recall at a dead run?

I heavily reward fast recalls because they are a good reflection of Zig's level of distraction and self control IMO. I think his ability to recall from a very attractive distraction is improving all the time as a result. He is released back to the lovely smell etc 9 times out of 10 so he doesn't think he has to make a 'choice'. He gets the treat and the smell!

ETA: Just want to add that if Zig doesn't come when called I walk up to him, put his lead on and the fun is over. Free running is a privilage! After a few minutes of him calmly walking (he has also started to offer behaviours at this point) he is released, recalled and rewarded. I don't have to do this very often anymore

Edited September 9, 2009 by The Spotted Devil

[Vickie]

Just out of curiosity...why do you (or anyone else) want a recall at a dead run?

Well it does look impressive in competition

not in a sheep competition it wouldn't...

[corvus]

Just out of curiosity...why do you (or anyone else) want a recall at a dead run?

I don't, it's just what we normally get. It does make me very happy, though! I can't say why, but it is just very satisfying to have your dog gallop towards you when you call them. I think he just likes the recall a lot.

My old dog went deaf, but by then her eyesight wasn't sharp enough for signals, either. I found sharp sounds worked best, but eventually she couldn't be let off leash at all.

Vickie, what is your recall based on if not entirely rewards?

[Kavik]

Vickie:

- I guess it is all relative

Edited September 9, 2009 by Kavik

[Kelpie-i]

Noice dog on the vid!!

I am having serious recall issue with one of my old boys at the moment...silly old bugger has gone almost deaf (literally). He still hears a whistle depending which way he is facing, but we are now having to go through a process of teaching hand signals.

I feel your pain Vicki....my old boy is also losing his hearing and when we are out in the paddock and he wonders off, there is no way he can hear us. My OH can whistle through his mouth and this usually works but if I am out there on my own...I've got buckleys. Luckily he panics when he realises I am not there and comes looking...he knows where the gate is to come inside, so he'll make his way eventually.. Luckily the fences are pretty good around the paddocks.....I would not leave him off the lead in a public park these days! Silly old dogs...gotta love em!

[Kavik]

Interestingly Diesel has the worst and slowest recall of my lot. Kaos and Zoe are responsive and fast. Diesel is improving, but I don't get the same response as with the others. Sometimes a correction helps speed him up considerably. Don't need to do that with the others. Completely different temperaments.

ETA: Diesel is also the least body sensitive - which can be both good and bad.

Edited September 9, 2009 by Kavik

[Erny]

Just out of curiosity...why do you (or anyone else) want a recall at a dead run?

A dog that is zoned into returning to you as quickly as possible is not as likely to be distracted by other things in the environment along the way by comparison to a dog who ambles or trots back.

Apart from which, I'm impatient and when I call I want my dog with me asap

[Vickie]

Vickie, what is your recall based on if not entirely rewards?

I'm sure I start with rewards, but I use a variety of things:

I start by only using a recall word when they are coming, never when I want them to come until it is trained.

I use older trained dogs to teach the puppy...when I whistle, I know my dogs will come running, puppy follows & learns whistle.

I like my dogs to check in regularly when we are walking & encourage this from day one.

Maybe the most important thing for me is to enforce every recall & only recall when I am in a position to enforce it. If the dog doesn't come, I go & get it, take it's collar & pull it gently in the direction I am walking. I immediately release & then verbally reward when they continue.

They learn that when I call, they WILL come. I am not at all mean about it, but I don't want a recall to be a like a trick, I want compliance, not fun. I'm never going to have rewards on me everywhere I go, I'm just not organised enough.

It is impossible to use food/toys rewards to get a dog to recall off stock when they start training. With Trim I recalled her & then sent her around the sheep again as a variable reward, it worked, but I won't do it again. With Shine I used a rope...when I called she came, if the rope wasn't on, I went & got her. She has never been given the opportunity to blow me off. She now has a fantastic recall off sheep, whereas Trim is compliant but sometimes she needs to get them where she needs them to be first.

Edited September 9, 2009 by Vickie

[Vickie]

Just out of curiosity...why do you (or anyone else) want a recall at a dead run?

A dog that is zoned into returning to you as quickly as possible is not as likely to be distracted by other things in the environment along the way by comparison to a dog who ambles or trots back.

Apart from which, I'm impatient and when I call I want my dog with me asap

That makes sense..especially the last bit

[corvus]

They learn that when I call, they WILL come. I am not at all mean about it, but I don't want a recall to be a like a trick, I want compliance, not fun. I'm never going to have rewards on me everywhere I go, I'm just not organised enough.

This is true also for Nelson's method. She teaches you call once, if they don't come you do something that looks or sounds exciting, and if they still don't come you go and get them. I had Kivi on a long line for a few months while his recall was still pretty unreliable. We made a few mistakes, but recovered pretty well in the end. Just took longer.

Nelson encourages you to give them off leash time while the recall is still in training. I've never really understood this, and know one avid Nelson supporter who still doesn't do this. Her dogs stay on long lines until the recall is as good as she wants it to be.

Here's a video of Kivi's recall a few months ago:

He was still on the long line in that video due to him being totally in dog obsession mode that day. I think we should have used the long line a little more heavily in earlier stages. I will do so with Erik.

[Aidan]

I think my confusion mostly lies in not knowing whether a slow recall is worth reinforcing because I'm meant to be reinforcing every recall to get a conditioned recall, or if because my dog is capable of and usually performs a beautiful recall at a dead run, we should consider a slow recall not a recall?

You need to be very clear on what your criterion are.

In my classes we start off with pattern learning so that we can control all the variables and just build something to put a name on later. So by the time the behaviour is on cue we have already sorted out most of the speed, distance and even some distractions. When we put it on cue we work on latency (time taken to respond, which ideally is as close to zero as practical), improve or maintain speed, distance, and more distractions.

For speed I like to use set distances, e.g have the dog run between two people at either end of a tracking leash on a tether (about 20m in total between each person). Let's say we have a slow dog, he takes at least 6 seconds to make the journey 70% of the time (we will assume latency is not a problem, if it were we would need to reduce that first). We test this by performing 10 trials (all reinforced) and timing each one. From this we establish a time that we know the dog can beat about 70% of the time. This becomes our criteria, in this case, 6 seconds.

So we work with this time, and any trial that takes longer than 6 seconds is not reinforced. Then we test it, we perform another 10 trials, if 8 of them are within 6 seconds we can reduce the time to say, 5 seconds for our next round of trials. (For the record, 4 seconds is acceptable, 3 seconds is better).

(NB those skilled in drive training can incorporate their methods to improve these times very quickly, keeping in mind that this must be maintained when those antecedents are not present; e.g when you haven't been swinging a ball on a rope around)

This is obviously in a class setting, not in the real world. I think the idea of having a "casual recall" and an "emergency recall" is sound, so long as they have different cues. The ER is the one you use in class and you don't stuff up by using it inappropriately or casually. Later your ER becomes very strong and most of your criteria are sorted out, you're just working on distractions, at which point you should use it as often as possible in the real world.

[corvus]

Ah, thanks Aidan. That clears that up. Kivi's recalls are typically at a run. I think it's the odd delay that is where we can put work in. It does seem that as Erny says, when he is running he is pretty hard to distract. One time he broke away from a game with his best canine pal to recall, and she intercepted him. He dodged her, then shook her off when she got him anyway, made a substantial detour to run around her, and then ran even faster than usual once he had a clear path. We were pretty impressed.

It sounds like we need to keep on practising with different distractions.