Motivational Rewards

[Erny]

Yes, perhaps it depends on the dog.

My boy used to be hell bent on chasing birds. I've noticed over time (he's never caught any) that his interest in the chase is waning and his chase is no where near as strong for it as it was.

[Kelpie-i]

I was once told, by a very good source, that the training of Greyhounds does involves "wins", despite what people want to believe.

Interesting about the "never win" idea, will have to look into it a little further. I believe it would depend on the dog you are working with.

[Aidan]

Interesting stuff!

Working from first principles, chasing is an operant. Therefore it must be maintained by reinforcement.

Let's say a dog does reliably chase wallabies, but has never caught one. What is the reinforcer?

It is possible that other learning events are reinforcing this behaviour - e.g chasing a ball is a similar activity in which the dog does win. However I would posit that chasing wallabies must be innately reinforcing alone, otherwise the dog would quickly learn that he does catch tennis balls but never catches wallabies (two different discriminating stimuli) and the behaviour of chasing wallabies would extinguish eventually (it hasn't in 9 years with my boy, but maybe will with Erny's).

Anecdotal reports from owners of dogs who have hunted for several years unsuccessfully, then do eventually catch something, suggest that chasing often becomes stronger when the dog has had success. So I would consider the possibility that a schedule of "wins" would create stronger chasing behaviour (hence it's apparent use in training greyhounds).

It would follow that, in theory, a variable schedule of reinforcement through wins should produce the strongest chasing behaviour.

At the end of the day, from a practical viewpoint, if you can increase or maintain other behaviour by giving the opportunity to chase then it is a reinforcing consequence. If letting the dog win every so often makes it even more valuable, why wouldn't you do it?

[Erny]

Anecdotal reports from owners of dogs who have hunted for several years unsuccessfully, then do eventually catch something, suggest that chasing often becomes stronger when the dog has had success. So I would consider the possibility that a schedule of "wins" would create stronger chasing behaviour (hence it's apparent use in training greyhounds).

It would follow that, in theory, a variable schedule of reinforcement through wins should produce the strongest chasing behaviour.

At the end of the day, from a practical viewpoint, if you can increase or maintain other behaviour by giving the opportunity to chase then it is a reinforcing consequence. If letting the dog win every so often makes it even more valuable, why wouldn't you do it?

My highlights.

I agree about the letting the dog win ..... that was my point in the first place. But if the dog wins even more often than just "every so often" then this has to increase the dog's drive even more, no? Just to be clear, for training purposes, I'm not talking about making every 'win' easy (and "easy" is subjective .... it depends on the dog you're working with), as with the work I and many others do with their dogs, we want to inspire them to work for the 'win'.

Are you talking about variable reinforcement, or intermittent schedule of reinforcement? - I am presuming the latter. (It's probably just my termonology - but I tend to refer to variable when I'm talking about low value reinforcement compared to high value reinforcement, as opposed to how often the dog acquires reinforcement.) Intermittent schedule of reinforcement does make the behaviour less prone to extinction. But "intermittent" in the sense of 'rarely' isn't necessarily the ideal for training purposes and I thought that Corvus put up the video clip to show the (her) use of the "Flirt Pole" as a training aid. Which is what got me puzzled.

Edited December 2, 2009 by Erny

[raineth]

well one of my dog's favourite game was to chase other dogs (or me). Now I guess that makes him sound out of control - but it wasn't. The chasing of other dogs was often invited by the other dog (play bow and then run away), and chasing me was by invitation only.

But all he did was chase - when he got to the other dog he wouldn't do anything with it other than smooch it; that couldn't be called a win, could it? Yet he never stopped loving that game.

I suppose chasing me was rewarded by attention and pats when he caught up so I can see that that would be a reward.

[Erny]

I think the answer to this is working out what the dog's goal is.

Edited December 2, 2009 by Erny

[Aidan]

But if the dog wins even more often than just "every so often" then this has to increase the dog's drive even more, no?

Yes, well sort of. The more times you reinforce something, the more probable that something becomes (up to a point). The question is - what is that "something" that you are reinforcing?

If you reinforce a range of responses (from low to high intensity) you will get a similar range of responses. If you reinforce the strongest responses only, and ignore all weaker responses (put them on an extinction procedure, i.e withold the reinforcer) you will get stronger responses (this is known as Continuous Reinforcement [Differential]).

Here's the thing - once you start extinguishing the weaker responses, you hit an extinction curve which will produce stronger responses. If you reinforce those responses, you will get more of them.

Once you have a very high rate of strong responses, you can switch to an intermittent schedule. In this case (being a duration behaviour) you would use a Variable Interval Schedule of Reinforcement which would give you a very high rate of responding that is resistant to extinction.

ETA: I will not take any responsibility for any brain explosions that may result from reading my post. Sorry if I sucked all the fun out of it! Bottom-line - let your dog win when he's putting the most effort in!

Edited December 2, 2009 by Aidan

[Erny]

If you reinforce a range of responses (from low to high intensity) you will get a similar range of responses. If you reinforce the strongest responses only, and ignore all weaker responses (put them on an extinction procedure, i.e withold the reinforcer) you will get stronger responses (this is known as Continuous Reinforcement [Differential]).

I'm thinking there's two different things you are describing there, Aidan.

One is "Successive Approximation". In other words, rewarding for the responses that are the closest to the desirable behaviour that you want, ignoring the behaviours that are furthest away from that goal.

Continuous Schedule of Reinforcement is reinforcement that occurs for every time of the desired behaviour.

You can use Continuous Schedule of Reinforcement with Successive Approximation.

You can use Continuous Schedule of Reinforcement without using Successive Approximation.

Yep - it's all this training termonology again .

Here's the thing - once you start extinguishing the weaker responses, you hit an extinction curve which will produce stronger responses. If you reinforce those responses, you will get more of them.

Once you have a very high rate of strong responses, you can switch to an intermittent schedule. In this case (being a duration behaviour) you would use a Variable Interval Schedule of Reinforcement which would give you a very high rate of responding that is resistant to extinction.

I know what you're saying, but maybe because we are talking different words for different meanings, I'm not understanding your point. IE I'm not sure how all of this goes back to the "Flirt Pole" and its use as shown in Corvus' clip, and the point she made about not giving wins, and me querying that.

[Kavik]

Just to add in the mix:

We know that allowing them to chase light is bad - laser pointers, shadows, reflections - as they can get obsessed with them. And they can never catch them so can't get drive satisfaction. So how does that fit in with this chasing/catching argument? Since it seems they get more worked up because they never catch them?

I think I'll put chasing birds in this category too - Zoe loves to chase birds. When she was younger I'd take her to a park where there were swallows and it was her favourite pasttime - and she never caught any.

[Erny]

Just to add in the mix:

We know that allowing them to chase light is bad - laser pointers, shadows, reflections - as they can get obsessed with them. And they can never catch them so can't get drive satisfaction. So how does that fit in with this chasing/catching argument? Since it seems they get more worked up because they never catch them?

.... Go away, Kavik !!!! LOL

On a more serious note though, and perhaps relevant - using the laser pointers etc. can lead to obsessive compulsive "disorders". And perhaps the same might be said of chasing balls, or birds or whatever, from whence they do not get a drive satisfaction. "Disorder" might be somewhere different to the discussion of what we would consider "normal" responses (be that a cessation/reduction of behaviour -vs- increase of behaviour, due to no-wins -vs- wins).

Edited December 2, 2009 by Erny

[Aidan]

You can use Continuous Schedule of Reinforcement with Successive Approximation.

That's the one I'm referring to, I just used a different name - Continuous Reinforcement [Differential] They are synonymous. I'm not sure how widely used that term is, but it's what I learned first

I know what you're saying, but maybe because we are talking different words for different meanings, I'm not understanding your point. IE I'm not sure how all of this goes back to the "Flirt Pole" and its use as shown in Corvus' clip, and the point she made about not giving wins, and me querying that.

We veered away from that, I suggested that some wins could make chasing a stronger behaviour, you suggested that more wins rather than fewer would make it stronger again, I agreed but added that making the wins coincide with the strongest responses would be the way to do it. In a nutshell.

My personal opinion on this is that whatever Corvus does, whether her dog ever wins or not, probably won't matter much for the purposes of the exercise that she is doing (if I am to understand it correctly), which is teaching "on" and "off", switching the dog in and out of drive into a default down, then reinforcing that with more play in an almost continuous loop.

[Erny]

We veered away from that, I suggested that some wins could make chasing a stronger behaviour, you suggested that more wins rather than fewer would make it stronger again, I agreed but added that making the wins coincide with the strongest responses would be the way to do it.

Yep.

My personal opinion on this is that whatever Corvus does, whether her dog ever wins or not, probably won't matter much for the purposes of the exercise that she is doing (if I am to understand it correctly), which is teaching "on" and "off", switching the dog in and out of drive into a default down, then reinforcing that with more play in an almost continuous loop.

Sorry - but I missed that as an explanation of what Corvus was doing.

Thing is, I didn't see her dog switching out of drive. A dog can still be in drive and drop simultaneously, so in the exercise as demonstrated by Corvus, I don't see drop as an "off" switch to anything, other than merely a position the dog has learnt that will get him what he wants (ie for the ball to come alive again) .... and with the way it is being taught I can see her achieving the opposite of what she is seeking.

If what Corvus seeks is merely an automatic drop as a classically conditioned response, then yes, I see that the exercise she is doing doesn't much matter. But if what Corvus seeks is a switch out of drive, then I'm having trouble seeing how that exercise, with that timing, is going to achieve that.

Edited December 2, 2009 by Erny

[Aidan]

We veered away from that, I suggested that some wins could make chasing a stronger behaviour, you suggested that more wins rather than fewer would make it stronger again, I agreed but added that making the wins coincide with the strongest responses would be the way to do it.

Yep.

Just one other thing, I had been saying earlier that a variable schedule would produce stronger responses, which generally it will regardless of any differential reinforcement, but I think someone will shoot me if I go down that path I just wanted to point it out because it looked like I was contradicting myself.

Thing is, I didn't see her dog switching out of drive. A dog can still be in drive and drop simultaneously, so in the exercise as demonstrated by Corvus, I don't see drop as an "off" switch to anything

Honestly, I'm not too clear on the exercise, can only go off what little Corvus has written about it. She has also said that this clip was just demonstrating play with the flirt pole. I would suggest to Corvus that once you put a behaviour into an animal's repertoire, be careful using it two different ways if the antecedents haven't changed. I'd be putting that barking on cue, too.

[corvus]

For Erny as she seems to have missed it:

In a more specific sense, Erik is practicing default downs at the moment because I love them and I'm trying to teach him to think through excitement. This is kind of Leslie McDevitt's off-switch game, but today Erik was a bit too "on" and if I wanted to practice that it would have been smart to have left the flirt pole behind and used something less intensely exciting. The off-switch game is played by doing something exciting until just before your dog loses it, then disengaging, waiting for a default behaviour, and then starting the game again. This was going great for us until Erik got the hang of it and now he just downs without thinking about it most of the time, which kind of defeats the purpose of the game as I understand it. So I've started introducing other commands to ask him instead of just waiting for his down. He was doing all right this afternoon, but it was so not the exercise as it's meant to be done.

..with emphasis added.

Man, I've been meaning to put that barking on cue for months and I keep forgetting. He does shut up on cue all right, though. Fascinating thing, since he's learnt the default down, when he gets the "Erik want x y and z!" barkies, if I tell him "shh" he shuts up AND goes into a default down. He came up with that one all on his own. It's interesting to see when default behaviours come out. With Erik, when he wants something and when he doesn't know what to do. It's a nice neon sign for me if I'm training with him and instead of doing what I ask he downs. I'm like, "What's that Lassie, you are confused??" He also does it when he's too revved to do anything else. Another nice neon sign for me. I need neon signs sometimes.

Incidentally, my other dog is generally quite disinterested in catching. The only time he will do it is when he's as drivey as he gets without losing it. Any other time he loves to chase, but the moment it stops moving he walks away. Sometimes he doesn't get to a ball before it rolls to a stop and he just wanders off. If I throw it again, he'll bound after it and if he gets to it before it stops he'll pounce, but then he walks away again. He doesn't even pick it up most of the time. I have two herders and they both very much like to chase and stop whatever is moving. My old corgi also liked to chase roos. Being a corgi, she never caught one. She chased for about ten metres and then they were well and truly gone. In contrast, Pyry the Vallhund does not like to chase. He only likes sure bets and as far as he's concerned, it's not worth his energy unless he knows he can catch it and kill it.

Can I just stress again that the flirt pole is just a toy. The clip was basically just me playing with Erik and milking off some of his excess energy. Just fun and exercise. I'm a little bewildered that suddenly it is a "tool" and needs to have a "purpose" and be used in a "proper" way. It's only what you make of it. It is, however, a very exciting toy for doggies.

[Erny]

Hope it goes well for you and serves your purpose (your training in general, Corvus .... not just the flirt pole game you play).

[Aidan]

Can I just stress again that the flirt pole is just a toy. The clip was basically just me playing with Erik and milking off some of his excess energy. Just fun and exercise. I'm a little bewildered that suddenly it is a "tool" and needs to have a "purpose" and be used in a "proper" way. It's only what you make of it. It is, however, a very exciting toy for doggies.

Sure, no problem. It was only a suggestion if you're using the same toy/tool for two different outcomes, you can either use a cue or a different toy on the end. </bewilderment>

It would actually be good to see the exercise that you described earlier if you have a clip of that?

[woodbyne]

Hi All,

Ok my sheepie tug arrived today so just had a play and she LOVES it not that keen on the one with the pouch to put food in might need some tastier food but the sheepie shewas going nuts for it. We have training on friday so i'm going to take her on the other side of the oval play over ther and get there before anyone else does to see what reaction i get.

[corvus]

Can I just stress again that the flirt pole is just a toy. The clip was basically just me playing with Erik and milking off some of his excess energy. Just fun and exercise. I'm a little bewildered that suddenly it is a "tool" and needs to have a "purpose" and be used in a "proper" way. It's only what you make of it. It is, however, a very exciting toy for doggies.

Sure, no problem. It was only a suggestion if you're using the same toy/tool for two different outcomes, you can either use a cue or a different toy on the end. </bewilderment>

That wasn't directed specifically at you, Aidan. Everything's cool. Erik is very good at cues, whether I'm conscious about them or not. Gotta watch myself with him! He seems to learn cues for getting very very excited everywhere he goes.

It would actually be good to see the exercise that you described earlier if you have a clip of that?

Yeah... I realise that, if only because I'm getting tired of trying to explain it. I only have early ones where we were both fumbling around a bit, but Erik did a beautiful off-switch for me for his dinner this evening, blinking, tail down, trying to pretend to be sleepy. He was either doing a good job of pretending or he can ramp up in a millisecond. Either way, it's the thought that counts at this point. He can't act relaxed without being relaxed forever! He's not that clever. I think.

I'll see if I can get something that's more like what I think an off-switch game should look like, but I'm on a steep learning curve, here, and I make a lot of mistakes. :D

Good luck, woodbyne!

[Vickie]

What's the purpose? It's fun? What's the purpose of any drivey game?

Sorry, I didn't mean to sound harsh. Out of curiosity if he had fun chasing dragonflies around the garden, would you be happy about it based on the fact that he found it fun?

I understand people have different goals but personally it would bother me if any of mine were using all that drive in a game that didn't really involve interacting with me. For the most part, (although they do play with each other) most of the things that they get enjoyment out of are with my interaction.

[corvus]

I kinda think that flirt poles don't work without a person on the other end, but maybe that's just me. I used to take it to the dog park, but I stopped doing it because he wouldn't leave my side or stop staring at me and throwing behaviours at me as long as I had it in my hand. So it stays at home now, at least until we master the "Go play" suggestion to go be a dog for a while without stressing about whether I'll let him play with the flirt pole or not. Any toy that doesn't work without a human partner involves interaction. Even if all he does is a default down, he doesn't get to play for free.