Schedules of Reinforcement

[corgipower]

ClickerSolutions Training Articles -- Ratios, Schedules -- Why and When
^clicky^

Then, thoughts on continuous reinforcement vs variable?

 

[PWCorgi]

Didn't read the link, will tomorrow.

I use variable reinforcement schedule for training tricks and fun stuff like that.

With the reactivity work that I am doing right now it is definitely what I would consider continuous reinforcement. Heck, I don't care if I have to carry treats on me/be ready to run away (increase distance) from the stimuli FOREVER with Frodo, as long as I can get him to feel calm around people/dogs, I am a-ok.

I definitely think for most training, especially like recall training, once the dog knows what is expected of them then variable reinforcement scheduling develops a stronger response. Heck, this is why people become gambling addicts, maybe THIS time I will get a treat!

 

[Fran101]

Wow.. I really had no idea it was all that complicated. but interesting link!

sometimes he got a treat, sometimes he didn't (thats only because he got in the habit of ONLY doing it when I had a treat in my hand)

and he gets a treat everytime he looks cute lol so for no reason at all..

 

[motherofmany]

Well, what are the relative advantages of continuous reinforcement (CRF), ratio (FR or VR), or interval (FI or VI) schedules? Why and when would we use them? What are the advantages of CRF? When and why should we reinforce every response of a certain type, say, a proper SIT? First of all, the only way you can be sure that each response will be "proper," that is, that it will meet your criteria, is to reinforce EACH AND EVERY RESPONSE that is proper, correct according to your own criteria. If each correct response is NOT reinforced, and you start with a ratio, even a "two-fer," you are apt to allow less than perfect responses to acquire strength from that final reinforcer after the second response

This is the part that really caught my attention. I think we have been unintentionally reinforcing sloppy sits and downs with Abby by trying to use variable reinforcement as a "phase out" strategy

I'm going to discuss this with our trainer when my son goes back to classes after the new year. But in the meantime I think Thomas and I will go back to continuous reinforcement and try to tidy things up.

Abby is truly a teenager right now and I think we would all benefit from this approach. Maybe we can eliminate that space cadet look followed by a slooooow response LOL

 

[lizzybeth727]

I'll be honest upfront, I didn't read the entire article. But I do know a bit about schedules of reinforcement.

Ian Dunbar says that we shouldn't use variable reinforcement - reinforcing after a certain number of behaviors - but that we should use differential reinforcement - reinforce only the better behaviors. I think this is probably what most trainers do, and certainly (as I did see in the article) how you effectively shape behaviors.

Ken Ramierez says trainers usually don't use variable reinforcement, we use varying reinforcement: every behavior is reinforced, just some are reinforced with food and others are reinforced in other ways. This idea makes the most sense to me: after all, most of the time we ask our dog to do a behavior and we don't reinforce it with a treat, we do reinforce it some way - petting, toys, praise, or some kind of "life reward" (going outside, playing a game, etc.). Even if we don't do any of that, then most of the time we ask for a behavior we'll immediately ask for another behavior. And according to Premack, getting to do one behavior can reinforce another behavior. So even if you cue two behaviors in a row with no "reward" in between, the second cue reinforces the first behavior.

So yeah, I don't think of it as variable reinforcement (reinforcing some behaviors but not every behavior) I think of it as varying reinforcement (giving different reinforcers for behaviors rather than food every time). So technically, if you do varying reinforcement, you're reinforcing each behavior.... so you're on a continuous reinforcement schedule.

 

[corgipower]

Very interesting Lizzy. More to scratch my head about.

 

[PWCorgi]

Very interesting Lizzy. More to scratch my head about.

Ditto!

 

[JustaLilBitaLuck]

I didn't read the link either, but this takes me back to my Psychology class, when we were learning about Classical Conditioning and Operant Conditioning.

 

[elegy]

Ken Ramierez says trainers usually don't use variable reinforcement, we use varying reinforcement: every behavior is reinforced, just some are reinforced with food and others are reinforced in other ways. This idea makes the most sense to me: after all, most of the time we ask our dog to do a behavior and we don't reinforce it with a treat, we do reinforce it some way - petting, toys, praise, or some kind of "life reward" (going outside, playing a game, etc.).

This, totally, though I've never thought about it in so many words. I don't reward with food/toy every (correct) behavior I ask for, but I do acknowledge every correct behavior, and I know that that acknowledgment is reinforcing to my dogs. Not as rewarding as a hunk of cheese, but it's not nothing, either.

And like Ian Dunbar says, I tend to reward the "more correct" responses- the straighter fronts, the faster sits, etc.

 

[Doberluv]

It kind of depends on what I'm training and what dog. Some dogs don't seem to require as much fastidiousness. Often times, I have used a continuous at first until the behavior is quite reliable...regular, if imperfect. Then onto a fixed schedule (every 2 correct responses or every 3, say) for a short time, (a few sessions) which helps "stamp" it on their brains. Then onto a variable, without worrying about which responses are better than others at first. I just want to get the dog to do the behavior pretty well, but everytime he's cued. Then, finally, we go onto choosing the best examples and only reinforcing those. I want him to keep on trying. It is when he tries harder (because of a variable reinforcement schedule) that his odds of throwing a better example go up, thereby giving me more opportunity to reinforce those more perfect examples. But you can't wait too long before going to a variable reinforcement schedule or the dog can get "stuck" at the last level. In other words, I want to raise the ante fairly quickly but not sooner than the dog is showing some reasonable regularity. I wonder if that makes any sense. LOL.