No, this isn't a continuation of the other thread.... I was out on the land just now, delivering a cup of coffee to the old boy, like the good wife I am. W&S were being ignored while we discussed the delights of the sewage system, so they wandered off for a mooch. Willow, as always, keeping a beady eye on us whilst Shadow got caught up in the lovely smells of bunny. As I headed back to the house, I gave them a call and they whiplashed to me so well that I wanted to reward them. However, I didn't have my treat bag on me. Instead, I gave them that magic phrase, "do you want a treat?!" (rising in pitch until I'm ultrasonic by the last word), which sends them into outer space in terms of "ohmygoshohmygosh somethingGOODiscoming!" and they galloped ahead of me back to the house to collect their reward. I know that by the time they got their yumminess, they'd have forgotten what it was for, but it got me thinking a bit about longer-duration event markers for when you need some breathing space before you can deliver the reward. Rather than a click which marks that a reward is coming in a couple of seconds, it seems to me that giving them a word such as "treat!" is very rewarding in itself, and, because they're used to there being a pause between the word and the treat (normally I'd say it when I'm sat working, and would have to get up and walk over to the treat stash or the freezer to fetch it), it is able to maintain its effectiveness over a longer duration. I'm not saying you could use it when you're ten minutes from home, because you'd end up breaking the connection between the word and the reward, but it seemed to work well for what must have been a minute to ninety seconds - they were immediately excited by the word when they recalled, and they kept that level of anticipation until they received their treat.
Now that is interesting. I used to be a bit quick of the mark, so I would click, then instantaneously give a treat. Now I slow it down, and it's around 2 seconds. I think, as long as nothing is happening/distracting in between the event marker and the payout, it doesn't matter to some extent. I have a separate cupboard for chew treats and sausage, bedtime biscuits on the kitchen windowsill by the sink, dog food in the hallway, freezer for kongs. Benson has a magic word "sausage". From the gate on the driveway, to our front door its about 5-10 seconds at a gallop. I have used his magic word,(to stop when he is following marcus to work, or the nice postman is leaving.. ) and he has gone straight to his sausage cupboard The time between me yelling "sausage!" as he legs it down the driveway..to coming inside to the cupboard and waiting for me to open it probably around 20 - 30 seconds or so?
I think that's very important. If I had asked for another behaviour before giving the treat, I think I could have damaged the link between the word "treat" and the payout. Your dog needs to be confident that you're not going to break your promise. If I had said "treat" then asked them to walk to heel back to the house, it wouldn't have been fair. Cue - behaviour - marker - reward, not cue - behaviour - marker - cue - behaviour - reward.
I think that's only important when it's key you reward a specific behaviour which is generally during training or if they do something amazing. There are plenty of behaviours we build into chains or are established that we don't have to reward. I'm having another think about trained secondary reinforcers for those times I want to say "yes! well done" but I'm not by treats.
Oh yes, I'm not talking about chains. If you have a chain of behaviours, you don't mark partway through, do you? What I mean is once you have marked, you shouldn't ask for another behaviour before paying up.
My "treats!" word is almost a secondary reinforcer, in that they obviously get happy when they hear the word, but that's because of anticipation of the actual treat, so I don't think it can count as one really. Whereas games that have been taught using treats (tug etc) are now reinforcing in themselves and I don't need to keep "topping up" the reinforcer like I would have to with my "treats!" word. Sometimes I've used a quick game of chase as a reinforcer if I haven't had anything to hand, but play (of the type that they intrinsically find fun rather than a game they have to learn) is a primary reinforcer, I suppose.
I was fascinated by Ken Ramirez use of clapping as a secondary reinforcer. He spent time training it as such.
What did he pair it with? Food rewards or play? Mine do respond to a clap, but I'm sure it's more the body language and verbal praise that comes alongside - "Yay! You did so well!", and the fact it generally means there's a tennis ball coming. I'm sure that, at this stage, a clap wouldn't be reward enough that they would work for that alone.
He paired the clapping with food and would revisit that to maintain it's status as a good reinforcer. He used it if food had run out or if something special happened away from treats. In his videos it was often paired with food in general training as well as being trained in it's own right.
A cue itself becomes a secondary reinforcer - a cue is an opportunity to do a thing that earns a reward so it's reinforcing of whatever behaviour came before. That's how behaviour chains work. The cue to do the next behaviour reinforces the previous behaviour and so on.
I use clapping too but I didn't 'charge' the clapping exactly as you would a clicker. I give a cue, get the behaviour, then clap, then reward with food (which may be at a distance like in my treat bag hanging on the fence). The idea is that I can use clapping to say 'good job' when I can't use treats (like if I ever get round to doing an obedience trial when no food is allowed in the ring...).
Yes! We watched some video of a demo at Ken's seminar this year and you could see that exact thing happening
I use a short, sharp 'good' as my marker - I'm hopeless with clickers (lose them or don't have them handy) It works well for us and is a different 'good' than my usual way of talking.
I use "good" when I don't have my clicker and I want a short event marker. It's a completely different tone to normal speech. I've noticed that when I say it, the dogs immediately lick their lips! Good old Pavlov The thing with that (normal) kind of event marker, though, is that the dog expects the reward in short shrift; otherwise, you run the risk of muddying your marker. So, in this instance, I couldn't have used my "good" and then gone to fetch a treat, which would have taken me over a minute; that's just too long because of the way the marker and reward have been paired.
Also - with GD pups we have a system of random rewards - once they know a command - so they don't always get the treat anyway, sometimes just the marker. I have to write my random treat schedule down as I would soon develop a predicable pattern! It works really well. When GDs are working dogs they always get treats - but always randomly unless a new behaviour is being taught. ...
Emily Larlham uses a drawn out good to keep behaviour going which isn't paired with treats. She showed a video clip of one of her dogs doing spins and he noticeably speeded up when she said goooooooood and she got further than if she'd just rewarded at the end of the behaviour. Secondary reinforcers are fascinating I think
I also do that, Barbara! I use "goooooood" for, for example, walking to heel. It means, you're doing well, but it's not the end of the behaviour, so you may or may not get a treat immediately, but even when you do, it's not a release. Again, subtly different to the initial incident, because that wasn't an ongoing behaviour, just a delayed reward
