windhorsesue wrote:
So, I'm wondering, how would you handle it when you ask Blacky for an exercise, and he performs it, but at a lower standard than what you know he's capable of. Do you withhold the reward and ask him to perform it again better, so thereby only reward for attempts that meet your criteria, or do you still reward any attempt he makes, and use other methods to encourage him to excellence?
In some CT lit I'm hearing that in later stages of training, as criteria is raised or new criteria added, it's neccessary to make decisions about WHICH attempts to reinforce. I'm wondering if it's possible to raise standards without doing this?
Great question! The funny thing is that I don't really know the answer on how we do it either.
I only rarely decide to not reward stage 1 anymore because now we're at stage 2. Now I come to think of it, in all honesty I'm probably 99% of the time not differentiating between reward-no reward.
As I'm such a big softy who just loves to reward, I only differentiate by giving huge jackpots next to the regular rewards when Blacky or Sjors is doing something new or better. I still also reward all the old style attempts, but the improved version gets the big ones. The good thing about that is that I never have to consciously stop rewarding a version as the ponies themselves just extinguish that completely for the more profitable behavior.
This is actually quite interesting, because I've wondered quite a lot how on earth the ponies were teaching themselves all the improvements all the time, when I never stopped rewarding the old attempts....
I guess that when Blacky or Sjors does something on a really too low level (for examply bow really shallow), then I'll just ask him to make a transition from that not-good-enough bow to piaffe and then reward for that transition or new exercise. But there's no moment of silence where there consciously is no reward or something like that.
So I guess that instead of saying with a big 'no reward'-pause, 'do it over' cue or vocal correction, I just let the ponies compare their behavior themselves: 'this piaffe I tried better than that bow, that's why the piaffe was the one that earned the reward'.
By the way, about two weeks ago I actually started to differentiate actively in order to teach Blacky to stay longer in the levade: I give him the cue for levade with my bodylanguage +voice and hold that pose myself, also when Blacky already lands again in order to ask him to go up again in the levade, after which of course I reward big time.
Maybe it's just in my head, but here too it also feels more like I'm extending the old exercise, than that I consciously stop rewarding an older version. Because if Blacky would just land and not go up again even though I asked for it, I would still cuddle him for his attempts, or turn the levade into fun again by letting him chase me through the paddock out of the levade rear.
Maybe I'm not that good in making straightforward decisions on what not to reward, because my main goal isn't perfectioning the exercises as quick as possible (although there are advantages in that too). My main goal is to keep the ponies enthusiastic about every movement and have them really enjoy it all, and also to keep a certain freedom for them to experiment in older ('less good') movements as well. For example, when I taught Blacky to rear, he started with a very low forwards hop, and I was trying to teach him to rear up higher into a real pesade, standing in one spot. However, I also kept rewarding (not jackpots but regular treats) for the lower rears, and the ones that weren't in one place but in which Blacky took a step forwards or backwards. Out of all those so called faulty pesades, Blacky now has developed the levade (out of too low rears), the laufcourbette (from those steps forwards or backwards during the rear) and the terre a terre (jumping forwards in the rear).
I think that if I had carefully fenced off the real pesade by not rewarding for all the lesser attempts, I probably wouldn't have been able to teach him all those pesade-varieties this easy, because Blacky would have been taught that there was only one good rear in the first place. So I like to keep the boundaries open by just rewarding as much as I can, for good and bad attempts and just differentiate in the amount or quality of the reward. It also diminishes the risk of the ponies going over their physical limits in order to get that reward, because with two half attempts within their limits they earn just as much rewards,
and if they choose for that option, I know that I shouldn't go there for now.
However, I do think it might be good to start clickertraining/reward training with more fixed borders or reward-no reward untill your horse has learned the basic rules of training like this. And then you can just reward as much as you want.