Tuesday, March 03, 2009

Double Helix - Pandora, Learning Theory, and Variable Rewards

I've been away from my studio for a couple of weeks - and I'm going to try and catch up on some of my testing results and get them posted - before my notes become completely unreadable and I have to re-do everything. You think jpeg is a lossy compression? You should see my handwriting!

I was away at a dog-training seminar. This is pretty advanced stuff, and we use learning theory to teach dogs - on the principle that if you understand how an animal learns, you can teach it much more effectively. This spills over into the rest of my life, of course, and if I understand how people learn, I can teach that much more effectively too. And if I understand how I learn, I can learn much faster.

Anyhoo - one of the principles we discussed was "variable rewards." By varying the reward for a behaviour (once learned) - and sometimes even omitting the reward, you build desire in the dog for a "jackpot." Much like gambling in humans, where the payoff doesn't come every time, dogs will try harder to get that "big" reward. Where this becomes really important is both in trials and in the real world. If your dog expects a treat every single time he sits - when you take him into a competition ring, and he doesn't get one, he shuts down. Likewise, the police officer with his k-9 can't always praise or reward a behaviour - "good boy" might give away his position and get him shot. By varying the reward schedule in training - these dogs know that the next behavour might get an even bigger treat, and so they just try harder.

How the heck is this relevant to glass? Well - that's my theory as to why this category of "silver-saturated" glasses like the Double Helix, TAG, and Precision (R-4) glasses are so popular. Variable rewards! You just never know what you are going to get next!



For instance, this is what I got the first time I trying Pandora. How awesome is this? Blues and greens streaked with purples. Oh boy!


Next time, this is what I got. Mud and Algae. Ok, it's pretty cool mud, but still - ouch.

So I have been trying even since to get those first results again. Variable rewards.

This is what Double Helix says about Pandora:

"Pandora yields rich purples and greens when kiln struck. There is no need to flame strike it. Recommend kiln striking at 1000 F for 4 hours or anneal at your regular temperature for a longer length of time or for multiple cycles."


And I can add - whatever you do - don't hit it with a reduction flame. Ugh.

1 comment: