The Room Do I dare disturb the universe?

Rambling On Cognitive Theory And Poker Bots

Rambling on Cognitive Theory and Poker Bots

Okay so I was reading The Way We Think which is a book on cognative theory … specifically blending theory. They posit that a Blend is (if I’m reading this right) basically pattern recognition (either physical or historical or both) combined with creative scenario / prediction (creating a mental space where your pattern recognition of the situation and knowledge of possible outcomes can combine) … this isn’t very clear but the idea started me thinking about the Poker website again.

The Bots that are good at poker perform a blend between their knowledge of the odds of specific combinations of cards appearing, the probabilities each of cards has at winning, and past history of performance in the hand (and previous hands) by the opponent. (Actually this is the way people are good at poker, I only assume bots are good the same way). I think a blend only becomes better when there is more knowledge informing the parts that make it up.IE if a Bot has more knoweldge of how a given opponent plays, they’ll be better at playing against that opponent, and if a bot can generalize over many opponents they’ll be better in general.

This combines with the article Geoff (Young) posted to use.perl about two types of Geniuses in the world. One type have a sudden shattering insight, and re-define their world. I suspect this kind of person has the ability to make a blend (or blends) from two radical points that nobody else can create. These types of people have one big flash, then fade away. Picasso with cubism is the specific example I remember the article talking about. The other type of genius slowly builds to become a master of their form. Pollock for example, or Rembrandt. This kind of person has a more traditional blend, but develops a deeper and deeper understanding of their medium. Rather than developing a radical new blend, they find a familiar blend that appeals to them that they can explore over and over finding new thought.

It would be interesting to figure out how to simulate the radical blend type of insight, I suspect that I fall into the second category of genius (ha!) and will eventually figure out how to simulate it on my own. (Should I be so lucky and gifted).