(Brit) I have a little simulation here that saves us time. It does the random picking for us. And it allows me to start with four whites and two blacks. -And I'm going to have the computer do many selections for us.
-Sal: So every selection, the computer is doing the equivalent of, randomly picking out a marble. It's shaking it and randomly picking it. And you see that the [inaudible] percentage of white, it's getting pretty close to the actual percentage. Or, four to two ratio, I should say, of white to black. (Brit) So, I think I believe you now because at the beginning I see a lot of variation on this graph over time, and now it's kind of flat-lining. (Sal) It feels like it's converging to the actual proportion, and once again there' some chance that it didn't. You know, it could have been one of those one in a quadrillion sort of instances where it just kept picking white every time but, this is consistent with the intuition as the more trials you have, it just becomes harder and harder -- the probability of it being a significant departure from the actual proportion becomes lower.