UNDERWOOD (to Hackett) You look kinda familiar. HACKETT I was at the ceremony when you were awarded the 15 million dollars two years ago. You hit Super Lotto with 2, 7, 19, 23, 31, 41 and mega ball: 13. UNDERWOOD How do you remember that? HACKETT They were all prime numbers. That hasn't happened since the Mega Millions Miracle of ë98.

A table of prime factors, with the primes indicated in red.

CHARLIE I was right. ALAN Still, some of the money helps fund schools and support teachers. CHARLIE That's the irony. The money funds schools that teach, among other subjects, mathematics. But if more people understood probabilities, fewer would play the lottery. ALAN People give more if they think they'll get something in return. CHARLIE They're buying into a game that's designed to make players believe the odds are better than they are. The lottery makes you pick six numbers between 1 and 49. But it's the same as asking people to pick a number between 1 and 14 million.

Suppose you have a lottery ticket. The ticket shows your six good balls, and there are 50 bad balls. Six balls are picked from the 56 balls
in an urn. What are your chances of getting exactly 4, 5, or 6 matches? Many lotteries and gambling games are based on this concept of picking
from mixed good and bad balls.

CHARLIE This is a three dimensional representation of the lottery's numerical relationship between ticket serial numbers and prize amounts. You're looking at 50 data points. Each one represents a serial number of a stolen scratchoff. As you can see, they appear to be random. HACKETT With a limited amount of information, there's no way to determine the algorithm used to encode the tickets. CHARLIE But the robbers accumulated 10,000 serial numbers. Hackett hits a button, the model's filled with 10,000 points. They all line up to form a very distinct, repetitive pattern. DON Not so random after all. CHARLIE Once you know the pattern, you can crack the code. HACKETT The robbers have enough data to identify the serial number of the jackpot ticket.

The pseudorandom methods (such as the Mersenne Twister) seem random, while the quasirandom methods (such as Sobol) seem to have a pattern, with
less clustering. As an example of where quasirandom methods might be better, one method for estimating the area of a shape is to bound it, then
to pick random points from that area. Using pseudorandom numbers gives the Monte Carlo method. With quasirandom numbers, the method is called
quasi-Monte Carlo. Due to the relative evenness of the quasirandom methods, sometimes they give better estimates.