Lotto flaws there

[oddlots]

Kind of a fascinating story of mathematical sleuthing by a statistician:

http://www.wired.com/magazine/2011/01/ff_lottery/all/1

I love this moment of realisation:

"Delighted, he decided to take a lunchtime walk to the gas station to cash in his ticket. “On my way, I start looking at the tic-tac-toe game, and I begin to wonder how they make these things,” Srivastava says. “The tickets are clearly mass-produced, which means there must be some computer program that lays down the numbers. Of course, it would be really nice if the computer could just spit out random digits. But that’s not possible, since the lottery corporation needs to control the number of winning tickets. The game can’t be truly random. Instead, it has to generate the illusion of randomness while actually being carefully determined.”

Hat tip: Felix Salmon

[photon555]

Re: Lotto flaws there

Very interesting article. It reminds me of something that happened to me about 30 years ago. A friend of mine was a high school math teacher with a Master's degree. He started working part-time for a consulting firm that did statistical analysis to defend clients in court cases alleging discrimination, among other things. He mentioned to me that an ongoing case involved a person who had collected a lot of winning tickets in a game and was being investigated. He further explained that the producer of the tickets was using an IBM PC, then a newly released machine, to generate the random numbers. By coincidence I had recently examined the manual for the IBM PC, even though I had no hope of affording one. I told him that in the manual it explained that the PC generated psuedo-random numbers, and that the string of numbers generated depended on the initial "seed" number. If no "seed" was used, the PC used the same default one over and over. Apparently, the producer was restarting the PC regularly and not inputting a "seed". So the same string of numbers was being generated over and over. Obviously, the person had detected the pattern and taken advantage of it. But I wonder if that is actually illegal.

Actually, as was explained in that old manual, it is impossible for any computing device or algorithm to generate truly random numbers. Only natural processes are truly random. A true random number generator could be built using a radioactive sample and counting the number of decay incidents. It would still be necessary to be sure that the code that massaged the raw data into a useful form did not destroy or compromise the randomness.

After I explained this my friend got a somewhat sly look on his face and said goodbye. I never heard from him about the matter again, nor did I receive any fee from the consulting firm.

[nedtheguy]

Re: Lotto flaws there

redacted