41.   For the fifth, which turned out to be the jackpot winner, Edwards said he decided to try something new and conjured up random numbers in his head.

42.   The risk of that type of error was once every nine billion random number pairs, less than IBM claimed, according to Intel.