Tonight (March 30) is the drawing for the Mega Millions $640 million prize, the largest lottery jackpot in world history. Americans are lining up for tickets in hopes of winning big tonight, and with staggeringly remote 1-in-176 million odds, they're employing all sorts of number-picking strategies to improve their chances of hitting the jackpot.
But will any of these tricks work? Or does the randomness of a bouncing ball trump any scheme?
In the Mega Millions game, players pay $1 for a ticket printed with six two-digit numbers — numbers they can either choose themselves or have randomly generated by the ticket-seller's computer. Theories abound as to which is the better option, and, for those players who make their own picks, whether certain numbers are more likely to win than others.
However, when a group of researchers in California tested these theories against actual lotto statistics, they concluded that strategies are useless.
In 2010, Helio Yang, a professor of information and decision systems at San Diego State University, used computer simulations with his colleagues to test the effectiveness of the three most common number-picking methods: the "random strategy" — using randomly generated rather than chosen numbers, in consideration of the fact that lottery drawings are inherently random events; the "high frequency strategy" — choosing the numbers that have most frequently been selected in past winning drawings, and thus may reflect some inherent bias in the lottery machine setup toward that number; and the "low frequency strategy" — choosing numbers that are underrepresented in past drawings, and thus might be considered statistically overdue. [8 Weird Statistics about Daily Life]
Simulations based on winning number data from the California SuperLotto, which has the same rules as MegaMillions, showed that no strategy was more effective than the others in producing numbers that would go on to win in a next drawing. Furthermore, for the high frequency strategy, it didn't matter how much historical data was taken into account; picking numbers that have been hot for 10 years gives no better results than picking numbers that have been hot for only one year, proving that there is no timescale over which it makes sense to consider numbers "hot."
The success of the low frequency strategy, on the other hand, increased ever-so-slightly when more historical data was taken into account. That is, numbers that have been underrepresented in a huge number of past drawings are infinitesimally more likely to win in the next drawing than numbers selected using the other strategies — but the effect is so minuscule, the researchers wrote in their paper, that "one would have to play lottery forever, or for a very long period of time, in order to have some minor gains over other strategies which yield very low return anyway."
So, whether you pick 1, 2, 3, 4, 5 and 6, have your numbers assigned randomly, or carefully select numbers with noteworthy histories, yourchances of winning Mega Millions are pretty much the same, and very grim indeed.