Statistics PhD wins lottery four times

Joan R. Ginther has won the lottery four times, each time with million dollar payouts. Either she’s the luckiest woman in the world, or her wins have to do with her statistics PhD. I suspect the latter. (And I need to get my hands on a lottery dataset.)

[Thanks, Ryan]


  • This seems to crop up once in a while (last one I remember, though very different circumstances). Inevitably, it’s because winning tickets are somehow treated non-randomly. In this case, I would be surprised if there’s some kind of algorithm for how the winning tickets were spread around to different locations over time, and Ms. Ginther deciphered it (likely from publicly available information).

    I suspect she could have made even more money selling her system to the Mafia or such, though.

    • Rereading my linked article, Joan Ginther was also mentioned there (written at the beginning of the year). The new information on her background certainly adds credence to the suspicion voiced by the article.

  • Fascinating article in the August Harpers magazine.

  • Peter Krenesky August 10, 2011 at 1:38 pm

    The scratch tickets, like the one pictured, have been widely reported to have security flaws allowing you to predict winning tickets with a high probability. Coincidently (yeah right), the flaws were found by another statistician.

  • A fellow who has won many times gave a tip on how to win often. Pick a frame of reference like 3 or 6 months and log how many times each number in the sequence of numbers played fell into the win. One can tell which numbers come up most and least often leading one to think that the most often hit numbers are the best to play. I found that in a succession of 6 numbers, 2 will be from the most hit group, 2 will be from the least hit group and 2 will be from the mid hit group. There will be many possible number combinations to go through but a manageable amount. Playing on 5 different combinations I hit 4 of the 6 every time. One set of numbers usually comes up with 5. My problem is picking the bonus numbers. There’s not much data to go on.

  • Yma Sauvignon September 15, 2011 at 1:04 am

    Lotto – A million people lose small to medium amounts so very few people can lose their way.

    The smoke and mirrors used by Camelot (UK) divert most people’s attention from this sad fact.

    Further – I have done a statistical analysis of the number of jackpot prizes versus the reported number of entries, and I have decided that these “independently scrutinised” lotteries are rigged by the left hand while everyone watches the right.

    You are better off going to a casino and playing a few minimum house bets on the 35:1 roulette numbers. If you can win once, compound your bet as long as you can within the house limit – your odds are better and the system is fairer.

    Press Association – Mon, Sep 12, 2011
    http: //