Is the South African Lottery Rigged? A Hands-On Exercise for Bored Blog Readers

South African reader David Drew pointed me to this report, stating that one of South Africa’s political parties has called for an investigation of the “extremely suspect” March 3rd South African National Lottery after it yielded a record nine jackpot winners (as opposed to the typical 0-3 winners). The evidence given to support claims that the drawing was rigged is as follows: “According to the party’s ‘informed sources,’ there is (sic) never more than three winners in any lotto draw such as those in the USA and Britain.”

Relying on “informed sources” does not seem to be quite the right approach here. Relying on statistics is more sensible. To make things easy, the South African Lottery provides all the data you need right on their web page. (On the right-hand side of the page there’s a link you can follow that lets you download a CSV file with the data set.)

If you have some extra time and some knowledge of statistics, you can determine for yourself whether it’s likely that the March 3rd drawing was rigged. When you figure it out, be sure to let your fellow readers know. (I haven’t formally done the analysis, but just from glancing at the data I have a strong hunch about the right answer.)

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.

 

COMMENTS: 132

View All Comments »
  1. frankenduf says:

    sounds like a dumb scam- why not win first and split the loot later?

    Thumb up 0 Thumb down 0
  2. frankenduf says:

    sounds like a dumb scam- why not win first and split the loot later?

    Thumb up 0 Thumb down 0
  3. wallinbl says:

    If I’m reading the file correctly, 9 winners is only the fifth highest total for this lottery. March 15, 2003 shows 33 winners on lower ticket sales.

    Thumb up 0 Thumb down 0
  4. wallinbl says:

    If I’m reading the file correctly, 9 winners is only the fifth highest total for this lottery. March 15, 2003 shows 33 winners on lower ticket sales.

    Thumb up 0 Thumb down 0
  5. dantheman says:

    i think it’s rigged. the probability of having nine or more winners in the march 3 lotto is about 0.0000868. (binomial distribution with p = 1/(49nCr6) = 1/13,983,816; n = (total sales)/$2.50 = 30,193,421).

    march 3 is not the first time this has happened. the ides of march, 2003 seems to be the worst (33 jackpot winners), but also check out october 31, 2001 (19 winners), june 23, 2001 (17) september 30, 2000 (14) and march 11, 2000 (6).

    Thumb up 0 Thumb down 0
  6. dantheman says:

    i think it’s rigged. the probability of having nine or more winners in the march 3 lotto is about 0.0000868. (binomial distribution with p = 1/(49nCr6) = 1/13,983,816; n = (total sales)/$2.50 = 30,193,421).

    march 3 is not the first time this has happened. the ides of march, 2003 seems to be the worst (33 jackpot winners), but also check out october 31, 2001 (19 winners), june 23, 2001 (17) september 30, 2000 (14) and march 11, 2000 (6).

    Thumb up 0 Thumb down 0
  7. hcivic91 says:

    Its been a while on the stats but my calculations show P~0.00004. I think that violates even the strictest or rare event rules.

    Thumb up 0 Thumb down 0
  8. hcivic91 says:

    Its been a while on the stats but my calculations show P~0.00004. I think that violates even the strictest or rare event rules.

    Thumb up 0 Thumb down 0