Wednesday, June 17, 2009

Fun with statistics: Wellington edition

I like to think of myself as an equal-opportunity news-rager, and as such I would like to scupper my chances of working for any mainstream news organisation in the country by critiquing this awful article from the Dominion Post (which, like the Stuff website and everything else in New Zealand not owned by APN, is a subsidiary of Fairfax Media). I know it's a lifestyle article, but I hold that nothing is beyond my purview.

An outrageous fortune of more than $25 million is set to be won in Lotto's Big Wednesday draw tonight as ticket sales near record levels.

After an unprecedented 20 weeks of jackpot drought and millions of unlucky tickets the odds are now firmly pointing to the windfall being won this week.

The luckiest Big Wednesday number is 44, followed by 10, 16, 12 and 22. The number 7 comes up the least. But the chances of a single ticket scooping the whole prize are so slim they confounded a maths academic's computer.

No wonder people don't understand statistics. Why is it more likely to be won this week? Is it because of the weeks of jackpot drought, and the millions of unlucky tickets? No - as far as I can tell, it's just because more people than ever have been convinced to flush $5 (or whatever lotto tickets cost) down the toilet, making each combination of numbers more likely to be picked. As for the numbers, how can number 7 be last when everyone knows that 7 is lucky in gambling? It's enough to make you go to a lotto shop and pick 1-2-3-4-5-6 out of statistical stubbornness.
Sure enough, she [VUW maths professor Stefanka Chukova] reported that the chances of victory were microscopic about [sic] one in 16 million for every line of numbers and one in 2.7 million for a lucky dip ticket.

But even if a person won this week, the chances of keeping the whole haul to themselves were tiny, she said. "Even the software on my computer can't handle it."

Is the software on the "maths academic's computer" just the Windows calculator? Because I wouldn't be surprised. But it gets worse.
Despite knowing the odds Dr Chukova still plans to buy a ticket "$25 million is a lot of money".
Sigh - I guess this just goes to show that common sense and fancy degrees have little or no correlation. Here are EtH's official recommendations for how to spend your $5 other than on Lotto:
  • go to your broker and buy $5 worth of diversified shares.
  • buy $5 worth of vegetable seeds at your local garden centre, plant them and sell the produce at a roadside stall.
  • pay an unemployed victim of the economic crisis to do your housework, freeing time for you to write your blockbuster screenplay.
  • buy a used shovel on Trademe and use it to dig up your back garden; should you strike oil, you will become wealthy beyond your wildest dreams.
  • write a love letter on the $5 note to a wealthy eligible bachelor/socialite in the hopes that they will marry you and share their fortune. (Although, to be honest, you could use the back of a lotto ticket as well.)
To further fill out the concept of statistics for confused readers, journalist Tom Fitzsimons has produced a handy table of statistical probabilities to compare with your lotto chances:


-Dying in a car crash on New Zealand roads: 1 in 11,000

-Having caught swine flu by yesterday: 1 in 40,000

-Being struck by lightning: 1 in 280,000

-Being elected Prime Minister: 1 in 4,312,532

-Winning a share of Big Wednesday: 1 in 2,715,020

Of course, your chances of dying in a car crash, like your chances of becoming Prime Minister, are significantly lower if you just sit around home all day in your pyjamas. That's why I imagine bookies - people who know a thing or two about probability - would not have have been offering odds of 1 in 4,312,532 on John Key last year. The less said about the odds of "having caught swine flu by yesterday" the better.


  1. You'll be kicking yourself if 1-2-3-4-5-6 are the winning numbers.

  2. You wouldn't be rich if you found oil in your back-yard; the Crown would own it. Better to buy a lotto ticket!

  3. Will YOU do my housework for $5? That would be a sweet deal.

    Also, I would like to see a statistical analysis of the possible correlation between degrees and common sense before you make any big calls based on a sample size of 1.

  4. The article seems to have changed..!

    It's missing the luckiest numbers bit.