Pick 3 Coincidences

by

196Nebraska has a lottery.  One of the games is “Pick 3”, where each player selects 3 one-digit numbers, and once a day a computer selects a winning sequence.  The player wins if they have the same three numbers in the same order.

Monday (Jan 19, 2009), the winning numbers were 1-9-6.  One lucky Nebraskan had bet the same combination, winning the top prize of $600.

Tuesday (Jan 20, 2009), the winning numbers were ALSO 1-9-6.   This time, three people had the winning combination.  (No, the person who won on Monday didn’t play the same combination on Tuesday.)

The news accounts of this event say that the two drawings were done on different computers (preempting the unspoken suspicion that there was some systematic error in the process).

According to the Associated Press, the probability of this happening are one in a million.  And indeed, there are 1,000,000 possible combinations of winning numbers over two consecutive nights, since if you just concatenate the digits you get all the strings from 000000 to 999999.

But to say the probability is one in one million means you’re measuring the probability of having 1-9-6 come up a winner in both drawings.  Most readers of the news story probably aren’t struck by that specific combination, but merely from the fact that the two numbers agree.  That sort of coincidence is far more likely to occur, with a probability of 1 in 1000 (since there are 1000 possible draws on Tuesday, and only 1 of them will match Monday’s combination).

The probability that two consecutive draws do not match is 999/1000.  But the probability that after n+1 drawings, no consecutive pair will match, is (0.999)^n.   For n=366, we find that the probability of no matching pairs to be roughly 70%, which means that over the course of one year of Pick 3 drawings, there’s a 30% chance of having the same numbers win two nights in a row at least once that year.

In two years of Pick 3 drawings, there is a more than 50% likelihood that the winning numbers will match on some pair of consecutive evenings.

As one of my Philosophy professors used to say, “Some surprises are not unexpected.”

3 Responses to “Pick 3 Coincidences”

  1. Barry Leiba Says:

    The bottom line, of course, is that you have a 1-in-1000 chance of winning on any given day that you play, no matter what number you pick: whether it’s the number that won yesterday, the number that won on your birthday, a number that’s never won before, or your cousin’s IQ. If you play “666” every single day, or “000” or “123”, you have the same chance: 1 in 1000. Every time.

    And yet one just can’t get that through most people’s skulls.

  2. David Petersen Says:

    Here’s another interesting thought about the lottery. Basically, we are assuming equal probability for all possibilities. And for a computer picking the numbers, maybe that is true, but not quite so for the old ping-pong balls. The ink used to print on the numbers will weigh more for 2 digit numbers (and more for 8 than 1, for example), so there is a very tiny physical disposition for those to fall to the bottom of the pile.

    Similarly, we are *assuming* equal probability, but maybe there are other small factors that could affect the outcome, so I would guess that numbers that have already won may be slightly more worth playing as we know it’s a combination that *can* work. If a combination has never come up before, maybe there’s a reason!

  3. laurence Says:

    i keep track of pick 3 numbers in kentucky and thru feb 27, 2009 the ‘2’ has not come up in the 2nd position – i.e. x – 2 – x – in 108 draws this year. strange indeed.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: