Winning the Lottery

Date: 01/30/97 at 06:33:30
From: Zink, Steve
Subject: Winning the Lottery

I would like to know what branch of mathematics I need to study in 
order to be able to write a formula that will approximate the outcome 
of the lottery.  I know that it sounds far-fetched, but please hear me 
out.  My lotto is comprised of 6 numbers: A, B, C, D, E, F.  In 
reveiwing the history of all the draws, I can see a type of pattern in 
the A's column. For instance, approximately every 10 to 13 draws A 
will equal 10. The other numbers in A sort of do likewise, but there 
are too many for me to keep correlated in my head. Could you please 
tell me where I might go to learn how to write a formula that predicts 
the outcome, if there is such a thing?
                                                                  
Thanks,
Steve Zink

Date: 01/30/97 at 10:13:44
From: Doctor Mitteldorf
Subject: Re: Winning the Lottery

Dear Steve,

Scientists are always looking for patterns in physical data, and 
market traders are always looking for patterns in stock prices.  
Sometimes our eyes make up patterns in data that's really quite 
random. So mathematicians have come up with rules that can tell 
objectively whether a pattern is really there, or whether it's just 
a chance combination of numbers.

The rules aren't straightforward and simple, because they depend on 
the context. If you're serious about this, it will be worth your 
while to take a course in first year statistics, or read an elementary 
statistics text; there will be lots of examples of real number 
patterns and false patterns that look real.

This should be an important concern of yours before you actually put 
money down in the lottery based on your ideas: check that the pattern 
you are seeing is "statistically significant". That's the term that 
means it's really in the data, and not likely to be just a series of 
random occurrences.

In the example you give, suppose the first number A can be anything 
from 00 to 99. Suppose you see the number 10 come up once, and then 
again 13 days later. This might seem a remarkable thing to you, but 
it is not "statistically significant" because in any set of 14 
numbers, there are 91 different pairs of numbers, so the chances are 
91/100 that one of the pairs will be the same. If the number 10 came 
up in the same slot a third time within 13 more days, that still 
wouldn't be statistically significant; only after the fourth such 
occurence would you really start to suspect that it's a pattern you 
might be able to bank on.

Statistics has a lot to tell us in situations like this, and, as in 
the case I just cited, much of it isn't really in line with our 
intuitions. I strongly recommend that you learn enough to understand 
statistical significance before you risk any money on a pattern that 
you've found in lottery numbers.

-Doctor Mitteldorf,  The Math Forum
 Check out our web site!