I've noticed that TX lottery winning numbers are not evenly distributed, for example since 1992, the values 51, 53, and 54 are very rare while the values 1-44 are fairly consistently recurring (approximate same number of occurrences). Starting at 45, there is a dramatic progressive average decline in frequency of occurrence. How best to make use of this information in prediction? Of course it is easy to predict the most frequently winning numbers, but this bias is interesting, although maybe not very useful.
> I've noticed that TX lottery winning numbers are not evenly > distributed, for example since 1992, the values 51, 53, and 54 are very > rare while the values 1-44 are fairly consistently recurring > (approximate same number of occurrences). Starting at 45, there is a > dramatic progressive average decline in frequency of occurrence. How > best to make use of this information in prediction? Of course it is > easy to predict the most frequently winning numbers, but this bias is > interesting, although maybe not very useful.
Try posting to one of the sci.stat.xxx newsgroups. Even better would be to include
a table of the number and its occurrance rate. Use a couple columns to keep
the visible length down. You will most likely find someone who has either
studied similar situations or at least used it as an example.
There were numerous studies of the US draft lotteries when that was a burning
issue of the day.
> ... Starting at 45, there is a dramatic progressive
> average decline in frequency of occurrence. How best to make use of this
> information in prediction?...
In the crudest sense, simply remove those "low probability" numbers from your sampling population.
As a practical matter, what would the odds of winning be even if there weren't those numbers in the overall game? Would you think your individual chances suddenly got all _that_ much better? (Yes, they're (mostly) rhetorical questions. :) )
I just got a spate of e-mails from another fella' who claims he's got the magic formula, too. I'll be happy to share it; all it'll take is your supplying him w/ your credit card information just like the UK lottery official w/ the inside track or the Nigerian diplomat. <vbg>
Gary Scott <garylsc...@sbcglobal.net> wrote:
> I've noticed that TX lottery winning numbers are not evenly distributed, > for example since 1992, the values 51, 53, and 54 are very rare while > the values 1-44 are fairly consistently recurring (approximate same > number of occurrences).
You also should study the distribution of numbers chosen by
lottery ticket buyers. They are well known to be non-uniform.
That affects the probability that if you win, someone else also
wins and you have to share the jackpot.
Otherwise, you need to compare the probabilities with the expected
distribution.
>> ... Starting at 45, there is a dramatic progressive
>> average decline in frequency of occurrence. How best to make use of this
>> information in prediction?...
> In the crudest sense, simply remove those "low probability" numbers from
> your sampling population.
> As a practical matter, what would the odds of winning be even if there
> weren't those numbers in the overall game? Would you think your
> individual chances suddenly got all _that_ much better? (Yes, they're
> (mostly) rhetorical questions. :) )
> I just got a spate of e-mails from another fella' who claims he's got
> the magic formula, too. I'll be happy to share it; all it'll take is
> your supplying him w/ your credit card information just like the UK
> lottery official w/ the inside track or the Nigerian diplomat. <vbg>
I understand that the balls are lined up to check them, prior to insertion,
so there would be some kind of bias against the intention of the air
pressure mechanism to actualy randomly disperse and select the balls.
My guess is that the later (higher-numbered) balls inserted, stay higher on
the pile, whearas the selecting mechanism is down below.
In theory, a computer program, using a fairly high-speed video camera (I've
done that part and have patents) could detect the numbers on the balls in
real time (just like fighter or tank images) and use a high frequency sonic
beam (this time) to influence the selection by diverting towards the edge
any unwanted number (enemy unit.....).
You know, you invent a tool and they go and use it as a weopon..... ;o)>
