Message Boards Message Boards

2 Replies
2 Total Likes
View groups...
Share this post:

How would one go about analyzing Lottery draw statistics in Mathematica?

Posted 11 years ago
So, there are probably plenty of different things that COULD be analyzed about historical lottery draws (as a 'for instance': PowerBall & Mega Millions).

I'm interested in some general statistics, I guess? & progression over time. All "just-for-fun," of course? ^_^

So things like:
  • figuring out and/or visualizing how many times a given number has 'hit' either within ALL draws, or within a given timeframe, or on a given draw day-of-the-week or during a particular calendar month.
  • figuring out the longest, shortest and average streaks of a given number either being 'drawn' or 'not drawn' on consecutive draws or on consecutive same-day-of-the-week draws. Maybe mean / median / mode type information too?
  • perhaps graphing the change in 'total hit distribution' over time or 'hit distribution over the last X draws' over time?
  • finding the 'intersection' between different 'hit distribution' sets (for instance, numbers that both have a history of not being drawn overall AND whose actual 'hits' have a history of not falling on drawings on a given day-of-the-week)
  • What the typical distribution, in an average draw, of numbers with a prior record of high hit rates, medium and low hit rates?
While I'm not precisely new to Mathematica, there's a lot of stuff I just haven't done and don't know what the best methods / functions to use would be. Is there anything specifically built-in that would do this kind of analysis automatically, or must we custom-build solutions by hand for each such thing we're interested in?

I assume we'd start by in some manner specifying a file to import or use as source data? Probably an Excel .xls / .xlsx spreadsheet or a .CSV file? If so, do we leave things in the source file and simply do maths in Mathematica, or can we actually "import" the data in same way, so we don't need the source data file anymore (though we might want to analyze subsequent draws, to see continuing trends and whether any predictions(s) end up being true of future live data or not, so perhaps we'd want to leave it as an "operating on external source data file" situation)?

So, I guess I don't have just a "single" question, but a bunch of related "How would I do X" questions... I guess this could be a thread for discussing all things "Lottery statistics" in a general sense?

What would be the best way to handle these multitudinous question in this thread? Tackle one problem at a time? Folks offter suggestions on which ever problem suits their fancy at the time? Break out individual problems into their own threads (seems wasteful and potentially fractious, making it hard to find the various solutions)?
POSTED BY: Michael Gmirkin
2 Replies
The questions which are practical and narrow are easier to answer. I would research online if something was done already, if there are typical algorithms and data. Then I'd start working one step at a time and ask if a specific issue is encountered. Also take a look at this MathGroup thread and this Wolfram Demonstration: A Canadian Lottery Wheel. I would also take a look at some WolframAlpha data, for example.
POSTED BY: Vitaliy Kaurov
Data analysis usually starts with some sort of data.  Perhaps you could at least do the footwork of finding a particular chunk 
of information that you would like to investigate.  For example, the winning lottery numbers for Illinois can be found here:

Once you have a file in hand or at least a web page of interest, I'm sure that further questioning will be more fruitful.
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
or Discard

Group Abstract Group Abstract