Wish I wasn't a newbie so you could see the site I've been working on (can't post a link according to the site rules). For my base games, I have Powerball, FL Lotto, and Megamoney up and running. It pulls down the new numbers automatically and updates the base statistics for each game. I designed the framework so new games can be added with only a few hours of work (the site pulls in all of the historical data and crunches the numbers). So now I'm trying to expand the other way: What other interesting statistics can I add for each game that would be beneficial. The base stats I have are:
-Last games out: Statisticsly speaking, there's a 40.64 chance of a zero day out (number that came up on the last drawing) number occurring again for Powerball; based on almost four years worth of draws. So if you pick from the last 5 numbers (not including PB), there's a 40% chance one of those five numbers will hit vs. trying to figure out which of the other 54 balls have good odds.
-Number breakdown data set: This is a quick view of each number that shows the draws out, previous draws out, how often the number has occurred over 7 days, how often the number has occurred over 30 days, occurred year to date, occurred over 2011, and occurred over 2010. It's just a nice way to view what numbers might be hot or cold.
-Sums (not including PB/MB): Based on the historical data, we can see there's a sweet spot for the sums. If someone were to pick 1-2-3-4-5 for powerball, that's really bad odds (it's never won before). The sweat spot for powerball sums is 116 to 190. That's a nice gap, but having a sum within that range gives better odds.
-Odds/Evens: This data set shows there's another sweet spot for picking the right amount of odd and even numbers (again, the PB and MB are not included in the stats for those games, it would only sqew the data). I have over 400 records in my powerball table. Out of those 400 draws, all even numbers has occurred 9 times; not good odds. But, 2 odd/3 even has occurred 135 times. 3 odd/2 even has occurred 128 times. That doesn't help much with picking good combinations but when I finilize my numbers, I always make sure I have a good split; better odds that way.
-Matching numbers: Again, using powerball as an example... out of over 400 draws, a winning number has repeated zero times. So I always check to make sure my final play numbers haven't hit before.
So that's why I have for each game so far. Now I want to add more data sets. I think I'll add an 'out' data set to show what's been popular over the last 30 days or something. I'm curring writing a query that will tell me which outs pair well when a zero day out number hits (since there's a 40% chance of that happening for only five numbers). After that, I'm not sure what other data sets would be useful. So I was hoping a collective such as this would have some ideas. Looking for anything really. No such thing as a stupid idea.