This weekend, I started improving the simulations (under Simulations in the top menu). The Most Played and Least Played simulations now have updated search options to fine tune the data.
The big change was a shift in thinking. My original thought, when I build the first simulation versions, was to use historical data to test my assumptions about whether least played or most played numbers are the better strategies. The first version waited for 7 historical draws before starting generating evolving more or least played numbers. The advantage of this approach was that I could actual historical data for calculating my most and least numbers AND use historical data as draw numbers to validate the model. My issue with this approach is the limited historical numbers to work with.
The big data theory is more of a long game approach. The historical data includes a draw history of 3742 unique draw numbers (or 0.027%) out of 13,983,816 possible combinations. That is not enough, I believe, to fully test my theories. Instead, I am using the historical data as the base for determining most played or least played numbers and then will use quick picks as future draws. The advantage of this is that I can look at the strategies over a very long time.
We’ll see how this plays out Already, I see a bunch of ways to keep evolving this simulation. My most recent tweak will be to allow visitors to base the data on all lottery data, just the weekday draws, or just the weekend draws. That tweak should be ready early in the coming week.
