How to choose best Combined Statistics

[mindy20]

Hi Stan,

There are roughly 180 stats available to choose from, we can choose any 5 statistics to combine them. That gives us around 1,488,847,536 combinations of statistics to choose from. Choosing 3 statistics to combine from 180 statistics gives us 955,860 combinations etc. Obviously, some combinations are more useful than the others in order to predict and filter the total tickets' package.

My question is, how do I find the best set of the 2,3,4 or 5 Combined Statistics? My aim is to find the set of statistics which has the maximum frequency, yet the minimum number of tickets after this filter is used in the package.

Would it be too much work to ask for a list of all possible combined statistics, so I can use an automated programme to go through them to find me the most ideal combined statistics for my purpose?

[stan]

Well, it depends what you call 'ideal'. If there are stats that you're comfortable with and have a good record of predicting their next values you can try adding a couple of those together.
But I'm afraid it's not technically feasible to evaluate all those millions of possible stats combinations to find the best one - even if I knew what 'the best' is...