How much does precomputation (matching a series of strings and their permutations with a set number of other strings) help, especially over a network ?

Question

Consider a typical slots machine with n reels(say reel1: a,b,c,d,w1,d,b, ..etc).
On play we generate a concatenated string of n objects (like for above, chars) We have a paytable which lists winning strings with payout amounts.

The problem is a wild character (list of wilds: w1,w2) which can replace {w1:a,b,c},{w2:a} ..etc. Is it really worthwhile to have all possible winning strings permutations with the wilds precomputed and used or simply at the time of occurance, generate all combinations with the pattern in hand accordingly.

I did'nt really see much difference initially, but now if I need to scale the machine to handle 11+ reels with a much higher concentration of wilds than previously, I need to figure out the exact approach for this particular bit.

Any ideas will be really appreciated :)

Answer 1

Your problem isn't quiet clear to me. You have reels - up to 11, which are filled with symbols. How many per reel? And some symbols are wildcards for 2 or more symbols?

Let's keep the wildcards beside for a while. If you have 10 symbols per reel, you get 10^11 possible combinations. That is: 100 (US-) Billion.

I guess some symbols are repeated on the reel?

What do you have to calculate? The winning probability? Whether a special pattern wins?

Example:

a a (a, b, c) (a, b) b  

This could be aaaab or aaabb ... - to aacbb. I guess, the more wildcards you have, the more possibilities you would need to precompute, which will need a lot of memory.

While the number of patterns to check will be done pretty fast, on a recent machine. Won't it?