I am currently developing a boxing pickem system for me and my friends in coldfusion i plan to release it on riaforge but thats another discussion all together.

I really need some direction with the psuedo/logic code for the scoring system, I have everything else done. My user profiles, event listing, match listings, result listigng,etc....

Once again im not looking for code, just the logic behind the scoring system so i can wrap my head around the actual programming problem.

Do I start each user with 20 points per match and subtract points based on the actual results... so if they get the match winner wrong i subtract 10 and if they guess the round subtract 5 and so on. So if they chose all the wrong outcomes they get 0 points?

Or should i great a function to compare the record with the actual results and users picks then add points for things matching and base the point system this way?:confused:

Scoring systems are a lot harder than i thought... then again I have had this problem in my mind for the last three days as I have been programming this thing for school. So it could be i am over thinking it. :confused:

Thank you for reading.

There are two ways to approach this type of scoring problem:

1. Rack your brains for several days to try and think through the best algorithm, or

2. Come up with a number of likely algorithms (like the two you've already defined). Collect a decent-sized sample of real-world data and run it through each algorithm. Go with the one that yields the most logical results based on input data. Tweak if necessary. (Otherwise known as the Mann-approach :stickout:)

There are two ways to approach this type of scoring problem:

1. Rack your brains for several days to try and think through the best algorithm, or

2. Come up with a number of likely algorithms (like the two you've already defined). Collect a decent-sized sample of real-world data and run it through each algorithm. Go with the one that yields the most logical results based on input data. Tweak if necessary. (Otherwise known as the Mann-approach :stickout:)

Thanks! I am taking the second approach... i am basically taking the results row and comparing it to the user pick row and +1 for all the correct ones.... Looks like its working so far with the sample data.