Score Calculation - NEED EVERYONE'S HELP PLEASE

[BustaCap]

ok, this is going to require some serious help and thinking. I have a program that's going to be used by a math teacher teaching a very basic math class. The student can choose the range of radomization for both first and second digit of a math problem and can choose the operator (+,-,*,or /) and the number of problems. i.e. x + y = ? What I need is a formula for finding a score in integer form so that it can be more of a game for the kids than a learning tool. It might need to incorporate:

-total time taken to finish the certain number of problems (maybe)
-average time taken for each problem (measured in hundreths of seconds)
-operator used
-the highest and the lowest inputed limits for both first and second numbers of the problem (maybe)
-the number of problems that were answered correctly (out of total number of problems)
-anything else you might think would make the score make sense and be fair


Hopfully someone on this forums has done something like this before, I know it is almost and art form to incorporating this many variables into it and still make it logical. The thing I have been having problems with is the fact that (until know) someone could sail through the problems (putting in the wrong answer on purpose and answering the last two correctly) and with an average of less than a second, could come up with a much better score than someone who got ALL the problems right, even if THEY got a good average time.

Thanks for anyones troubles in advance. I would greatly appreciate a working formula.

-Matt

[David_Jones]

Hi Matt,

No specifics here, but if you find that someone getting most of them wrong is getting a better score than someone getting them right, then I suggest you just re-weight accordingly.

In a (not entirely) general form, if you have n variables (V1 through Vn inclusive) that you want to take into account, then you could weight thus:

Score = K0 + ( K1 V1 ^ P1 ) + ... + (Kn Vn ^ Pn)

Where:

K0 to Kn are constants to be changed in order to get the weighting right (see above your point on getting the answers correct versus getting it completed quickly), and P1 to Pn are powers to which the variables are raised to add an additional weighting mechanism.

Really, I suggest you play around with some sample data until you are consistently getting scores that 'feel fair'.

Hope that helps, and apologies if it was too general / basic for what you were after!

David.

[BustaCap]

I have attached my half (maybe less) finished program. Interface will change and there aren't many error handlers, but this might help to better understand my situation.

on the results window, the four numbers are: number correct, total time, average time per problem, and score, respectively.


David, I'm not sure how or if your formula will work in this circumstance. Thanks, but do you have a more specific example?