So I've been doing some fooling around with the
track_difficulties values in my database file and I found that, for me, using the 0.85 value rap99 provided wasn't quite challanging enough. I tried setting them all to 0.95 but noticed very little difference between that and the stock Legend. Seeing that the pro difficulties are not all equal across tracks, I decided to come up with a bit of an algorithm to generate new series of track-specific values for something between pro and legend.
I first tried taking the average between pro and legend for a given track, but most of the values fell between 0.84-0.88. That's higher than rap99's 0.85 value, but I wanted to try something a bit different as those values weren't all that dissimilar from rap's work.
I devised the following formula to calculate series of new track_difficulties values for each track:
So the new difficulty value is the average of the pro difficulties (the summation expression) plus some fraction of the deviation between legend and pro difficulty. The
d's represent the different difficulty values over
i number of tracks. The parameter
rho (the p looking greek letter) is some fraction between zero and one, though only values above 1/2 will give you difficulties greater than rap99's 0.85 value.
Using this formula I still preserve some of the unique differences in track difficulty present in the lower difficulty settings through the deviations. However, because I am using deviations, it means that tracks with a lower pro difficulty now have a relatively higher difficulty level. I did this to even out the AI performance. However, without testing I'm not sure if it'll wreck stuff up.
Anyway, I'll do some testing today and let you guys know how it goes.
PS: And yes, this is what a grad student in Economics does on the weekend for fun