The wizard make target is a slightly out of date thing for the instructions to refer to really, since it's the default anyway now.
reaver wrote:Second, I don't know how closely I need to match the old formulas. Currently, I've worked based on the 1% rule- if the new formula changes the result in less than 1% of cases, It's fine. Please tell me if this is too large a margin (so I can adjust the work I've already done.) or too small a margin (so I can make the code even simpler at the expense of slightly different behavior).
That seems fine to me, from what I've seen of makeitem a lot of the probabilities would be extremely awkward to replicate exactly. Also as lots of new items have been added it's doubtful the probabilities there are perfectly balanced anyway. If the file is better organised it should be easier to subsequently review and tweak the probabilities anyway.
I'm not sure how you're implementing it, but what would be quite nice is a big table of data like we have for mon-pick so the relative weights of different items can be seen at a glance. It's even worth considering using the mon-pick code (that I refactored into random-pick.h for exactly this kind of purpose) so we can weight items e.g. by depth/branch. But at least getting things into a form where we can
see what the probabilities are is a good start
Something else worth checking - have you run any item generation stats? Would be good to get a sample of many item generations (there's probably already a script that does this) both before and after your changes.