Specifically, please look at what happens with actual damage for different combinations of opponents and modifiers. Arguing from a few IRC discussions is not likely to yield good results, IMHO -- damage calculations generate lots of hot air (the forums have several hundred postings on this topic, many with reasonable sounding arguments), but difficult to get right in practice. Generating some damage tables and checking that they give desirable results is likely to be more productive in this case than throwing something at the wall to see what sticks.
I agree with you regarding the pertinence of having some solid data to back up one's assertion. The latest version of 'dr' I've found is 1.17. It does not seem to be compatible with 1.1.2 units files anymore. Am I correct when I assume that 1.17 is the latest version?
I have read the discussion about the RATE system and understand why so many people feel that changes to the additive multipliers will unbalance the game. One problematic case seem to be 0.75 * 1.25 == 0.9375. Also, 1.25 * 1.25 = 1.5625 (which does seem close enough to 1.5 though). Taking into what Sapient said about CMULs being the same as FMULs, the formula compatible with Xan's changes become damage = (((base_damage || SET) + ADD_1 + ADD_2) * MUL_1 * MUL_2 * (1.0 + ToD + Leadership)) This is (much) simpler than my previous version (phew!) and the user can understand it eventually. Yet it is still more complex than an all-multiplicative solution. I realize that there is a strong opposition to changing the damage computations for fear of breaking the units & campaigns balance. As Ott, Soliton, Xan and others pointed out, any change to the algorithm should be validated with real data showing the consequences of this change. Ott's 'dr' script seem to be the best solution we have so far to provide this data. I can try to put it up to date if 1.17 really is the latest version... That way we can test proposals such as Darth Fool's suggestion. About this argument: > The algorithm is already complex enough that the > argument of saving time by calculating them in your > head seems a little unrealistic; after all, the computer > calculated answers are shown to you anyways. That's not always the case. The computer calculated answers are not available when you plan an attack in a hypothetical situation. Some (most?) people could care less about the actual damage value; they use a rough estimate. However, for some serious MP players, getting the real damage value *is* important to answer questions such as "if I move my leader there, can my opponent actually kill it by using this and that unit at night with leadership?". In this case, an algorithm which is easy to understand & to compute by hand is a real benefit. Laurent Birtz -- a serious MP player :) _______________________________________________ Wesnoth-dev mailing list [email protected] https://mail.gna.org/listinfo/wesnoth-dev
