Aiming algorithm

[Arthanor]

A larger deviation (larger success area) makes cover more effective as the likelihood of there being some cover in the area is larger. The smaller the deviation, the more likely you are to snipe out toes from under a bush.

Otherwise, I agree that a precision property for weapons would be really interesting. It would combine well with the firing accuracy property that we currently have to represent the whole process. A sniper rifle could be a really precise weapon that requires a high accuracy soldier to use (a sniper should be a marksman), whereas other, less precise weapons would be more forgiving (down to the shotgun which is not very precise).

It would also deal with the issue above, someone who wants cover to be more important would simply have to make weapons less precise (= larger success area = cover is more meaningful) through a mod.

[vlad]

I would like to point out one more thing. I dont particularly like the idea of aimed shot getting special treatment.
The difference between auto, snap and aim shots (when ignoring amount of projectiles) is in TU needed and chance to hit with the specific shot. Those two are interchangeable in meaning - you can't expect great chances to hit when pointing gun in general direction of enemy and pressing the trigger in split of second and you also can't expect low TU usage when you need to aim carefully and guess whats target going to do next, predict his movement.
I dont like the idea volutar's proposition would be applied only to aimed shot. If anything, you are always going for the most visible part whether aiming or just spraying. The part where we differ the shots is in TU/chance to hit, not the aiming points. That part is already covered by hit/miss system. If chances for auto and snap needs to be adjusted, so be it. Or if only agents are affected by artificial number reduction for some shots. Or the larger/narrower deviation. Your call.

[Arthanor]

Aiming for the center of the most visible target area would indeed make sense for all kinds of shots. Whether you have time to aim or not, you would aim for what you see: the visible part of your target.

Maybe what volutar was saying regarding snap and auto shots is that since they are rather inaccurate to begin with (low hit %) it won't make as much of a difference as for aimed shots (which can often reach 100% to hit)?

It would be silly to keep 2 different aiming mechanisms when one is obviously inferior to the other.

[volutar]

It's not yet implemented, it's still in preparation stage.
And obviously noone have tested that, and noone can judge by practical results. Neither and I can.

[hellrazor]

I don't think it should be optional, volutar's target placement is clearly superior to the current way of doing things. (And I've mentioned that in every post I've written I think?).

Indeed it is. So it shall be done!

[vlad]

If that was implied before, I just had urgent need to point it out explicitly. No other intent behind it. Tbh from what i have seen so far this idea has support, with some adjustment for "cover" supporters.

[Dioxine]

Yeah, even a cover-loving, vanilla-hating enemy of freedom and democracy like myself understands the importance of Volutar's work here. This is not about "optional" system, this is about basic hitting algorithm which should be perfect. I don't care much about vanilla'ess (it was then and now is now), but an imperfect algorithm makes the cover relevant for all the wrong reasons.

What I think should be optional is the "see by the eyes, shoot from the shoulders" idea which, saying it bluntly, I consider retarded (as is hip-shooting). Even when firing bursts you first catch the target in your sights, so every shot should be coming from the eye POV. And that's not because I hate missing or seeing the "no line of fire" message, but because this makes much more sense.

Only when we have a more or less perfect underlying algorithm, we can start adding an optional mechanism of random spread (with the maximum spread either based on weapon, a d100 + displayed hit chance roll, or some combination thereof) - as long as there is any brave cover-lover who is going to code it down

[Warboy1982]

this is a re-creation of the xcom engine. we're here to imitate, not to innovate. the firing/accuracy model MUST be accurate to vanilla, or we're missing the whole point.

i appreciate all the weird and wonderful things you guys want to do with it, really i do. if you want to alter the deeper mechanics of the game, please do. there's nothing to stop anyone from doing so, apart from dealing with C++, and the openxcom codebase is written in such a way that it's actually easy to mess around with. hell, i cut my teeth on C++ by creating my own codebase and mod so that i could do all the crazy things i could come up with. it was fun, and a good learning experience for me. i realize that not all of you are coders, so this may seem somewhat unfair, but i'd remind you that yankes is working on an "extended" version of the engine where he IS doing features by request, so maybe talk to him? (and maybe buy him a beer)

but when it comes to the master branch... there's only one major goal in mind: make xcom. openxcom is all about recreating the original, warts and all. the accuracy model isn't a bug we can fix, it's a well defined mechanic, and it's really not the responsibility of the project to be the "swiss army knife" of tactical turn based games. it's highly specialized. it's impossible to cater to you each individually, as you all want different things, and by trying to incorporate more and more features and options, we end up with a quagmire of code that's difficult to navigate and work with (see the "range-based accuracy" fiasco).

and i don't mean this to come across with any bile, i fully endorse the "a la carte" philosophy, and think it's really the best way to acheive exactly what you want, exactly how you want it.