The curious case of accuracy_vs_x.

[The Great Khan of Rome]

accuracy_vs_units and it cohorts has puzzled my as unit modder. Every since I unleashed unreal into the modding space, I had some questions about that stat. It's counterintuitive; smaller numbers give more accuracy, higher numbers give less. I've had to try figure this out someday. I don't know if anybody else in unit modding knows what this stat exactly does. I don't think most unit modders know what this does. I've come up with two theories in mind when looking at this:

Target Theory: Units target each model of a unit and they use accuracy_vs_x for a percentage of projectile deviations from the target.
Possibility/Truthiness: Possibly accurate, at least for short range (about javelin range, so 55-75 in EDU). Over longer ranges, this theory seems to make less sense. The range thing could even illusory or part of a bigger picture, like the angle, velocity, or missile type as specified in the EDU. This may also explains how it's used; smaller values indicate smaller chances for induvial projectile's to deviate and vice versa. The amount of deviation and how that's calculated is unknown. It may just point to random spot around or near the target

Area Theory: Units target an area around the unit and they use accuracy_vs_x for a percentage of projectile deviations from the target.
Possibility/Truthiness: About the same as Theory 1, with accounting for longer ranges. Another version is that the accuracy_vs_x are for the percentage of the area the unit will target. Basically, a units will target an area, usually the units center, and then target random points around that when aiming. The accuracy_vs_x measures how much area they will use for firing. This further evidenced by later total war titles and how they do missile accuracy. This is the theory the modding community accepts, with good reason; it works in most cases. I'll bring up two cases where it doesn't seem to work however. Individuals unit platoons and small unit sizes. These cases seem to support the 1st theory, unless one is to postulate that firing radius moves depending on size. This seems to be correct, according with personal experience of mine, so these cases may not support the 1st theory.


I will postulate another theory of mine that might take into account all potential cases.

Synthesis Theory: Both Target and Area are true, but used in different circumstances.
Possibility/Truthiness: A bit of conjuncture. Basically, when missile units/models target a unit, models that need to arch upward in order to fire their projectiles (how I don't know), their projectiles seem to follow Area Theory. When misses unit models have sight of unit and not need to arch upward, are firing downwards say in the case of using hills or walls or are in charge range, their projectiles seem to follow Target Theory. I can't vouch for this that much. This is coming from experience and and intuition based on said experience.

I would like to see units modders take a look at what I have to say. I would greatly appreciate some professional opinions on what I have to say. Personally, I think Area Theory is true, like the rest of the community here. It seems to best account for what's there. Theory crafting and figuring out med2's battle engine are what is fun about trying to mod this nearly 20 year old game.

EDIT:
http://www.twcenter.net/forums/showt...ctile-accuracy by Gen.jamesWolfe

now until relatively recently, this was not a well understood value, However, after myself and Paedric did some research on the topic a while back, we were able to figure out that the accuracy represents the maximum angle a missile will deviate from where it is meant to go, in radians (assuming the projectile is flat in trajectory).

It seems somebody has put in the effort already. This happens every time. I don't look hard enough I guess. Thanks James! It seems to be like later total titles where they had a value for angle of scatter. So accuracy_vs_x is misnamed, like a lot of M2TW values. It should be called maxium_spread (in rads according to Paedric).