The Importance of Inaccuracy

Seriously? I just had a game of TA last night, and that wasn't even close to my experience.

Big berthas and Timmy's miss targets. Quitch hit the nail in the head when he mentioned about the amount of targets. No other unit misses, you just need to tell it where to attack as shots don't lead.

The main problem in pa is the units being unreliable. I go into a 1v1 hummingbird fight and haven't a clue what is going to happen. Whether it hits or misses is completely random And game changing.

No unit misses in TA since all you have to do to stop them missing is micro?

sheesh...

tired of homing shi t. I want my sim pro back! and not "gimmik" sim pro; I want my Game-making Sim Pro.

Apart from air missiles, nothing homes in this game. What are you talking about?

Apologies for the bad English man. No other unit is inaccurate.

Sheesh yourself.

so you guys want chances to miss? go play a game with dicerolling mechanics ... such as bloodbowl ... do you realy want nurgle to mess with your play ...
i sure dont ... giving a artillery unit a percentage to miss or field of inaccuricy is no different from dicerolling that adds randomness to the game .... randomness that is not wanted ... units can dodge .... it is just a matter of how the projetile launched towards it behaves ... is it direct, low arc, high arc, tracking, hitscan, quick, slow?

Not "chances" to miss. We would like simulated (in)accuracy; there's a difference.

Anything with a projectile had variance in its shots. Guardian/Pulveriser, Battleships... hell just plasma weapons in general... even EMGs were inaccurate; they all missed regularly on a moving target and if the target was small enough they could even miss a stationary one. Considering that about 50% of all the units in a game wielded plasma weapons, you're asking a lot for me to not believe my eyes every time I boot up a game of TA.

As I said if you manually aim they are accurate. This is not the weapon, this is the user/game code. Saying a sniper rifle is inaccurate because you can't shoot it properly is not a good argument.

Just remembered terrain height was an issue with the guardian/pulverizer on certain maps. Edit: and again this is not a problem if yo know how it acts.

I'm afraid that is not the behaviour that I observe when testing your hypothesis. Even when manually aiming, ballistic arc weapons fire is not 100% accurate. Close I'll admit, but not complete.

Don't they have AOE?

All weapons in TA have AOE (technically). Damage fades over distance linearly based on increments of pixels, so even the tiniest variance in the accuracy of a shot has implications.

While there is some "missing" in PA for sure (especially in regards to artillery), as a whole, most of the units are pretty tight accuracy wise, especially in comparison to TA. Tell a Brawler to attack a unit, and it would spew plasma all over, maybe a third of the shots would miss. These sorts of mechanics allow for stronger use cases and more strategic leeway.

Right now all the ranges are hard; get in range of an enemy unit and even at max range, for most units their accuracy percentage is very high. This forces the game to be about strong range management, adds more micro to ensure effectiveness, and adds more of a hard rock-paper-scissors component. Combined with the fact that a lot units also die almost instantly, this means if your attention is anywhere else for a split second, you basically have no recourse, and are punished hard.

Some people want that I suppose, and they're welcome to that opinion, but I'd prefer that given the size of the game, a bit more leeway was given in this regard to make confrontations more interesting to watch and play. Don't get me wrong, I don't want this to be a tanky game where units never die either, but right now I feel that strategic options are a bit too limited combat-wise due to the current balance design, and a stronger focus on inaccuracy would help fix that.

@nanolathe and what exactly shows you weither projektiles between supcom ta and pa are realy simulated or not ? how can you tell that for each and every projektiletype?

well I can answer for supcom =


you could even zoom in and see the freakin projectile's 3D model.

this applies to even the t3 PD which fired at an insane rate per second.

plus the collision box was the exact model.

projectiles would go through the commanders armpit or in between his legs (no damage delt) and hit the ground further behind.

A projectile that is simulated means that, once it is fired, it follows a trajectory that is independent of the unit that is being targeted. If the target changes direction, teleports, vanishes, or dies, the projectile will not care, and will happily continue on its path (whether that path is based on real world physics or "magical" video game physics is irrelevant). If something solid moves in front of the projectile, the projectile will collide with that object instead of moving on to the targeted unit (i.e. collisions are calculated). Even if the projectile is "homing" it will still have a trajectory that is limited by some form of "physics". It will have a turning limit and velocity limit, and collisions are calculated.

Non-simulated projectiles do not behave like this. The game essentially rolls a dice, and this determines whether it hits or misses (or some other possibility). From there, the game visuals try to show as best as possible that this is happening. If the game rolls a hit, and the target teleports to the other side of the map, the projectile will follow suit, changing direction immediately, clipping thru terrain, and going potentially enormous distances.

Very well put @krakanu

I'll add that the two have never EVER (as of yet) been at a stage where they were visually indistinguishable

you can always see the difference very clearly.

so what i generaly read out of this is that someone sets a projektile to work to certain way ... "this is a artilleryshell, it gets shot out of a gunbarrel and flighs that high on that degree before it falls down in that angle but shall hit anything in its path and disapear because that is generaly how people do expect it working"
isnt that just setting a couple of parameters and then just run the simulation?
by that logic isnt what sorian does setting parameters to the ai and run the game to see what it does a simulation too?

You mean, apart from the game code itself? Oh I dunno, not much except also being able to physically observe the projectiles moving independently, irrespective of whichever unit fired it.