On the topic of scale...

I was referring to punishers. And while they may not be able to hit an army of peewees traveling full speed, as soon as the army of peewees stops, they're ALL dead. And if they have a decent sized army, the shot is aimed at the front and still decimates the back. Especially since armies become strings in larger groups, so the shot gets aimed at the front of the string and hits the middle/back.

And in TA:T, there's just units that are so powerful they completely invalidate T1 units... Like fat boys. A large group of fat boys and demolishers could kill almost any amount of T1 units. Lets not even get started on experimentals. Maybe 600 T1 units could do something, but you have to remember that the simple fact of having 600 units IS a detriment. Even if it was balanced so an amount of T1 units equaling the cost could kill it. It's pretty darn annoying to build 500 of something.

My main point is that the cost/size/power between the weakest and the strongest unit shouldn't be absolutely crazy. And we need to be really careful with splash, so as to not completely invalidate hordes of weaker units.

I am pretty much fine with the balance of orginal TA. A goliath/bulldog can get killed by T1 units. That's how it should be. Guardians always seemed a bit OP to me, but that's mainly because computers aren't smart and I never had to pit them against human opponents.

Punishers were designed as an anti naval weapon. You're not meant to use them against land units. Unless you do. Then it's totally ridiculous.

There are two problems happening here.

The first was one of size. Punishers were really small with long range, so they stack very well with any army in any arena. They were difficult to approach and difficult to destroy as a result.

The second problem is a design disconnect between land and navy. Both theaters happen at the same altitude, and share a common projectile physics engine. So when you design one arena to use that system differently from the other, it gets all screwed up at the coast line. Land defenses become useless against water, and simple coastal defenses end up overrunning everything on land. That's bad.

In PA, any planet can have any arbitrary layout of land and water. It's important to keep a sane interaction between both arenas.

Don't fret over the balance of mods. Most mods are not built with the same strict game design principles as professional devs. Michael Bay guides custom mods to ensure large explosions and super mega oversized kill engines.

Actually, TA:T was derived from Absolute Annihilation, which derived from Uberhack, the original balance mod for TA. For the most part, it's a LOT more balanced than vanilla TA, and also gives the CORE and ARM greater unit variety.

The super units are very powerful indeed, but have you considered the cost of building them? They're even less cost effective than the Krogoth, and most can be taken down disgustingly easy by cheaper units. They have their uses, but for the most part they should never become a focus of gameplay, because otherwise you're wasting resources that could be put to better use. Here's a (very old) forum post about some of the original balancing work going into those units:
http://www.tauniverse.com/forum/showthread.php?t=36736

Also, with regard to the static defenses, the very first question on the TA:T FAQ is about that particular topic:
http://www.tauniverse.com/forum/showthread.php?t=39562

Bear in mind that AoE weapons, assuming they have large AoE, are designed mostly to kill T1 units. One of the reason larger, bulkier units (read: The Bulldog) are more survivable is because they aren't as heavily affected by AoE even when they are all bunched up. Yet another reason more survivable and more heavily armed units should be just straight-up bigger than lesser units.

That begs the question as to whether aoe weapons will have a damage drop off instead of the more traditional radius of damage x.

TA seemed to have a linear damage drop off. It was Supcom that used flat damage across the entire radius.

Starcraft used a layered drop off system. Center point takes full damage. There were larger 50% and 25% points as well. This gives a decent enough falloff curve, without having to worry about problems of extreme precision.

Should be quadratic decrease.

Bear in mind though that quadratic or exponential drop-off requires a lot more computing power than linear or flat. PA will already be taxing, we don't need MORE math for the processor to do while it is already simulating orbits, projectiles, laser beams and exploding planets in 3d space, in real time.

It really should depend on the size of the explosion. Small explosions like rockets, plasma shells, and tank guns should be flat. Larger explosions, like tactical missiles, artillery and flamethrowers should be linear. Only stuff like nukes, experimental artillery, super flamehrowers and space rainbows should be quadratic.

What can a quadratic explosion do that a linear or stepped explosion can't?

The whole point of having damage drop off is to limit the AoE lethality of explosives. Pick a system and make it work.

Quadratic explosions are more realistic. This was the reason the Cybran T2 gunship was so good, it used a flat explosion that murderized grouped units.

Well not really, computing wise they are pretty much similar. Also you could use >> for exp of 2 which is pretty fast.

Of course. They're realistically terrible. Most of the area deals very little damage, with one big spike in the center. That pretty much defeats the purpose of AoE.

You do realize that you can scale a quadratic explosion right? The point of them is so that you don't have a HEAT weapon killing 4 tanks at once, and you don't have the background radiation from a nuke half-obliterating an outpost on the very edge of its damage radius, and leaving the other half of the outpost unscathed.

And I'm also all for linear explosions. On the scale we're talking about, linear is usually better in every single way than quadratic.

But again, quadratic is more realistic. For example, Fragmentation Grenades have a 70 foot kill radius on flat land. Past that, you will be inflicting possibly lethal wounds, but not instantly fatal. Due to air friction, the fragments slow down exponentially. Thus quadratic is the most realistic, if not the most balance-friendly.

You know that the overall damage output of linear decrease is lower then that of a quadratic decrease?
Integral of 1-x = x-1/2x^2 that from 0 to 1 is 1-1/2-0+0=1/2
Integral of 1-x^2 = x-1/3x^3 from 0 to 1 is 1-1/3 -0 +0 = 2/3



General formula:

m == max damage at impact point
r == range of the damage
x == actuall distance of unit and impact point

linear : m-(m/r * x)
quadratic: m-(m/r^2 * x^2)

Difference of functions quadratic - linear:
m/r * ( x - x^2/r)
remember m/r are positiv cuz there is no negative damage or negative range
So for comparision >0 we can skip Case differentation and also eliminate m/r
Now for the interval from 0 to r, x is lesser or equal to r
So we can say x/r < r/r = 1
with - :
-x/r > -1
We now get:
x+x*(-1)<x+x*(-x/r)
Equals
0< x - x^2/r


Integrals:
linear: mx-m/2r * x^2 from 0 to r : mr- m/2r * r^2 - 0+0= mr-1/2 * mr = 1/2 mr
quadratic: mx-m/3r^2 * x^3 from 0 to r : mr - m/3r^2 * r^3 -0+0 = mr-1/3mr =2/3mr

please note that this is only for visualization as it shows the distribution of damage from the impact point to the surrounding in on straight line.

If needed i can calulate the whole damage output over the area also.
And we didnt talked about protection yet, maybe units behind units wont get as much damage as if there aint anything between them and the explosion, or maybe hiding behind rocks or buildings.

Edit:
For the Area it is the volume of a cone for linear and that of a homogenouse parabloid for quadratic decrease.
which leads to (overall damage output):
linear: 1/3 pi r^2 * m
quadratic: 1/2 pi r^2 * m

Edit2:
For interest i include also the damage output of a non decreasing AoE.
For the straight line it would just be
m*r
the area of a rectangle.
For the whole area it would be
pi r^2 * m
the volume of a cylinder

To come back to the scale topic:
Lets compare one big unit with the same health as 5 little units accumulated.
If a artillery with AoE hits the big bot it does 1 times full damage.
If it hits the bulk of the 5 little units it may do 5 times lesser or equal to full damage.

Which can lead to: spread out units by micro or scripts, or the conlusion that bigger units can be better against AoE
if bigger also means more health.

That logic is somewhat flawed. Bigger units *can* be better if we're just talking about tanking explosives, but Shields are most cost efficient for that. Not that PA will have shields.

But remember, assuming PA uses the system SupCom2 used for unit movement, small groups of units will fare far better vs. artillery because they can hide behind cover, take up less space than a megabot, and are almost always faster. Big blobs will be the ones that suffer the most. Which is a good thing, because big blobs are boring.

I fixed my post.

in terms of scale I think a more pressing matter is that the shortest unit is again almost 10 meters tall, which makes it very hard to sympathize with the scale. Big things that are a lot bigger than other very big things don't really mean much compared with big things that are big compared to something familiar on a human scale.

So I say tone down the minimum scale a bit so that you can at least see the humans on the scale chart.

Actually, the tanks that have been shown are basically the same size as modern abrams tanks, and the bots not that much taller than a tank (the largest T1/T2 bots shown are 3-4 meters high tops, most are around 2-3m).