Friendly Fire on Splash Damage

Actually is pretty easy let me write it here using a bit of math and code

X=Friendly units
Y=Enemy units
If X<Y Then
Shoot
else
Dont shoot
end

Thats it the code i use for most part is LUA

At the moment you do not have enough control over firing so maybe later.

So the general consensus is people would like friendly fire. and some minor counter argument is that the explosive units should know when friendlies are in the area and won't shoot.


Seems over exaggerated considering the number of times anyone has dox's lose in a base, the idea of your pelters destroying more of your base then the dox is a little too theory crafting to prove points on either side.

I think leaving this conversation to the point that " yes we would like friendly fire" is good enough for the dev's.

If they try it out and like it , then good. If it hits this Theory, then we will see, otherwise. these conversations are going to get no where unless we put the rubber to the pavement.

In that case who were you arguing against the whole time?

If that's the case then we agree, friendly fire galore!

Units never shot at their friends in TA or FA, at least they never did anything to make their projectiles hit a friend.

if they got in the line of the shot wasn't much to do was there?

Several people have stated that area of effect units should fire regardless of whether friendly units are in the way.

I never once stated that there should be no friendly fire. At the beginning of the thread I said I was undecided, and since stated that I would support it only if units were made intelligent.

I simply argue that units should determine whether they'll cause more damage to a friend or more damage to an enemy.

If the Hornet flying over head determines that the carpet bomb will do more damage to the friendly base than to the single dox below standing next to the line of friendly power generators, the Hornet will keep on flying and not destroy my own base.

why not. unless you can find a way to calculate it without killing the servers.

Why not? Because of everything I have stated multiple times in this thread.

Yet never did you address the effect it will have on performance. I rate performance higher than "intelligence" in this case, PA already being the beast that it is.

Because that is above my expertise. If I were to guess, I'd say it probably wouldn't be a huge increase in demand. Units already determine if an enemy is within range and how to shoot it.

Unless you or I are a developer who knows the ins and outs of the PA code, it isn't something we can comment on since we have no clue how it will alter PA's performance.

A reasonable stance but you would surely agree that a line-of-sight check is less computationally intensive than a line-of-sight check with a risk/reward or even only fire/don't fire calculation on top.

@brianpurkiss, A point was brought up earlier that was never addressed.

How does artillery work out if damaging a structure to destroy an enemy unit is the wrong thing to do or not?

If an enemy force gets inside a base and clusters around your structures, you're advocating that none of your artillery fires and instead lets your base be destroyed? Wouldn't some damage to your structures to destroy their units be better?

And this isn't limited to enemies in your base. If you have a small force attacking a massive enemy force in range of your artillery, should your artillery sit idly by, precisely when it's the perfect time to fire, in order to "save" your few friendly units which will die anyway?

Having your artillery stop firing when the enemy enters your base would be, in most circumstances, purely frustrating. Automated behaviour needs to be easily predictable. "Your structures will take damage if your artillery fires inside your base and it lands next to them, which they will do if an enemy unit is there" is utterly predictable, consistent and can be taken into account in your strategic planning and base construction. On the other hand, "Your artillery may or may not fire and may or may not lead to even more damage being caused to your base in some set of circumstances" is a very bad kind of automated behaviour, because you can't predict when it will do what, and can be just as exploitable as the other scenario.

BUT - there is a difference between friendly damage due to splash, and friendly damage due to projectile collision. The above argument deals with the former situation. For the latter however, it is the opposite - if there is a scenario where a projectile would hit your own unit before the enemy (therefore doing no damage to the enemy), then you would never want them to fire.

You wish it could be that simple. And even if it were, it would negate you a timely understanding of the situation, and the possible consequences.

It will be impossible to test it, but I suspect that given two armies, one that follows your rules of no friendly fire if, against an another that will not care about, the latter will be more likely victorious.

The reason being in how much easier would be for the second player to quickly grasp what's going on on the battlefield. Simple as that.

You see the problem?

Human vs human, everything is right and dandy.


Human vs AI agent and suddenly you have an easily exploitable weakness in the AIs code.


AI takes it into account - 9 AI FFA, plus a human. And the system performance argument is now invalid.

Which just leaves us with artificial stupidity.

Not actually sure which way you are arguing, as both ways (in addition to a lot of other things) are exploitable when the opponent is an AI.

Having said that, friendly fire by artillery really wasn't very exploitable in Total Annihilation.

Oh I knew this was coming, allow me to confirm, your "expertise" shuns even the term expertise.

calculating a million units PLUS their potential dozens of millions projectile's trajectories is already hard enough without adding to that running the simulation ahead of time, determining what each of those dozens of millions of projectiles hit then accordingly cancel or authorize the shot.

how could you even do it "ahead of time"? you'd probably have to halt time, the user would probably change the unit's direction at any moment.

Tell me with a straight face that those 9 AIs don't impact system performance and I tell you to do a bit of profiling.

Always a good thing to keep the KISS principle in mind.
http://en.wikipedia.org/wiki/KISS_principle

Actually I did address that.

"Will this do more damage to my opponent or more damage to my enemy?"

Simple math. If the total damage to my enemy is greater than my allies, then secrew 'em, fire anyways.

And naturally, you'd be able to manually fire or switch firing mods to shoot at everything.

Over what time period? Any unit can do more damage than an artillery round if left there to fire on your structures.

There is no "right" answer to this. If there was I'd be with you. But I don't want my units pretending to be smart and instead not being reliable or consistent, to the point they need to be micromanaged anyway. If I know they will hit my own structures if I let enemy units in my base, I can account for that. I can't account for unknown behaviour.

Uh. Not like that.

"Will this shot do more damage to my opponent's hp or more damage to my hp."

Or any unit can cause more damage to my own base thanks to Holkins.

I still don't get why people want their defensive structures to destroy their own buildings.