Friendly Fire on Splash Damage

Damage should be the same for friendlies and enemies, but I support the difference in damage between self destructing (due to an ordered attack) and being destroyed.

How many RTS games have produced a successful walking bomb? Hard mode: Not counting Starcraft.

This game series has a unique balance lever of letting any unit have a death explosion. It serves multiple functions on the field, causing unit clusters to crumple under fire, breaking synergy across similar range classes, denying wreckage drops/clearing wreckage fields, and encouraging battles at extreme ranges (melee to deal explosive damage, long range to avoid chain reaction). A vast number of units can be both enhanced and given vulnerabilities by equipping them with some form of death explosion.

Do we need an explicit walking bomb to get this point across? No way. Is it even possible to make a walking bomb that isn't completely broken in some way? For example, a unit cannon with walking bombs can deal extreme damage. That bomb may be completely useless on the field and only work as ammo for the cannon, defeating the purpose of having it walk in the first place.

Besides we already have a functional walking bomb. The Commander fits the role quite well.

Theory crafting is all very well and good, but 20 doxen have a combined DPS of 400. Which is the same DPS that a Holkins is able to supply. It also would kill the factory sooner than it could produce reinforcement units. And you're making the rather bad assumption that someone is going to use their entire army to attack

You actually haven't proven anything. You're arguing the same point of view - that the unit should fire when more damage is inflicted to the enemy than the enemy is capable of inflicting on you. You're factoring in "potential damage", which is pretty hard to determine, and you're assuming that there is a single event. I.e. the bait unit is with the massive army blob, and so artillery does as much damage to massive army blob as it does to your own defensive units.

You're not factoring in the idea that the opponent is a smart human, and so moves a bait unit into the defensive line to remove it due to friendly fire artillery, so that he can move the rest of his army in without casualties, or that the opponent doesn't land a cheap air unit on top of the artillery/target of choice, hence removing it, or that the opponent doesn't rush Shellers with scouts so that they destroy themselves, or otherwise land air units on top of an army so that the supporting Shellers suddenly totally annihilate friendly forces. (Which would clearly completely screw up the AI's neural network, because it would start getting anomalous data based on how units were used and positioned)

You're not thinking about the friendly fire damage as being on the last surviving unit of an attacking army, giving it a force multiplier.

You're not thinking about the logical effect of artillery being inaccurate - it won't hit units that are appropriate microed. Meaning with appropriate micro, you can quite happily lead artillery shots into targets you want to destroy.

And please, think about what you say. If there is a unit that static defences cannot deal with, the only solution is to chase it down or trap it using static defences. Artillery is inaccurate enough that you can quite happily run units inside the net all day. A hard defensive line remains a really bad way to defend your base.

I sort of wonder what Uber thinks about our thoughts in this thread. Could be good to have something else on the front page other this this topic that is sort of beating itself to death with no real confirmation from the big dogs.

400dps will kill the factory in 5 seconds, leaving no time for it to produce reinforcements anyway. They will then move onto the next structure, and kill that too. And then the next. At what point does it become better for your artillery to fire anyway? This lack of a concrete answer is why such a system is bad. Consistency and reliability is far more important than trying to be smart but getting it wrong. I don't want the game to lose me the game because it decided something that turned out to be wrong. If I let friendly fire be an issue it's my own fault.

I'm not arguing that at all. I'm arguing that artillery shouldn't try to work it out, because it can't - it either always fires, or never fires when any friendly unit is within the possible impact area, or has the ability to move into the possible impact area. That's a big area, and there will still be plenty of edge cases where you'll take friendly damage anyway. It can't work out whether it will do more damage to friendly or enemy units, because it doesn't know when the shell will land exactly, or if enemy or friendly units will move.

Out of those two options, the former is consistent, reliable and able to be planned for. If you want an option for the later, I'm all for that, but in no case should such an unreliable mechanic be the default.

Correct. And everyone who is against such an idea and saying "it should calculate" is thinking of the case where the friendly artillery overkills a unit that could have been dealt with by other means and causes more collateral damage.

Yes, you are. You're just saying that the tolerance should be extremely low.

Exactly. It can't work it out. So the AI can't work it out, can never work out when it should hold fire or not, can't plan properly, and ends up with a very obvious archilles heel.

One thread, found here, brings up the idea of combining bomb bots and land mines: https://forums.uberent.com/threads/bomb-bots-and-mines.56018/

...but you already posted in it. Still, if friendly fire and bomb-bots coincide, and variable formation spread isn't enough to keep them from blowing themselves to pieces, a land mine function would be a great way to give them a use until unit cannons are available within a match.

And those people are using an imaginary algorithm that magically knows whether it's a good time to fire or not.

So simply, when should artillery fire that won't result in inconsistent behavior under any circumstance?

Again, no, I'm not. I advocate that artillery should never not fire unless told to hold fire.

And hugging enemy units or structures to avoid artillery fire is just as, if not more, exploitable against the AI, since the value of units under a players control is inherently worth more than those of an AI (ie., saving your units is more of an advantage than killing theirs).

Exactly. So the tolerance is incredibly low.

Once again, you disregard heuristic approach to the argument, and ignore the fact that the AI is capable of controlling moving units.

No, I'm saying there should be no tolerance, as it shouldn't try to calculate anything. If it's in range, fire. Very different to any kind of tolerance

Please actually respond to the questions I asked, instead of claiming I'm disregarding everything. Or at least explain how what I'm disregarding affects the argument. How does the AI being capable of controlling moving units invalidate my post, and how do heuristics result in dependable behavior? And as per the previous post, when should artillery fire that won't result in inconsistent behavior under any circumstance? What are you actually proposing?

IIRC in TA the Big Bertha (Long range artillery) had this as the default setting.

Again showing that TA wasn't as perfect as people say it is.

Why? I think this was to prevent the berthas from firing at aircraft all the time. I don't see this as a disadvantage.
In PA the holkins cannot fire at aircraft anyway, so just let it fire all the time.
Also in TA the artillery had a natural minimum range due to massive projectile speed, I think we might need a minimum range in PA aswell.

We still did not have the chance to play with bomb bots, but I assume they work in a logical way. They detonate when the enemy is in range. Or they self destruct at the player order.

Both explosions should deal Friendly fire. Specially the latter 'cos the players has full control on self destruction of their units.

Apart that an unit that has been destroyed doesn't necessarily has to detonate (think about the Nuke Launcher), Bomb bots could keep a wider formation than other units, so that not to interfere with each other in case of detonation.

However, when the Bomb bot detonates, it should deliver Friendly fire. If the player wants to mix units with Bomb bots, that should be the risk. It sounds like a cheat otherwise.

I remind you all that the Nukes and the Commander deliver a ton of Friendly fire when they detonate. Why should we apply a double standard across different units?

If something like the Bomb bot doesn't deal nicely with Friendly fire, isn't the concept of Friendly fire to be wrong. It's how the Bomb bot has been designed that has flaws.

Friendly fire should be a fundamental requirement, and upon it we should design the units. Not the other way around.

That was useful to prevent giving away the fact that you had a Bertha until you could choose an appropriate target. Otherwise you lose the element of surprise when it fires at an insignificant unit.

Then again, you could always just switch it off when you placed the building.

Otherwise you get instances where you place one, forget about it, and then eventually remember that you built a gun a while ago that hasn't done jack since it was built.

But then again, you could say the same as where you could switch it on once you place it, but adding in micromanagement like that is really kinda silly when shooting at enemy's should be the default.

To be honest, every time I built one which fire setting I wish was the default changes . It was very dependent on how safe it was.

well it'd be ridiculous if they didn't harm friendlies but maybe we can nerf the damage to each other so that they at leas function correctly.

That sounds pretty much having to fix something that's not wrong, when the original flaw is in the assumptions that have been made before thinking about the unit.

Instead to fix Friendly fire, we should accept it as a fact, and only later designing intelligently units like a Bomb bot. Not designing a Bomb bot and then discovering that huh we would like to have Friendly fire... and now what? I dunno... I'm confused.



So, I'm afraid that the simple existence of a Bomb bot in this game will negate Friendly fire *by design*. We'll see.

Well it depends on how the bomb bot is designed.

You could make it immune to it's own weapon by making it more of a bomb pumped laser, making the unit non-volatile when killed.