Planetary Annihilation's Economy System

I wish I could have figured out how to say it like this.

Did you ever try out some of the various ai-programming games out there? You might find them pretty enjoyable. You basically predefine everything and just watch your ai fight it out. That's definitely a fun thing to do, but it is quite different from what an RTS is to most people, including myself.

Starcraft 2 is so different from what you are asking for. It is basically a different genre.

I guess we all have our own idea on how the game should feel, so we all champion slightly (or radically) different 'feels' for the game. All to the better - I'm happy with anything between TA & Zero-K too!

I guess we'll leap the hurdle of what to do with powerful UI mods once we get there

Back on Neutrino's large-ish post a few pages back, the way I interpreted it was;

• If I make an attack with the intention of distracting the enemy human, the human's UI/game should not automate away the effect.
  
  If I make an attack with the intention of blowing up enemy tanks, then automation is fair game.

Keep in mind that multi-player is about beating another person, not beating a bunch of tanks. If all you care about is commanding your tanks and factories better than other tanks and factories are being commanded, single-player will probably keep you satisfied.

That depends - if Eco recovery involves decisions with multiple optimal choices then not automating in the sense I have to take an action for it to occour is naturally a must - I need to make that decision for the recovery to progress. however, once I have made that decision I would want the process of executing it to be as simple & accessible as possible.

If there is a single optimal way of recovering Eco, and its just a matter of me performing it, then I would want it automated. If there is a logical way of prioritizing what to power in these situations then I would want that automated.

I agree players should be forced to take action, when there is a decision between multiple valid options to be made. If its just a matter of performing a housekeeping action - turning things on or off, the same way every time - then I would prefer not to have to do that manually.

I have played a few and enjoy them, but I do like the pace and adrenaline of real time. I dont want to give the impression of being a bad player who dislikes responsibility of control, I drift in and out the ZK top ten as [GBC]1v0ry_k1ng and play an aggressive style. My biggest weakness of a player is that I micro quite poorly, and sloppy unit control often kills me early game in 1v1 against skilled opponents. My strength is being able to control large fronts (such as when other players drop from big team games), make the right unit compositions and commit them to the right locations. This macro style works very well with automation, as I can multi-task several different areas or types of forces (land units and bombers for example). It translates very poorly to most commercial RTS, which are focused around micro and performing lots of 'housekeeping' micro.

I should clarify that when I say make units programmable, I mean in the sense of behavioral toggles with fixed effects (skirmish/dont skirmish, hold position/manuver/roam, retreat on x% HP/dont, repeat action infinitely etc) rather than actual scripting.

I'm glad neutrino has common sense, I'd like to actually play the game rather than watch the AI do it's thing.

If the original post is still accurate, sounds interesting.

Basically what the automation supporters think is:
I should have my ways to prevent the opponent player distract me if I predicted that.
Actually any sustaining command would decrease the effect of distracting, like order queuing, if you have two bases and you queue a lot in one of them, when the opponent attack the other one of them, you would have less loss, I don't understand why automation is essentially different than that in this aspect.

However, I can attack your automation, and force you to respond to that.

That's me attacking the human, not me attacking the units.


Having an economy that automagically diverts resources to maintain production, or intel, means that I'm fighting the computer, not fighting the player. This is something completely different to having an infinite build queue, or telling engineers to assist factories. Queues and assisting are mindless actions; there's no calculations or decisions being made. I believe the interface should be made to minimise mindless actions.

Actions that require a decision; those should always be left to the player to do.

Since the RTS games are complex, making them into ai-programming games are impossible so that idea has no meaning, and no matter how much the player predefine things, human management is still far more important than automation.
If a player can implement some automations (with all knowledge of them, that's where googlefrog draw the line), he must be better than his automations except the monotonous tasks that require nothing but fast reactions.
If it is a platformer, then you should do everything in real time and any automation is bad, but we are talking about a strategic game.

As long as the player has knowledge of his automations, he's making decisions by toggling and adjusting it.
Say if a hunter set a trap to capture an animal, would it be the trap capture the animal or the hunter? I think it is both.

Queuing several move orders is less intuitive and transparent than implement them in real time, so should we remove the order queuing? Where the line draws?
I think if we can't make a useful automation tool super transparent, at least we should have it in UI. Non-intuitive stuffs do raise the learning curve in some ways but that's better than raising the reflex training curve by the powerless of UI.
And I think that the feeling of lack the ways to control your units is worse for the game experience than UI overlays that can distract you from the game world, usually a RTS player would already be constantly distracted by things like resource bar or way points anyway, some togglable automations won't make a big different.

I wrote this post before realizing that you must have seen this whole argument many times over.

So I shall revise my position and say "wait and see". I only ask that you allow UI mods (the sort that runs client side but servers can refuse to allow) that can do this.

You still hurt unit production. A widget for economy priority won't generate resources out of thin air.

Shoot enough power generators, factories slow to crawl.

And he'll have to rebuild the energy anyway.

Hmm? I don't see why a player would *not* pay attention if something in his base was getting shot? If there are units in your base shooting powerplants, your attention will likely be there.

Unless of course you are in his base shooting his powerplants.

That said, here's how I imagine the scenario working out in practice:

1. I produce excess energy because metal fluctuates; similarly I have excess buildpower
2. I build 1-2 metal makers to sink that excess energy (but not relying on them because MMs eat lots of energy and are inefficient)
3. I have metal makers on a hotkey because metal makers have priority over production

Now, an enemy decides to charge my defenses (which use energy to shoot). Or I win the war over a wreck field and twenty engineers worth of reclaim pour in. Or a lucky bomber strike takes out a fusion powerplant.

What happens next is me seeing my energy bar going down, pressing "1" to call up my metal makers and click off. (or if the UI allows, click enough of them off so that the energy stops going down).

I will build enough energy such that I have an excess buffer to handle most freak accidents or volatile usage spikes (eg. energy defenses powering up), and simply soak it up with overdrive or other economy boosting 'optionals' that I keep on a hotkey to power off as needed.

The micro needed to go through your base and pause factories requires zooming your screen, considering which factory to turn off (shipyard, land or air?) and/or which energy defense will not get power... that is far far too much when you are handling anything larger than a ZK 1v1. I will eat the inefficiency and just invest in excess eco to prevent ever having to do that.

Because if there is an enemy knocking on my door, blowing up my stuff, I want to turn off all my MMs and run a giant energy excess. Just so I don't have to worry about it and can move my armies in response to him instead of clicking my base.


Annoyingly, this is exactly what I did when playing Forged Alliance. Only it was even MORE of a pain there because getting MM usage to exactly match energy production basically never happens. And that the UI did not allow you to turn them on/off one at a time (you did them in groups). What this meant was that I was pressing '1' every 10 seconds to turn them on/off as my energy bar goes up/down. (and if I think I can't spare the time, I turn them off and lose the excess)


You can think of energy priority as going in tiers:
Optional inefficient economy boosters -> Factory production -> Cloaker -> Radar -> Shield -> Defenses

If automation exists at the top (eg. a metal maker widget), you will see people losing a bit of efficiency and just building an excess energy buffer so they won't have to get pressured.

If no economy priority automation exists at all, people will still do the same thing with a hotkey to take away the micro load. The value of attention and micro is high enough that I'm willing to eat some inefficiency to reduce the amount I need.

Have you seen what a soft energy stall looks like? Say +5/-15. In a ZK 1v1, getting into that situation, say by losing a row of windgens to a raid from a turret blindspot, even if you can run your mexes and radar with 5E income still results in "GG, E stall".

Hell, there still exists a tactic where you run a glaive/scorcher past an enemy wall of solars to poke them all closed to cause an estall. (although not one that results in GG) Running a glaive through a windfarm will cause an economy crash of game winning levels.

It is neither weak nor ineffective. (and you don't want to know what 'Hard' estall is like)

Yeah, at least give us the possibility to create client side unit AI widgets (you can disallow them in the offical server or the unmodified game), we can prove more automations is good for game experience by making them popular.

I hope you understand that popularity isn't a measure of goodness.

At several points in human history, slavery has been pretty popular.

And that's why the negative attitude towards automation can be so popular among the RTS fans and developers at the moment.
And I was not talking about goodness, I was talking about experience.

Hey there, I'm a Zero-K dev and was asked by our (shared) players to comment.

I'm fully sympathetic to Neutrino here. Zero-K has a very powerful interface, and full lua scripting so you can write an AI to play the game for you if you like. So we have to design the game to be complex, and the decisions non-trivial: IE, they simply can't be automated.

Sometimes this railroads our design. Neutrino is right: Energy drain like discussed here basically had to be removed. Turrets draining energy to fire, metal extractors (mexes) draining energy to function, etc are trivially circumvented by a reserve or priority system.

So instead, we have an energy 'grid'. Energy has to be adjacent (or linked in a chain) to metal extractors and turrets to provide power. If you break the chain, the structure can fail.

The total energy linked to a metal extractor determines its max Overdrive amount, which consumes all your excess energy to provide more metal with diminishing returns. This replaces metal makers, keeping metal extractors and thus territory important even when you have huge amounts of energy.

There are dozens of other design solutions to this but you can see how this actually makes the game more complex when your energy also has a spatial component. It is vitally important to emphasize that Zero-K is not any simpler because of it's automation. It is actually needlessly difficult in many places and I think this hurts our popularity, but all the devs are high level Spring players who have been playing for years, so sometimes we design for ourselves and fail to see this.

A good example is many of our projectiles move so slow that they can be dodged (like Quake, imagine this with 100 units!). So we have an AI (on 'attack move/fight') that jinks to dodge them automatically (less competently than a player). This is so needlessly complex it is sometimes ridiculous.

Google Frogs initial post on this topic covers some of these ideas a bit:
viewtopic.php?f=61&t=34222
(The core Zero-K developers are myself, KingRaptor, GoogleFrog and Licho).

I think it is always good to look at the decisions your players are making and ensure that they are non-trivial. Even SC2 tried to do this with their macro mechanics: not having multi-building select is just busywork, so they tried to replace it with things that are NOT busywork (and sometimes succeed, like Chrono Boost, or failed, like Inject Larvae).

What we have in Zero-K is actually a priority system for construction. If you want to finish a super weapon, set it to high. If you want your turret or expansion to not be slowed by your factory, set your factory on low. These are non-trivial decisions that give the player more power and I think that PA should include this, over the busywork of cancelling and re-starting all your factory queues every time you start a large project.

Whether they want priority for power-draining buildings is a different issue. Forcing an enemy to e-stall then walking over his non-functioning defences feels totally great and if you circumvent this you are forced to rethink or remove the whole mechanic because it has now been rendered obsolete.

P.S. Glad to see your economy has it's drain rate be flat based on the builder, not variable depending on the buildee. It's much better this way!

I feel that the complexity would always be there if the you want to promote the importance of simulation in unit interactions by things like slow projectiles no matter the tasks like jinking are automated or manual. Micro heavy spring games like XTA are not really simpler than 0K in this aspect.

I didn't want to get into this because it's fairly off topic, but yes. We endeavoured from the outset to create a game which takes advantage of the physics system in Spring and uses physical properties for all of it's unit counters, rather than things like reduced damage armour categories (which we avoid). So the fact these small, light units can dodge slow shots means they are an important counter to the units that fire them. This scales badly with purely human micro, but if the AI can do it at least competently you can swarm the big tanks to bring them down even into late game. This AI therefore becomes a part of the physical simulation that underlies the unit interactions (Units even circle-strafe!).

It's no Dwarf Fortress but it's that direction.

If PA wants to make a game "intuitive, direct and transparent", I certainly think they can learn from some of the interface innovations that have come from a result of years of an open source format where anyone can write AI's or interface additions. But ZK and PA are different games with different design objectives.

I put forth practical game examples to inform you what importance energy plays; what energy numbers a typical build fleet really requires; what exponential power percentages have; how devastating overspending on more power infrastructure can be; how much the player has to keep track of in any typical game. You haven't provided any hard experiences to show me why all this is wrong. I don't even know what kind of RTS player you are.

By now I'll just address this one notion, which you put forth early on:

This is rubbish. An energy deficiency caused by surplus metal and build power can't translate to a loss of existing energy production. The player is at worst still producing at their maximum energy rate, which is nothing equated to automatic lost games, although, underproducing power plants in general will lose you games regardless. It's so dubious and nonsensical the way people think they can just rid of newbie traps. I have yet to know an RTS game which doesn't have subtle mechanics that punish players competitively for not fully grasping them.

Actually, the concern of keeping radars and defences online is a very strong argument for unit resource priority states.

I was originally concerned about the potential harm of a particular unit behaviour change. But it would be lying to not admit what you describe hasn't already existed in RTS games in the form of unit stances(hold/return/fire at will; stand/maneuver/chase).

I asked this earlier, but if your feeling is that functions not visible on the map are bad for gameplay, doesn't that logically make unit move and fire stances bad?

Damn, this conversation exploded since yesterday! To sum things up,

I completely reject the proposition that every single level of damage has to be achieved through energy. A priority system hardens a player against minor stalls, but it can not save a player from a major stall. It is impossible. If the resources don't exist, certain things simply can not be done.

Various talking points are below:

No. No nonononono. ZK's priority system applied mostly to production, and you know what? There are abolutely zero game ending issues with prorating metal. Every unit gets its fair share, and none is the wiser. Players already handle priority by adding workers to their favorite projects. It is simple, transparent, and nearly impossible to screw up.

ZK's production priority doesn't add anything new or meaningful to the game. Not only does it not give any benefit to the player, it HURTS him by encouraging bad gaming practices (too many factories, too many fabbers, etc.). It will not be missed.

There are so many ways to hurt a player without energy that I see this as a non issue. Seriously. Just take a pen and paper, and write down every single thing that gets messed up(or CAN be messed up) by an energy stall. The list is going to be HUGE, over twice as long as the list for a metal stall. There will be a big enough issue with having an energy stall NOT deal incredible damage.

The #1 damage during an energy stall is the loss of ubergun and Commander abilities. If you rely on keeping your Commander cloaked, or in using his Ubergun to stay alive, he is now completely open to attack. That's a game ending threat out of the gate. Two seconds of lost power, and it's enough to lose the game.

An attack that is repelled is not an attack that failed to do damage. While you may not like Starcraft 2, it shows this simple truth with its own workers. When players stop their harvesting operations to stay alive(the PA equivalent of shutting down extractors to power guns), they are taking damage. Every single second of lost income is causing the defender to lose his ground, but he is given no choice. He has to disable harvesting to stay alive.

This hazard will be increased as the number of energy using systems increase. Unpredictable energy users don't necessarily demand extra generators, but they DO demand rapid changes in priority with every situation.

That's both wrong AND a lie. When a player suffers an attack or economic damage, his priorities change. The whole point of a good attack is to FORCE a response.

Economic damage will happen regardless of priority, and I disagree with neutrino's assessment. If a raid proves to be consistently ineffective, the solution is to improve the raider. Damage is damage no matter what, and there are plenty of ways to deal lasting damage. Blow up the wreckage (death explosions are good), reclaim/steal it for your own use(fast units are good), or target a player's more critical infrastructure (such as nuke/ordnance defense). How many of those losses will it take to lose the game? While the question is up in the air, one thing is certain. Energy is not necessary to deal lasting damage.

The simple fact of Total Annihilation is that its economy was more resilient due to priority. It may have not have been the most ideal system, but weapons and the "mana" based d-gun both had a lower "storage" priority than all the "prorate" items. At the very least, low energy caused the bertha and annihilator to shut off before everything else. Those were exactly the first things you wanted to turn off in an emergency. When a Supcom energy economy failed, you were 100% screwed. Not only did the UI make it exceedingly difficult to recover, but the game itself crushes a player for ANY sort of shortage.

A little resilience in the player's resources is not a bad thing. It is more fun to take some damage and champion on, than to stick the player up the butt at every first opportunity.