Planetary Annihilation's Economy System

auricomus I think we disagree on the meanings of some important words. It is pointless trying to argue like this. But it looks like I am going to try anyway.

Upon rereading I think I make some independent arguments in the second and third last paragraphs. You should reread and point out the bits where I am just telling you without justification. If you disagree with any justification then feel free to point it out and argue your point. Then finally how are your points not invalid?

This was probably a poor way to put it. To me the point of your UI is to let you make your units do what you want them to do. If it is difficult for someone to make their units do what they want them to do then the UI is hard to use. You are advocating that the UI be hard to use by this definition. It doesn't actually matter that I said "hard to use" because the point should come across directly from the context of the quote.

By "weak UI" I mean one which is hard to use by that previous definition. I felt "weak UI" was a reasonable term to use because the opposite is often referred to as a "powerful" or "strong" UI.

Below is a quote of you saying that a weak UI will make games smaller.

Ergh, can we not Quote Snipe? It's such a vicious and unfriendly way to have a conversation. Not to mention it's really ugly format-wise.

Ironically the birth of the RTS with dune 2 had stream out resources. While they were delivered in discreet bundles from harvesters the amount those harvesters contained could also be variable.

Its just warcraft spoiled so many people with its entirely discreet resource system. Personally i've never understood how this is a hard concept to get, although i can see it being hard to manage without knowledge of how much things go in and out. I ended up putting the resource monitoring UI mod for my sup com, forget which one but it lists the top 5 causes for resource use as well as mex status.

I did always enjoy customizing my UI because it made the game feel like it was mine. Someone else can sit down and just be in awe of the information and capabilities i am given by these extra tools. anyway i'm not sure where i'm going with this so i'll stop.

auricomus, it should have been clear that I was using the phrase to informally mean mechanical interface.

Then what you assert is a flat out lie.

Quote Snipe? I am not aware that it is considered a bad thing. To me quotes clarify the ideas you are referring to. People aren't held to what they say but I think it is useful to assume they think what they said unless they later said otherwise. Quotes are very useful when there is an issue with definitions (because you can focus on some misunderstanding). It is hard to have a discussion while such misunderstanding exists.

In this case auricomus is making a lot of statements which are contradictory from my point of view. Getting to the bottom of why I think they are contradictory sounds like a good way to reach some agreed upon definitions.

Googlefrog,

I said that you cannot assume that the UI will be hard to use because no-one has seen it yet.

I also NEVER said anything about a UI in the quote you took supposedly showing me saying ; "weak UI makes games smaller"

And i will reiterate;
a UI is not supposed to do anything for the player except allow the player to control the game. Ergo, an AI managing power, is NOT a UI feature, because that has nothing to do with the user.

I'm afraid that the meaning of UI has been corrupted beyond measure to some people here >.>

You seem to think that words have strictly defined correct meanings which can be corrupted. If the word "UI" looks like it has been corrupted then we're probably in a situation where we need to abandon that word and talk about more specific things.

I have explained what I meant by certain words. Read them and update your arguments to reflect this. You can use different words for what I mean as long as you explain them. I don't think it matters what I call UI as long as I explain the surrounding concepts sufficiently clearly.

I'll reiterate the most important part as a series of assumptions that I make. Point out the wrong ones.

• Here you say that extra-automation beyond the basic UI is not required because the other player is in the same situation.
• This situation results in both players being unable to handle more than 2 battles.
• I generalise this to "people won't be able to handle a large amount of anything".
• Games in which nobody has a large amount of anything are small.
• It is implied that a more powerful UI means that players will be able to manage more things.
• So a weaker UI (by the definition I just explained, a basic UI, no automation) will make the games smaller than a stronger UI.

On another point, what do you mean by "control the game"?

Googlefrog,

Words do have a strictly defined meaning, this is what is called a language, and allows us to understand each other.

UI means User Interface, a way for the user to interface with the game. Interface meaning a point where two(or more) systems interact.
Ergo, the UI is the 'point'/system where the player interacts (controls) the game.

Having a system that manages power or whatever without the users intervention is not a UI feature. The ONLY part of the UI that this relates to, is the system IN the UI giving the player the ability to program the priorities. The priority system itself is NOT a UI feature.

Moving on to my paragraph which you have corrupted/misinterpreted almost beyond recognition;

1) I highlight , that as a human player, they can only have one point of focus, we only have one brain sadly, one set of optical sensors, one set of audio etc.
This means, that as a human, we can only focus on one thing at a time (this goes for anything) we don't have multiple cores like a CPU. Well , we kinda do , but consciously we don't.
This means, that for a player to engage another on multiple fronts requires shifting of focus between these 'fronts'.
This requires skill and can limit a players effectiveness if they cannot switch focus as well as another player.

2) This situation results in certain players being better than others at focusing on different aspects of the game rapidly and effectively.

3) Generalized properly : "people who are better at switching their focus between situations and reacting effectively are better at the game than people who can't"

4) Games in which nobody has a large amount of anything are games in which nobody has a large amount of anything . I don;t see how this has anything to do with the size of the map.. if that's what you are referring to.

5) A UI that allows players to switch focus faster does allow them to manage more things.

6) A UI which limits a players effectiveness is clearly not as good as a UI which allows the player to be as effective as they can be.

There I've addressed all of your points and would like to return to the last one.
Yes a good UI is key (i never said it wasn't).
We don't know what Planetary Annihilation UI is like yet.
An AI controlled energy management system is not a UI feature. So please , stop calling it that.

Also, an AI controlled energy management system means that weaker players don't have to worry about power as much when playing . This is silly , because energy management is a key part of the game, and should not be handed off to an AI just because some players think wont be able to cope. If they cant cope then they need to get better , OR play in such a way where they can cope (turtling, rushing, HOTKEYS etc)

"Control the game" , i would have thought this is clear but ill try and clarify further:
A player controls the game by influencing/directing the behavior of the game program producing a result.
Ergo, having an AI control your power distribution is not the player doing so, and therefore is not part of the User Interface.

I probably repeated alot there, but it was only in response to your points ^.^

Perhaps, to reference Neutrino's earlier comment, you disagree that "pre-programming responses to specific situations" should not be allowed. IE. players shouldn't be allowed to give orders in advance.

Because, the way I see it, a priority system is simply a way for you to tell units: "Do X when Y happens". In this case, a player created "Decreasing Priority: Mex, Radar, Factories, MM" order is essentially telling all your MMs "turn off if my energy income < expense" and telling factories "turn off if MMs are all off and energy income < expense" and so on up the chain.

It is pre-programming responses so you don't have to do them later. It is NOT an AI because it does not have a model of the game world or consider alternative strategies. It is just a script that does many hundreds of actions per second based on whether a number is above some threshold.


Obviously, I disagree with that and think orders SHOULD be allowed to be given in advance and that scripts are a useful tool to implement player decisions. Essentially the player has decided on some strategy (prioritize mex over radar over factories over MMs, say) and the script lets him implement it with only a few clicks during downtime shortly after game start.
Instead of using 30-40 clicks/button presses every battle while never actually changing his overall strategy of which is better to keep on. And if he does change, making it cost only a few more clicks instead of requiring 30-40 different clicks/buttons the next time it comes around.


Now, what I consider an AI is something like this:
AI_Harass; you give it a number of light fast raiders who then split up to cooperatively cover unexplored areas to find the enemy. Undefended enemies are sniped if no resistance is nearby, defenses and force concentrations are remembered and avoided, etc. etc.

Now, THAT is an AI subroutine. It makes important decisions (do I kill that mex and sacrifice the raider?), holds simplified models of enemy positions, and automatically performs actions in response to guessed-at information (enemy patrols seen chasing the raider but are now out of sight will be assumed to be looking for the raider = decide to retreat)



EDIT: I might like to say that opinions on varying levels of automation/UI power/whatever you call it have been hashed out pretty well earlier in the thread as well as many other threads.

Maybe so but now I have a much better understanding of your point of view. We're getting somewhere.

Most words have quite a fuzzy meaning. This works for us most of the time but it breaks down when two people try to go into specifics over a word. But I think we understand what we each mean by UI so this is just an aside.

Now I think I need to explain the scope of my idea of an economy management system. In any case I think I still need to describe it precisely.

My idealised concept of UI in RTS games is that the player can only affect the game by giving orders to units. The player should be allowed to give complex orders to many units without breaking the balance of the game. The UI is simply a bridge between the clicks and keypresses that a player makes and the orders that are given to units.

So in short the UI is a machine that takes the keyboard and mouse as inputs, does some calculation and spits out a list of orders to give to units. It does this every game step (or whatever you call it). The UI has states independent of the underling game. For example the set of units which are selected does not directly affect the gamestate. It is just another piece of data the UI uses to determine what to do if you right click on the map. A player doesn't have to select units for the UI to give them orders. The UI could in theory give arbitrary orders and you could build a whole AI into it and ignore the player completely. It could do this without affecting the underlying mechanics of the units in the game, this is still in some sense fair.

(I will answer the question "what stops players from letting the AI play for them?" when AIs are good enough for this to be an issue)

So, given that point of view (which you probably disagree with) a priorities system is just some extra machinery within the UI which lets the player configure which buildings should be manually turned off by the UI (manual, as in not forced off due to game mechanics) first and which should be turned off last.


I don't expect you to agree with the central idea of my post; That the UI is a machine that turns arbitrary player input into arbitrary orders. But hopefully my point of view makes a bit more sense. Changing people's minds through forums is very hard. I have found it is much more effective to make a game with these ideas and then have people try them out.

Edit: I forgot to mention that the UI can access the data that the player can.

Umm... googlefrog do you mean just... hotkeys? In TA or SupCom it is very easy to hotkey every one of your Metal Extractors (or any other discrete selection of units/buildings) and turn them on and off with but 2 button presses. Pretty efficient. With my very simple keyboard setup I can hotkey different hotkey groups if I needed to.

What more are people asking for here exactly?

Tell your mexes/radar/factories/MMs when to turn off before the situation occurs. Conditional on some simple criteria.

That's Predictive AI. Not player input and certainly not a User Interface issue.

It doesn't predict anything though. It is exactly as I describe, I want to tell my MMs "do X when Y happens". It is an input because it is a strategy I have decided to execute and not wish to spend the time on later.

It's a strategy you want implemented for you. Not one under your direct, second-to-second control. It's not player input, it's a background script that is being asked for, and to be in the game as "standard".

I have no problem with player created scripting to increase input complexity, but not reactive scripting. That is making the game play itself for you.

---

You edited your post after I was done typing and had already posted, by the way. My previous statement of "Predictive AI" would now read "Reactive AI". I'm not editing it since it just breeds confusion.

I forgot to mention that the UI can access the data that the player can.

Yeah, we were going a bit too fast there, so I deliberately waited a bit this time.

I think our concept of "play" is different. I certainly do not consider "direct second-to-second control" playing the game or even necessary to play the game, it is just a pattern of mouse-clicks and button presses that I do to make my thingies do what I want.

Assessing information and making decisions are what I consider playing the game. Execution factors very little. Turning off a few MMs first, checking my resource bar, turning off factories, checking again, etc. counts only as tedious busywork (and don't say it is a skill of counting numbers quickly as MMs 'consume' more E than their listed rate due to the M they produce for factory usage which gets nearly unpredictable in an energy stall). It is a set of clicks and buttons I use to implement a strategy (a certain priority order I have decided on earlier) and the less of that I have to do, the better.

When and How I make the clicks and buttons to make the thingies implement my strategy is strictly secondary to deciding what the strategy should be and if it needs changing due to new information.
EDIT: tactics also falls under here. Deciding to put your rocket launchers on the hill is a decision, lining them up neatly is not.

" I want to tell my MMs "do X when Y happens". "

Exactly, you INPUT the command(Via the UI) for your base to run itself, and then you forget about it .
Running your base is a key part of the game, and you want to tell your base(basically an AI) how it should react and when, and then forget about it.
This eliminates potential for you to make mistakes when under attack, change up your response depending on specific circumstances, and most of all , reduces the skill ceiling for the game.
If the huge portion of the game 'Resource management' can be handed off to a bot, then that's a whole part of the game where players cannot exceed each other in.

This is a GAME , not a factory, there needs to be room for mistakes and room for players to excel and so; as much as possible should be left down to the player. Not the reverse.

So jseah, would you be happy with "Button A" turns off half my Metal Extractors, "Button B" halts all Air Factory Production, "Button C" turns off all Radar except on the planet I am currently focused on... etc.

Or do you want all of those things to happen automatically without buttons?

One is hotkey scripting, and is fine.
The Other is automation and is a crutch. The game is Automatically Reacting for you. The game is playing itself.

You do realise that AI in the RTS sense is just complex reactive scripts, right? You want your economy to be entirely run by complex reactive scripts... or "AI".

---

You further realise that these scripts would be sharable, right? So this game becomes "Who's got the best reactive script in his game folder" and not ""Who's got the best reactive Mind, Strategy and Tactics as a Commander"

I could just download a script to handle my Economy and never have to learn how to manage a core aspect of the TA and SupCom subgenre of RTS.

I know this is going to sound old, but this part really shouts STARCRAFT to me. Eliminating mistaken orders or button hunting (which hotkey was it again?) IS UI to me. You are never going to get rid of the fact that doing exactly what you describe you want is impossible, but the UI should try to hit as close to it as it can.

I want my UI to understand what I want my thingies to do. Having an interface limited in time... is well, a limited interface. Same with conditional orders or more complex if-then commands.

Besides, resource management is not just a priority system. Weighing risks and rewards of capturing a wreck field or a mex point is a key part of the overall strategy.

EDIT: a genuine mistake is a bad decision, along the lines of "oops, I should have kept radar instead of mexes powered".
Not something like "oops, should have pressed T instead of R".

Automatically without buttons, but there is a button that changes the details by which it happens. Yes, I am perfectly fine with reactive scripts. Hell, I occasionally write widgets for ZK. >.>
Some of them are stupidly complex and one even tried to do behaviour prediction (anti-bait for AA towers). It did not go well btw.

The way I see reactive scripts is as follows:
Each reactive script is a complex strategy. Deciding to use/activate a script is a decision to use that strategy. Knowing what the script does and all its details is key to knowing just what decision you made and how it fits into the overall plan.

This means that scripts are just another button you use to communicate "what you want to happen" to the game. Obviously there cannot be a button for every detail of every strategy as there are far too many of them; thus the scripts you choose (aka. the buttons you have) are tactics and strategies you anticipate you will need often.

And then when you need them and decide say "hey, it would be good if my skirmishers would kite the oncoming assault force" and you have a "move away from target when at less than max range" script AND you think what the script does would be useful, then you hit the button that says do X. Or in this case, you give the Kite command / push the hotkey that runs that script.

If you know what the script does, then you know its failure modes, when your strategy no longer applies (eg. surrounded or backed against a cliff), and then you don't use the Kite command or cancel it when you find that out. You do something else, like lining them up to deal maximum damage before they die (and lining up could be a script too!).

Thus the kind of Complex Strategies buttons you want to have are the kind you think you will use often. You don't want a script that does TOO complicated things (eg. automatically build mexes everywhere! and defend them! and etc...) as the situation for using it will come up rarely and it's button slot could be given to a script you will use more often and it will be of more use. Since when do you ever think "hey, 'really complicated multi-factor strategy' would be useful here?".
No, simple low-level stuff with simple and easy to understand conditions are where scripts are at.


The reason why I think "playing the game for you" isn't going to happen is because frankly, AI algorithms suck at playing games. Despite much ado about Sorian's AI on FA, I stomped it on my first try in an equal match; and despite the expectation that it will be better in PA, I fully expect to stomp it flat once I learn the game.