This is something I've been thinking about for a while. Why do so many RTS games have allot of clicking, and micromanagement? It seems necessary to manage every little thing, and it's kind of annoying. A real commander(to use the much maligned argument from Real Life) would not have to deal with setting up a square array of farms, she could just set some other officer in charge of it, and give that officer a plot of ground to build farms on. Micromanagement takes away from the overall strategy. Note that I didn't say that it was a bad thing, because some games thrive on micromanagement, and do well. I'm not talking about that sort of game. Each unit, or group on units should have an AI that makes them behave intelligently when the player is elsewhere. Now I realize that this is nothing new: RTS games have had soldiers fire on nearby enemy units(without orders) since before I was born. I've noticed, however, that most games don't take it much farther than that. You still have to micromanage your economy in the middle of a battle. Now I wouldn't say that the economy needs to be removed: disrupting your opponents economy is an important part of game play, and when the opponent disrupts you economy, it obviously needs micromanaging, but this should be minimal: make the enemy leave the economy alone, gain military control over the resources that your economy needs, and the filthy capitalists in your army should move in and start harvesting those resources, and using them as you ordered, without really needing to be told to. Buildings that produce soldiers should be controlled centrally: a commander should be able to tell her infrastructure "I need x troops at y location, and they need to be upgraded in z way" or "continuously produce these types of troops, in these these ratios", and a subordinate should take care of it. Why isn't this sort of thing found in a lot of RTS games? Why is it that micromanagement the exception, rather than the rule?

Generally speaking, it's hard to make an game that does the micromanagement by himself, but still let's the player play the game. The fire nearest enemy routine is quite easy: you just have to check for enemies, and if they are near, shoot them. But if you want them to cast a spell, then you have a lot of problems: should they cast right away or wait for more enemies to come? should it cast on the enemy with most HP or the enemy that deals most of the damage? This kind of decisions are left to the player, and also because it would be really really bad if you wanted to cast a spell, but the AI had allready used all the mana.

Of course, this could all be worked out, but how hard would it be to do it?

Another problem is that, if your AI is too good, it will just play the game by itself, and you never have to do anything. A human could be a general better than the AI, but how can you make the AI run an army without getting lost by the interference of the human player? Say the human wants to rush the enemy, but the AI sees that you lack the anti-invisibility protections, how do you comunicate?

This is all possible, of course, and I personaly would think it would be really cool to see something like that... it's just a matter of how do we do it. Personaly, I would make a game that is scriptable out of the box: you actually build your army AI, as if it were part of the game. Well, this is one of my crazy game ideas I'll never make: it's an RTS so big, so full of resources, that is impossible to be played by a human player! And this is still easier to make that an AI that heavily interacts with the player.

just my 2 cents.

Reminds me of "Too Many Clicks! Unit-Based Interfaces Considered Harmful". I don't know why higher-level orders are so rare, but I agree it's an interesting area to explore. I think the RTS genre is slowly moving in this direction though: squad-based control in Dawn of War for example.

My personal theory is that designers look too much at what is possible within the game (e.g. each unit can be ordered to do various things, every setting can be changed, etc.) and too little at how the game can be played, or what intentions a player could have (ambushes and multiple front attacks, coordinated flanking movements, etc.). The current interfaces are just like bricks and mortar. Sure, you could build a house with just that, but how about prefab walls - since we're going to build those walls everytime anyway? They're very flexible, but at the same time they're so time-intensive to use that you're never able to effectively put this flexibility to use in your grande strategies.

On the other hand, when people heard about the squads in Command & Conquer: Tiberium Wars, many players feared they would loose control, and they weren't so eager to try something new. Not everyone likes to play on a higher level, and I get the urge to micromanage too when I see my troops getting into trouble. It's a bit of a troubled situation: in reality, commanders are either not around the front-line, or they trust their soldiers in doing their job well. In games, we are constantly aware of the low-level situation, and we don't trust our units to act properly. It's not only about writing a solid AI, it's also about raising the players trust in this AI. And some people just don't like to play this way. :)

There's something to say for both sides here, I think, but there's certainly room for improvement. Tedious construction orders that have you pressing the same shortkeys again and again and waiting for that dozer to arrive, just to shave off a few precious seconds, has little to do with a commanders job imho. A macro system, something along the lines of Total Annihilations command system but extended with recording orders and labeling such orders, for example, could be a nice improvement. And perhaps integrated into a customizable interface...

Just my thoughts on the subject. :)

There is an old RTS game called "Dark Reign". In it you could build several unit construction buildings and then choose one of them to be main building. the main building would produce the units, but any other building of that type would, if it was not producing anything (you could set them to produce a different unit type), contribute to the construction (thus reducing the time to produce the unit).

So this kind of thing has been done in RTS games before.

Dark Reign is awesome! ... although the AI is a but rudimentary. And therein lies the problem: while the AI for enemy players being a bit crap in certain situations isn't really a problem, if your AI 'helper' does stupid things that tends to annoy the player.

(Edit: In Dark Reign you can pick a 'primary facility' for each type of construction building, set an exit point – or a 'rally point' in DR2 – for newly constructed units to go to and queue new units for construction without having to select the building. You can still micromanage if you like, building different units at different buildings; if you don't, multiple facilities make construction faster. You still have to manually pick units to build, but not where or where to go to.)

The economy doesn't generally need micromanaging, I find: you just tell resource collectors to mine a particular resource and they do that until otherwise instructed. I very much liked the Supply Lines suggestion posted here recently.

For group-based AI, check out Ground Control: you can't control individual units, just the squads (which are 1-8 units depending on the class). For a tactical AI, the algorithms are only just becoming good enough to make them work, but yeah, I'd like to see an RTS where you can say 'squad 1, attack the Red player from the west flank' in some way.

Quote:

Original post by Bob Janova For a tactical AI, the algorithms are only just becoming good enough to make them work, but yeah, I'd like to see an RTS where you can say 'squad 1, attack the Red player from the west flank' in some way.

How about a sort of drawing board mode? Instead of clicking on destinations, you draw a route? The unit will pathfind their way to the start of this route and follow it - if possible - to the end. Perhaps combined with the map overview... a chart full of red movement arrows... :) Of course, Total Annihilation already allowed something like this with it's command chaining system...

Dark Reign lets you set waypointed paths on the minimap too ;).

taking away micromanagment would be interessting (for the programmer), but don´t think of micromanagment as an enemy which takes away the strategie part

i´ve been playing warcraft 3 for 3 years.

if a strategygame does only depend on strategy it will 100% get boring.
all strategy game that i know are hell boring because at some point you know everything ... you start of with the same shit, building building ... scouting ... make counters .. blablabla ... and in fight? stand by whatching everyone dying?


think of micromanagment as an additional point of making your game competitive

i just love the possibilitys of a fast game like wc3 ... surrounding units, blocking off ... it´s so fast that even the best are not perfect ... great fun
(if you didn´t know: 200-400actions per minute in a fight are normal for mid-skilled players)

Spring (spring.clan-sy.com), uses a group ai system to great extent that provides automation for the player. Just about anything can be automated (in theory), and there are plenty of group ais floating around for some specific mods which range from clustering your troops when moving, to converting metal extractors to moho metal extractors (in the TA based mods), and then some.

The system is a little limited due to issues with the engine and the way group ais are currently constructed, but generally it works quite well. What it doesn't do is take away from the strategy. It only provides certain limited automation mechanisms for gameplay which isn't generally interesting (common when rts's "go big").

To contrast War3 which is a very, very small rts compared to TA or the up comming SupCom, it generally doesn't need the same level of automation since it was designed to not need it. Unfortunately it wont be providing massive battles any time soon either, which is more where automation is more helpful.

Interestingly automation also goes hand in hand with player aiding AI as well. Sometimes this can add micromanagement as well as remove it. Like the Majesty example above where the units don't need to be controlled... This could be added to allow units to perform autonomically (builders for example), and all you do is specify where to build stuff and give your build orders priority;)

Cheers,

Lorenz

Hmm... Alot of the problem seems to be that players don't trust the higher level AI. This seems to be because the AI might not work well with the player's way of thinking, or the designer didn't think of the situation that the player is trying to use the AI in, or for any number of reason's, the AI is "jsut a little bit off". Obviously, the player needs to be able to control the AI as much as possible. The solution, from m programmer's perspective, is obvious.

Scripting

But not all of the players are good at scripting. Only people with some programming experience are going to be able to appreciate that part of a game. What are the different ways of allowing the player to customize AI without requiring the player to know or be skilled at scripting?

Outside of that, what are the things that make a player AI useful rather than annoying? I'm currently working on a game(It's still very theoretical) Where there will be a lot of complexity, and quite a potential for micromanagement, But I also want the player to be able to deal with combat the whole game, and leave minor things alone and let them take care of themselves. I'd like to know what you guys think about how to do this well, and how to do it badly.