Defining micro

I did it all the time with big armies against artillery in SupCom and FA. Will the same tactic be advantageous in PA? I don't know. It might indeed be necessary in competetive matches.

That sounds like a really advanced AI. The automization I support is much simpler but it would be cool to see an AI of that complexity that can do what you describe.

People usually use the group movements for kiting, retreat, flanking etc, automate any one of them will make the manual group movements less needed.
Tactical decisions usually don't happen frequently in a rts battle, it's just that the players have to use many move commands to implement a tactic.

It would be cool if you could select a big group of units and set them to e.g. "kite" or "chase" behaviour too.

Although on the other hand there's an argument for behaviour being firmly tied to unit role, so that the raiders and artillery in a mixed group always act according to their strengths.

Retreat = move order in the right direction.
Flanking = move order in the right direction.

What's there to automate, really? I don't see any edge to be gained here.

Definite no to the latter part, if you tell a scout to kite it should try to do that, if you tell an artillery unit to chase it should chase.
Units should NEVER disobey you, no matter how foolish your move might seem.

With big armies against artillery? A big army wont react fast enough to evade anything for the most part.
At least not in the scenarios I know. Avoiding arty is something that is mostly important with the acu alone vs arty. Or with other single, valuable units.

Advanced: yes, interesting from a technical standpoint: yes, good for gameplay: no.
Letting an AI play the game for you is just not good for gameplay.

For me it at least it feels like they happen every few seconds, never counted them, though. Mostly they are caused by changed behaviour of enemy units.

I remember tight UEF vs UEF games on Winter duel. You would have to constantly dodge artillery both with the ACU and your own artillery. 10-20 artillery pieces and 30-40 tanks.
Now just because SupCom had half second delay between order and execution plus that only like 3-6 units at a time would recieve the order doesn't mean that PA have to be the same.
And if you gave the order prematurely you might actually even be able to dodge with a large army in SupCom.
Now imagine if all units moved at the same time as you gave the order. Alot of artillery fire can be avoided then.

How do you know? How many games have you played where you could control the units to a larger extent but just said "No, I can't do it much better then the AI anyway"?

I have to agree with Cola_Colin on this one. You are constantly trying to gain a concave on the enemy, focus fire and keep your units on a line. Offcourse there are some pretty obvious situations when your micro won't do much difference but that is propably because you have already outmanouvered the opponent or countered his units in other ways. Some of theese could be handled by automization or AI but when the opponent is aware of the AI he also know how not get in a bad position and will focus his attention on that.

Let's not also forget, as I brought up before, features like attack-move, patrol, and factory waypoint commands are already forms of automation. Not to mention any kind of pathfinding. So, I find it hard to imagine that anyone is a complete anti-automation purist.

The question is to objectively define at which point automation becomes excessive.

Like you say: since the units in SupCom reacted rather slow it wasnt very successful for the most part. Winter Duel sure is a quite extreme example of t1 arty spam, I admit that. I am more thinking about all the games on theta passage I played, arty wasnt that common on it.
Sure in SupCom2 vs Cybran Brackman artillery it was actually quite common to at least try to automicro them. That wasnt very much fun, I admit that. But that was a problem of a single unit, the Brackman.

I cant name any. But a game which has an AI that plays the game for me obviously isnt even a real game.

@luckywaldo7:
yes thats just about right.
For me automation becomes too excessive at the point where it interferes with the move-commands I give all the time while playing. Units that try to automatically react to changed situations are not what I want. Kiting is a form of that: Units would automatically do different things, depending on what the target-unit does. Evading shots by moving left-right is like that, too. Any kind of reaction to the opponent are a task of the player himself, not of some AI. And yes, patroul is kind of similar to this, units patroul between A and B and as soon as an opponent appears they stop moving and attack. But this behavior is pretty stupid and mostly useless, so it has very few uses -mostly for scouting units to observe a bigger area and for air units to stay in the air- and isnt really used for its automation-function.

So make it optional?

That's what I keep trying to say. ZK's implementation is entirely optional. If you have more confidence in your own ability to micro units (to the degree that it would be allowed in PA), then do so. But for those of us who aren't tentacled APM monstrosities with a million prehensile limbs, the ability to toggle LOW LEVEL unit AI will be a great boon when we're across the solar system, coordinating an attack on another planet.

I doubt that there is any AI that can come close to human capabilities with a physics based RTS game like PA is going to be. So even if they developers tried to make to the best AI for PA they could I would doubt that it would match a humans skills.
Some parts of the micromanagement could be handled by automization but I doubt that an AI would be able to determine when to use what kind of micromanagement.

Yes! It's not making the game play itself. It's about giving players higher-level tools, similar to the way attack-move and patrol give flexability for more control, but hopefully new innovations worth a decade of RTS and interface improvements.

Play smarter, not faster.

I think it is possible to make the automations more useful for some complex situations by having a fast way to configure some more detailed unit ai options.
Console commands might be a good way to achieve this, for example, typing "AVO PW 200" might make the selected units constantly avoid any enemy peewee within a 200 distance.
Some console commands might need mouse controls, like typing "NGZ AIR" then you would need to use the mouse to select an area as a no-go zone specifically for your air units.
Typing out multiple console commands at once should be a possible choose, and there should be some easy ways to know the current ai behaviors of the units and have some easy ways to clear them.
This seems a bit nerdish, but it might come out very useful.

I think this is totally impractical. Constantly typing stuff takes a LOT of time.

Constantly clicking stuff also takes a LOT of time. If theese commands saves you clicks you might get more spare time. There could also be the possibility to bind them to hotkeys or add them as UI elements by the player.

I'm a longtime Linux user, and even I think this idea isn't viable. The important thing to remember is that with an RTS, you're going to have one hand on the mouse most of the time, so moving that hand to the keyboard is a break in the control flow. (The opposite of a typist having to lift a hand from the keyboard to use the mouse.) Also, as you add these specialized commands to the units, it becomes harder and harder to remove them when they aren't needed anymore. In this case, it's more ideal to stick with standard UI elements to submit these commands.

They would have literally to replace all mouse commands, which isnt really possible.
The move of the hand between mouse and keyboard takes far far too long. Typing a whole word to give a command itself takes far too long, too. Doing a few clicks is way way faster.

No, they wouldn't have to replace all mouse commands. You are making a strawman.

It boils down to this:
You are against automatization even though you haven't really tried any games with automatization.
I'm pro-automatization and want to see how far it can it can improve the ability of the player to make tactical and strategic decisions.

And I'm pro-automization (on a low level), having played several games with automization. But I still don't consider console commands to be a good idea for ingame control. Script control, sure. But commanding units this way is cumbersome and unnecessary.

A good UI should be simple and intuitive. Enabling unit AI through a click of a button, or a hotkey is simple, and if the interface is good, it's intuitive. Typing a command requires removing your hand from the mouse, typing in a fairly long command (assuming you remember it), pressing enter (assuming there are no typos), and returning your hand to the mouse. Not a very smooth workflow.

Don't focus on console commands. I think we can agree there are bad ideas put forward by people for either side of the debate.