The use and abilities of assistant AIs

If you want help managing different aspects of your base, find more friends to share units with. Don't remove certain aspects which used to require substantial skill and replace them with AIs doing the work for you.

I know how to make the AI slightly worse at it than normal players... Don't have any AI.

And yet another thread about this topic...

I am against much AI-Help for such tasks and, while I have given alot of arguments against them already in some other thread, at my very core I just dont want these tasks to be done by an AI for a very simple reason. I dont want these things to be easy. They are hard to separate good players from bad players. I will probably play PA 1v1 alot as soon as it is playable and practice such things so I am more or less good at them, so it sets me apart from people who dont practice them. The threshold at which to set this "I dont think this should be easy"-level is totally arbitrary and subjective, which is why I didnt really mention it this directly so far. I.e. Starcraft2 has this threshold far lower: Even making constantly workers is hard to do and people like it. In fact I like it in Starcraft, too, because Starcraft is just aiming for a different kind of RTS.

Very simply put:

And that's why you should lose to somebody who is not lazy.

Clicking aloot and multi tasking is not the skill that should be in focus in a strategy game. The STRATEGY should.

Clicking: okay, thats not really part of it.
Multi-tasking: O_O. That statement shocks me. RTS are ALL ABOUT Multitasking.
If you want pure Strategy you are not supposed to play any real-time games.

You misunderstood me, im saying that multitasking shouldent be the focus of a rts, not that it shouldent be present.

With the amount of multitasking nessesary in for example subcom you dont get any strategy except very crude ones (You cant do more complex ones while doing every small thing needed), the winner should win thanks to the strategy, not how effective he is at multitasking.

Well thats my feelings on the subject atleast, it seems im alone.

This sentence sums up the typical AI-takeover advocate pretty succinctly. They're the kind of single-minded person for whom the concept of doing more than one thing at once is as daunting as dodging bullets.

Area attack command, not AI handover. Unless you want the AI to automatically pick out priority targets, in which case: too bad, that's your job.

Solved by better patrol priorities, not AI handover.

There's no such thing as an optimal build order isolated from external effects, so this is getting the AI to play the game for you. That's your job.

Target priorities and scouting are part of strategy. If the AI is picking them for you, it's playing the game and you're watching.

What? SupCom has no strategy because it has large amounts of multitasking? Are we playing the same SupCom here?!?

Man, if you think Supcom is bad you mush thing Starcraft2 is the Anti-christ or something >.>

Seriously SupCom is so damned easy to Multitask in, not just because of the tools like Strat Zoom and Queue commands, but also because unlike Starcraft1/2 it was DESIGNED that way, so say thet Supcom has not strategy because it has multitasking is just ignorant of RTSs in general.

Mike

Once again im missunderstod. Who said SupCom hade NO strategy (Did you miss the "Except very crude ones")? It has more strategy then most so called rts's on the market. What i did say is that the strategies you see is crude and simple, the winner of a game is mostly decided on the ability to multi-task (And effective builds) and not strategy, and in my opinion that shouldent be the case in a grand strategy game like this.

If i want multi-tasking and aloot of micro il play mentioned starcraft 2, i own the game, i like it, but this isent starcraft, they (the developers) keep mentioning strategy on a epic scale, that cant just mean "aloot of units" can it? Why not convince them to add any function we can that increases the amount of strategy you can pull of?

That reply is just wrong (And kinda insulting to be honest). The point i was trying to make was that: Not using better tools to prevent degrading a skill that with those tools wouldent be as nessesary, is foolish.

Im somewhat good at SupCom and at multi-tasking (Not great, but not to awful either). But just becuse i can win matches by multi-tasking dosent mean that i find the current system perfect or even good.

Strategy games should be about strategy, multi-tasking should offcourse be present, but when a tool that removes some of that multi-tasking is a alternative i fail to see the reason why it should be denied becuse you want people to multi-task more. Its called a Real time strategy game, not a real time multi-tasking game.

Strategy should be the top focus and any tool that can increase the amount of strategy should be there, multi-tasking is something that should be circumvented as far as possible to give room for more strategy.

Wait what? "Designed to do the thing slightly worse than actual people"? This is called a strawman, fyi.
AIs won't be as good as a human player for many things. That's a fact. Things they do well, however, like micromanagement with millions of APM, can be let to them so the player can concentrate on the strategy.

Play Wargame : European Escalation. You need something like 30 APM. I'm a lazy bastard as well, so 30 APM is fine. It's still a hard game.

Well sort of, but more because of the abysmal scenario and stupid unit design. Though it does strike me as lazy for leaning on the "high-APM" skill so much. You can, after all, make a great RTT requiring low (or lower) APM.
But I digress.

The good thing with SupCom is that it was far in advance compared to other RTS for the UI. The bad thing with SupCom is that it's still not enough. Without even speaking about the mass fabs, things like auto-kite is a bare minimum which isn't even there. Area commands are another one, or drag&drop queues, securing a zone, target priority or most things Zero-K did, for example.
Assistant AI is just another UI element in the end. You want to put a new firebase there. The basis of this kind of firebase is always the same. Let the AI do it, then. It will save you a dozen clicks for the same result. Obviously, you should be able to set parameters or change some elements manually. For example, put more AA, add a strategic missile defence or a TMD or what else...
Same thing, you want a rear base to add 3 air factories, tell it to the AI, saving you a few clicks and the need to zoom on said rear base while you are planning your strategy.


Having an AI playing for you and you spectating/intervening on what it does is vaguely related, but quite different. It would be more of sharing units with an AI, interesting to learn the game or for some mods/special game modes.
Here, it's about getting things done in less clicks. It's basically an UI concern.

Your analogy was dumb, and it deserved being mocked. Comparing doing more than one thing at once to dodging bullets implies that it's a very difficult task, not a relatively easy one. And, as everyone has been saying, even with AI takeover, all the tasks it does would still be necessary, because the AI would be bad at them.

No it isn't. Unless you only play on identical flat maps, firebase design is something that varies depending on terrain, the enemy's position, the enemy's force composition and your budget. Letting an AI build one for you is taking the strategy out of your hands.

You know what's not strategy? Letting the AI make important decisions for you.

No it isn't. This suggestion takes strategic decisions out of the player's hands and puts them into the AI's. That's not what the UI is for.

I agree with all of this. The type of automation I'd like to see is more on the order of better area attack commands, better patrols, with the addition of basic automated movement/fight and perhaps group formation settings for units. That would be entirely sufficient. The more configurable all of this is by the player (via intuitive UI toggles) the better.

Dumb or not, it didn't deserve a flat ad-hominem without any form of constructive criticism such as, you know, explaining why it didn't work.

You missed the "this kind", it seems. Though I'd recognise that it wasn't the best way to explain it. The point is that there aren't that many kinds of firebases, and their basis follow a few parameters. Give the parameters, let the AI plan it, tweak it a little bit and you're done. The firebase would be the same than the one you would painstakingly plan, but you obtain it in far fewer clicks.
Or, if you prefer, a more advanced form of FA's templates.

You know what's not suggested here? Letting the AI make important decisions for you. And you know what else isn't strategy? Trivial micromanagement.

I agree with automating tedious things, like getting engineers to build a bunch of deposits, or getting artillery to pick more appropriate targets when firing.

I'm skeptical of your claims to understand what you're talking about here, since you seem to imply that building firebases requires painstaking planning and large numbers of clicks, which just isn't true. And that you compress a lot into some nebulous "tweaking" stage which somehow fixes all the problems of your idea without being well-defined enough for it to have any identifiable flaws.

Building placement, base composition and build orders are strategic decisions, so I don't know how you can claim that this idea doesn't replace strategic decisions. If it's "trivial micromanagement", you don't even need an AI to handle it, you can do it with general UI improvements to things like queuing and grouping. The very fact that you're handing over to something that can be termed "an AI" demonstrates that the actions are not trivial.

Even if you think the analogy is dumb, mocking me for it is a violation of forum rules. Im not a ******* so i havent reported you or anything for such a minor thing, im just asking you to try not insulting people.

And no, it would not be nessesary for you to do every task, even if the AI isent as good as you.

Having the AI do it would result in your having time to concentrate on something else more (A 4 squad combined attack perhaps?), making the result there better then if you were doing everything at the same time.

No, doing everything yourself is NOT a good strategy since it would lead to you lossing in any real conflict ever fought.

Its a very good strategy to delegate your work making you able to concentrate on the more important tasks and leeting the others (AI in this case) take care the less important ones. Doing every task yourself, minor and major, is not a good move, its basic logic for any command position.

I agree that it isent a UI thing, but i want to debate that they remove control. If they were added you wouldent have to use them, they are tools. If you use them you would have to learn when its effective to use them and when you shouldent.

Your still in control. If you want TOTAL control, then maybe they should remove any automatic behavior of any unit in the game? So that the defensive turrets dont fire on a closing enemy unless you tell them too, since that removes control from the player. And attack moves and patrols should be removed too, they remove control from the player.

I think you want to play Gratuitous Space Battles instead of PA, it has no Micro at all and is almost 100% strategy!

Mike

I did a search, its some kind of space combat, not the same thing. But i might try it, looks somewhat intressting.

PA isent made yet so if i want it or not is up for debate, its gonna be a game inspired by TA with grand scale battles, thats all thats been decided. Supcom was inspired by TA and it added aloot of auto futures, i fail to se why PA wouldent. The poll is currently 10 against 10 so its not like your guys that want to keep your multi-tasking is a majority or anything (Atleast not in this thread).

Im not saying that the game should have no micro and multi-tasking, im saying that tools that lower the amount of micro and multi-tasking shouldent be denied.

Like the strategic zoom, the order que's, the repeat build que button on factories (Well, it was in SupCom so im guessing it will be present in PA to). Like automatic unit behaviors. Patrols. Attack moves. etc.

AI assistants would just be another step on the road and a very realistic step. Since as i have mentioned before, theres no supreme leader of a military force in history that has done everything himself without delegating tasks.

Of course it's a different type of game, that's entirely my point, once you start throwing in automation like this you end up with a game with the player as a spectator, not a participant.

SupCom didn't add any 'auto futures', it gave the Player a more powerful UI. Your AI idea isn't making it easier for the player to do things, it's doing things FOR the player, that's a big difference.

Again, I'm all for giving the player a powerful UI that makes doing things easier, but what you want is not something that's easier to do, you want something done FOR you.

Mike

So what? A lazy player should still not win vs a hard working player.

SupCom does it quite right. It makes things as easy as possible without introducing any kind of AI that is doing things that you should be doing yourself. PA should follow that path: Make things as easy to control as possible, but let the player be the only instance of higher intelligence that controls the game.

You want an AI to plan out your base and an AI to micro you battles.
So basically you just want to enter a few parameters into the AI and watch it play?

Many many many many mistakes made in real conflicts stem from idiots at the lower ranks not doing what the higher ranks want them to do. Also total realism isnt always the best thing for a game.

There is a difference between an infinite loop for the build queue and an AI that controls my units for me. The first does not imply any real decision-making, it just makes it possible to achieve something(make infinite amount of tank x) with as few clicks as possible. Controlling units or laying out bases however is something that involves decisions, you cant just automate that with a single button.
There are definitely things that can be further improved, mostly in the area-command area.

Patrol functions, auto attack, order que's, repeat build order on factories, a layout planner. They are all automatic functions. They all do things FOR the player with a push of a button, the AI does things for the player with a push of a button, you may argue that the amount of control lost is different, but other then that theres no difference.

No mather how you argue the functions allredy added between TA and SupCom is not only UI changes to make it easier, they are automaticly doing things for you that you dont have to do yourself. As mentioned the only difference between whats allredy been added and the suggestions in this thread is the amount of control lost to automatics.

But considering that it would bring more freedom on other areas i feel it to be a just tradeoff. (Being able to switch your attention to areas you feel are more important in the theatre of war).

Its not like the assistant AI mentioned would make you work less, it would just switch your work load away from the less important tasks and onto the more important tasks. You wouldent be able to sit back and roll your thumbs (Unless you want to be destroyed by your opponent).

I guess we will just have to agree to disagree. I doubt the argument will go any further without devolving into a repeated loop of the same sentences.