A questioned on a game concept.

[SerriaFox]

Not sure what forum to put this in so off topic it goes.

Is it possible for a 4x game requiring micro-manage to the n-th degree, with AI supporting the player to nth-1 degree and still be fun. The only game I can think of tried this was MoO3. I has been a long time since I played MoO3, and I did not play it long. To me the players own AI made his decision relatively unimportant. In fact I vaguely remember one game were i just kept hitting end turn and I kept expanding with no player decisions at all.

So Is the concept of a really complex game with AI support (to minimize micro management) a flawed game concept?

 

[Lillefix]

I'm leaning towards no, but there are all kinds of people I suppose. But I need to think over why exactly.

edit: I mean yes(to the question in the opening post).

You may want to change the question in the opening post or title.

 

[Cheetah]

I think it is possible.

I'm an ardent micro-manager myself, so I was actually able to play MOO3 to some satisfaction. All the controls were there, but the AI just made terrible choices. So every turn I went to all the right controls and forced them to the correct setting, played the turn, and then the AI set everything wrong again, and so on.

MOO3 really did it a horrible way, but also remember that they started out with some other untried concepts as well, and then ran out of time. They ended up cutting a lot of the feature, and other features were only half-way implemented. Add to it that they didn't bother to bug test the product they were selling, and of course it turned out bad.

BUT! I think that if the game is specifically designed around the concept, and the macro-controls - if not also the micro-controls - are available for the player to play with if he so chooses, it is possible to get a very complex, very large game to work without the player having to micro-manage everything.

And if I ever bother to complete my own Civ/MOO-clone, I might find out one day.

 

[Leoreth]

Depends on what n is in my opinion (duh). I think n = 2 is no problem, Civ with decent automated city AIs would work pretty well imo, for example.

There are actually two angles that should be considered:

1) AI capabilities. If the AI is ineffective, nobody would use it, no matter how many layers the game has. The problem is that 99% percent of all games don't use "true" AIs. The AI in Civ games is nothing more than a couple of heuristics that are hardcoded into the game to have the AI make decisions. There's no "planning ahead" or "evaluating enemy possibilities after X turns" like chess AIs do, nor is there any capability of learning. That makes the AI capabilities completely dependent on the gaming skills of its programmer (even if we assume they would go to the lengths of actually implementing an AI that follows the best strategies known to them, which most don't). And if experience shows us something, there are no game designers that come up with more effective strategies than the best gamers who play the game afterwards. Which means that these gamers will have to turn off the AI to gain an advantage. That is independent on the layer we're on.

Will 4X strategy games include true AI at some point? There's nothing prohibiting that, but to put things into perspective: the most complex game with a competent true AI is Chess. AI for more complex games like Go are already inferior to experienced human players. And 4X games are vastly more complex than Go. So I think we'll have to wait a while for that.

2) Being able to control things is fun. As is knowing what's even going on in the game. Taking MoO3 as an example, most of the time I didn't even know what was happening to me in my games. I was moving sliders etc., and then events happened that I assume were related to these decisions, but I had now idea how. You might say that was because of bad design and a horrible interface, and it certainly was, but I still think that's a fundamental problem: fun in games comes from the feeling of being in control, that it's your actions that lead to victory or defeat. Without that, both would feel hollow. And that needs a direct and evident connection between actions and outcomes.

 

[stormerne]

If the AI could support the player to the n-xth degree, where x is selectable by the player according to skill, preference, time available etc., then I believe it's still possible to be fun and consistently so. You select the depth of micro-management you want. The only problem is currently that x is rarely configurable in any meaningful way.

 

[Murky]

There's a fine balance between too much micro and too little. If you automate everything you might as well have the player hit a button that says go then watches the AI play the game for them. If you make everything micro then it wastes too much time on tedious little details like having to click each city screen every turn and managing each worker turn. I think CIV IV does a pretty descent job of finding a good balance. You can adjust how much you want the game to automatically run things. Generally, you still get better results the better you manage your resources and units by microing. MOO2 is probably the best 4X game at providing a good balance, but it's so outdated now.

 

[History_Buff]

Yes, with Paradox's titles (assuming they meet the definition of 4X, which is vague at the best of times), the one's I've played being Europa Universalis 3 and Victoria 2. Victoria has always had far more micro-management than your average game, but a lot of it is optional, at least on your first play through. To do well, you need to get involved.

 

[GoodGame]

Not sure what forum to put this in so off topic it goes.

Is it possible for a 4x game requiring micro-manage to the n-th degree, with AI supporting the player to nth-1 degree and still be fun. The only game I can think of tried this was MoO3. I has been a long time since I played MoO3, and I did not play it long. To me the players own AI made his decision relatively unimportant. In fact I vaguely remember one game were i just kept hitting end turn and I kept expanding with no player decisions at all.

So Is the concept of a really complex game with AI support (to minimize micro management) a flawed game concept?

Your description of the problem is kind of vague, but in general it's like you're describing a simulation in which the player's only make decisions at the top-most level. On one hand, I guess that would work, on the other hand, why make a game so complicated that the average player can't play it? In a way, an advanced AI aid is making the game have broader appeal---to appeal to casual gamers and players that like micromanaging details. But why not just abstract those micro-managed details into the ruleset, so they're present but as "chrome".

I'd agree that MOO3 is an example of flawed game were the AI does the work, and the player is just bored, but MOO3 has more problems than just that: overly complicated GUI that is like a spreadsheet and takes many, many clicks to navigate.