Author Topic: AI War state of the game  (Read 45139 times)

Offline Faulty Logic

  • Hero Member Mark III
  • *****
  • Posts: 1,194
  • Bane of the AI
Re: AI War state of the game
« Reply #105 on: November 09, 2012, 08:19:40 pm »
"I do things the exact same way every game, then complain about it."

Quote
What strategy is there in this?
As before: which planets you take, when you take them, how long you hold them. This can't be the same every game, unless you were playing the exact same map over and over.
If warheads can't solve it, use more warheads.

Offline TechSY730

  • Core Member Mark V
  • *****
  • Posts: 4,570
Re: AI War state of the game
« Reply #106 on: November 09, 2012, 08:58:45 pm »
Despite strategy and tactics are technically different concepts. But in practice, there is quite a bit of overlap; the distinction is not clear cut. I think this is why there is so much confusion, and why you can't completely ignore the tactical aspects even in a strategy focused game. (Not saying the devs are doing this, but just pointing out that severe enough tactical "issues"/"shallowness" can start hampering the strategic aspects of the game, and vice-versa)

Also, in even in most RTSs, if you playing against defensive focused (turtling or "turtle-ish" opponents), and let them do their thing and establish themselves, then you will almost require brute force, as their defenses will be strong enough to shut down all but the most advanced, micro-intensive non-brute force solutions.
In a traditional RTSs, when facing turtling players (in PvP mode), you can usually shut them down with good offensive tactics before they have the time to setup their defenses (turtling tends to be slow and expensive to setup, and gives your opponent lots of freedom to do whatever, so it is sort of a risk). In AI war, the AI has already established itself (by virtue of winning the war and thus already having an empire at game start), which means the turtles got themselves to this state before you can really do anything about it. This means "blobbing" will almost certainly dominate that game. This is sort of the reason why many players don't like the single player campaigns of RTSs, you usually are facing a pre-established opponent that is only sort of aggressive, leading to similar "blobbish" tactics.

If you don't like facing pre-established defensive players and their effect on the effectiveness of tactical options, then you can disable them. Facing an established turtle is usually one of the most annoying things to take out, and in AI war, your opponents are indeed pre-established.

Offline TechSY730

  • Core Member Mark V
  • *****
  • Posts: 4,570
Re: AI War state of the game
« Reply #107 on: November 09, 2012, 09:19:12 pm »
This said, I would like to see the planets that use the "preferred ship types" logic prefer them more strongly (aka, reduce the chance that it will chose a non-preferred ship type), and possibly even give some variety in the distributions, like:
Key: (Chance for ship type A, type B, type C, other)
(75%, 15%, 7%, 3%)
(90%, 5%, 3%, 1%)
(32%, 32%, 32%, 4%)
(50%, 35%, 10%, 5%)
(95%, 3%, 1.5%, .5%)

Things like that.

Also, I think the percentage of planets that use this logic could go up some.

And finally, the "rerolling" of the selections when the AI gets a new ship type can throw a monkey wrench in this scheme sometimes, causing much more mixed distributions than indended.

Possibly bump up the chance of a planet spawning with this logic

Offline Wingflier

  • Core Member Mark II
  • *****
  • Posts: 2,753
  • To add me on Steam, click the little Steam icon ^
Re: AI War state of the game
« Reply #108 on: November 09, 2012, 09:32:12 pm »
@Chemical Art -

Sorry for the confusion, but I'm not talking about tactics, I'm talking about strategy.

In every other strategy game I play, I have unique, dynamic, and ever-changing situations that I must handle quite differently every time:

Commanding in Natural Selection 2 - I must react to what the opposing Commander is doing, and the strategy he's employing, while also attempting to take the map on my own, holding onto mission critical resources and locations.  Since I don't have direct control over my marines (they quite literally have a mind of their own), much of the strategy is in presentation and communication with my squads.  Every trick my alien opponent uses, I have to react accordingly or he'll take the map and win the game.  I need to know what upgrades to research when, and make sure my soldiers are making full use of them.  No two games are even close to the same.

Wargame: EE
- The map layout, as well as the unit composition my opponent and I start with, and how we use them, make for a completely different game every time.  I have to react to exactly what he's doing and what he's doing it with, or I will lose.

Battle for Wesnoth
- The race he chooses, and the unit composition he uses, as well as how he uses them, makes for a completely different experience and set of decisions every time.  Every game I must counter and react differently, no two games are alike.

In these games, I can't begin with a "set plan" and expect to have any kind of success.  I have to constantly be changing and reiterating my strategy to fit the circumstances and the ever-changing scenarios I'm challenged with.  That's what a strategy game is!

Here is how I play EVERY GAME of AI War:

1. Unlock both MKIII Harvesters.
2. Take all the planets adjacent to my HW.
3. Put all my defenses on all the planets adjacent to my HW.
4. Blob my way to the Core Shield Generators.
4a. If there's an Eye, send a small raiding part to take out the guard posts, then back to #4.
5. Send ARS Hacking Ships to the Core-Shield Type A Planets.  Set my blob on top of the ARS Hacker and twiddle my thumbs for ten minutes.
6. Do the Champion Scenarios whenever possible and necessary (these are also very simple and repetitive...very little strategy).
7. Unlock the best ARS unit of the 3 (it's an easy choice every time because the bonus ship balance is so bad).
8. Place one of my free unlocked Zenith/Spire/Neinzul Modular Command Stations on each of the taken ARS worlds.
9. Blob my way to the closest Factory planet and take it.
10. Once all the Core Shield Gens are down, blob my way to the first Homeworld.

All I have to do is make sure I have my blob to defend the exo attacks every few hours, and some of the bigger waves, and voila, I win!

Can I ask you, where is the strategy to this?  How do I have to alter my strategy ever in any given map or situation? 

It is so formulaic and monotonous, for a game that prides itself on its "strategic value", I'm surprised that I'm the only person that sees how shallow the strategy really is.

The average AI War game for me is about 10 hours.  In 10 hours, I may make 1 or 2 important strategic decisions, the rest of it just goes according to plan.  The 3 strategy games I listed at the beginning last, on average, 30 minutes or less, and I make MANY more strategical decisions in that short amount of time than I do in a 10+ hour AI War game.

I've increased the difficulty again and again.  It doesn't change the strategy, or the tactics.  The only thing that changes are the wave sizes and the reinforcement counts; that's it.

Explain to me how this is a strategy game.
"Inner peace is the void of expectation. It is the absence of our shared desperation to feel a certain way."

Offline Lancefighter

  • Core Member
  • *****
  • Posts: 2,440
Re: AI War state of the game
« Reply #109 on: November 09, 2012, 09:54:23 pm »
You forgot scout.

Ideas? Suggestions? Concerns? Bugs to be squashed? Report them on the Mantis Bugtracker!

Author of the Dyson Project and the Spire Gambit

Offline chemical_art

  • Core Member Mark IV
  • *****
  • Posts: 3,952
  • Fabulous
Re: AI War state of the game
« Reply #110 on: November 09, 2012, 10:05:51 pm »
For the games you describe Wingflier, you are asking to react to the opponent.

You are already banging your head against AI world against this desire.

http://www.arcengames.com/mediawiki/index.php?title=AI_War_-_Controlling_The_Tempo

Others can describe more intricately why this is so, but from the start you are already running against the grain for the game design.

Secondly, the games you describe are PvP games, not player vs AI games. Maybe the games you mentioned have AI's, but you are not describing the AI of those games vs the AI of AI Wars. PvP is inheriently a more varied game, and no amount of programming in this generation at least is going to be able to compete in that regard. You are playing AI Wars.

In regard to the railroading of CSG games, I agree, which is why I don't play it. It forces you to do some things, period. It runs counter to the options heavy nature of AI Wars. But the thing is you don't have to play it, which is why I don't protest it (as long as it doesn't later become mandatory.)

At the core of every strategy game, you have a formulaic components. Take resource A, defend location B, etc. Formulas are not the issue, its how you accomplish said formulas.

Difficulty does play a role here in regard to your steps for AI Wars games.
Step #2 is difficult on 9/9 non-defensive AI's (in terms to winning a game), and almost suicide on 10/10 sans other factors like Fallen Spire. Step #5 is difficult past ARS hack #2 after 9/9 and almost impossible after hack number 3, meaning you have to choose which hacks you do in addition to what units you select.

Given that over time the "worst unit polls" are become closer over time, balance is certainly improving. In addition, some units are great for blobbing  but some units are outright terrible but still very effective (autobombs anyone?). What units you considered the best are not the best for other players. 

You don't mention defense at all, but it is very real. Timing of when to take a factory IV when you have exo-waves, for example. The moment you take a factory IV you have both a boost and a liability. Because it is a fact that the longer you hold a factory IV, the more likely (over time) that something will cause its destruction. The odds of this happening increases exponentially on high difficulty games because you simply cannot afford the AIP hit to have a buffer zone between borders and your factory IV. So on games that are difficult you must thread the needle about taking the IV so you can push toward the AI HW, but not taking too long  before you risk losing it... unless you put a disproportionate amount of defenses to defending said factory IV and in the process threaten other resources.


What makes high difficulty games more difficult somewhat is that your choice of strategic options is limited. You are more limited in what worlds you take (which is worth its own thread in the strategy section), what choices in regard to hacking, what ars to hack, and what ship choices to show.

The choice of what worlds to take is in the end the core of AI Wars.

If you don't like this mechanic, you are in the end fighting the very mechanics of AI Wars, which goes full circle to the beginning of this post.
Life is short. Have fun.

Offline TechSY730

  • Core Member Mark V
  • *****
  • Posts: 4,570
Re: AI War state of the game
« Reply #111 on: November 09, 2012, 10:19:32 pm »
Here is how I play EVERY GAME of AI War:

1. Unlock both MKIII Harvesters.
2. Take all the planets adjacent to my HW.
3. Put all my defenses on all the planets adjacent to my HW.
4. Blob my way to the Core Shield Generators.
4a. If there's an Eye, send a small raiding part to take out the guard posts, then back to #4.
5. Send ARS Hacking Ships to the Core-Shield Type A Planets.  Set my blob on top of the ARS Hacker and twiddle my thumbs for ten minutes.
6. Do the Champion Scenarios whenever possible and necessary (these are also very simple and repetitive...very little strategy).
7. Unlock the best ARS unit of the 3 (it's an easy choice every time because the bonus ship balance is so bad).
8. Place one of my free unlocked Zenith/Spire/Neinzul Modular Command Stations on each of the taken ARS worlds.
9. Blob my way to the closest Factory planet and take it.
10. Once all the Core Shield Gens are down, blob my way to the first Homeworld.

All I have to do is make sure I have my blob to defend the exo attacks every few hours, and some of the bigger waves, and voila, I win!

Can I ask you, where is the strategy to this?  How do I have to alter my strategy ever in any given map or situation? 

It is so formulaic and monotonous, for a game that prides itself on its "strategic value", I'm surprised that I'm the only person that sees how shallow the strategy really is.

The average AI War game for me is about 10 hours.  In 10 hours, I may make 1 or 2 important strategic decisions, the rest of it just goes according to plan.  The 3 strategy games I listed at the beginning last, on average, 30 minutes or less, and I make MANY more strategical decisions in that short amount of time than I do in a 10+ hour AI War game.

I've increased the difficulty again and again.  It doesn't change the strategy, or the tactics.  The only thing that changes are the wave sizes and the reinforcement counts; that's it.

Explain to me how this is a strategy game.


This is partially due to, as you mentioned, how unbalanced certain things are (about a 1/3 of the bonus ship types are a bit too good or rather bad, which I agree is not good balance, and econ is still rather messed up, especially with the harvester balance, though it is a bit better)

And, this is partially due to the nature of a PvE game, in particular how the AI doesn't care about you very much unless you start playing stupid and attract its attention too much. (This means you can set your own pace mostly, which usually means you can select your own strategy and make it work decently). The AI does not have the same objectives and progression; it does not have to try to grow and hold off your attempts to stop its growth while doing the same to you, as it came pre-established, and it can almost always overcome your production levels if you give it reason to; the AI has already won the "build-up turf and tech" war.

And, this is partially due to how you are picking the same options. (Have you tried a game without champions or fallen spire? Does it feel different?)

And, this is partially due to how the AI is rather dumb with its tactics (it rarely does anything beyond "build up blob to defend" and "send blob to attack", though steps are being taken to remidy this, but it is a slow process)

And, this is partially due to how the AI never really takes steps to regain "turf", only hamper and reduce yours. (What happened to the plans to let the AI take back planets? That would add a fun new aspect). Aka, the AI makes almost no strategic choices/actions (even less than its tactical actions), even when you start showing yourself a major threat.

And, this is partially due to how there is little in the way of physics/terrain effects and each planet is pretty much the same in terms of how units operate in them and what can happen in them.


There are many other factors going on too.

A few of them are rather deep faults of the game, but it seems that most fall into other categories. Some of them are just the game poorly advertising the effectiveness of other strategies/techniques. Some of it is simple balance problems. Some of it is just the nature of this type of game, especially with how stratagies (both yours and the "opponents") play out very differently in a PvE style game. Some of it is playstyle and preferences.

It's a complex subject, as this thread is revealing. ;)

EDIT: Many of these point chemical_art ninja'd me on.
That and read that article. It explains why the human seems to have so much freedom in making a strategy effective.
It may just be that this style of game isn't what you are looking for in a game.  :-\

Offline Wingflier

  • Core Member Mark II
  • *****
  • Posts: 2,753
  • To add me on Steam, click the little Steam icon ^
Re: AI War state of the game
« Reply #112 on: November 09, 2012, 10:25:23 pm »
@ Chemical_Art - I haven't seen that the higher difficulties change the strategies or make the decisions any harder.

I disagree with you that just because this game isn't PvP, it can't force you to make unique and important decisions every game, more than once or twice a game - maybe several times every hour.

Honestly, I don't believe that what AI War has turned into is what Chris originally designed it to be.  I don't think he wanted the only real danger of the game to come during huge waves and exos, and the rest to be a formulaic grind to the finish.  It used to be much different than this.  AI War has improved in a lot of ways, but it has also gotten worse in others.

To me, there's not much strategy left.  If I can make a plan from the very beginning, and not hit any road bumps or obstacles along the way, then I'm no longer playing a strategy game, I'm just rehearsing a build order basically.  Why are counters in the game?  What difference does it even make?  You take your blob, you move it against their blob.  The counters become so pointless and superfluous - if you have a little bit of everything you beat everything.  What's even the point of it all? 

Maybe you're right, maybe AI War just isn't the game for me.  I just see a game that could be so much better.  It could actually be about strategy instead of what it is now, which is rehearsing build orders and blobbing to victory.  If the community is fine with that, I guess I just need to move on.  I'm not fine with that.  It wasn't always this way for sure.  There were times when blobbing from planet to planet was the worst decision imaginable, back when it was an actual strategy game.
"Inner peace is the void of expectation. It is the absence of our shared desperation to feel a certain way."

Offline Lancefighter

  • Core Member
  • *****
  • Posts: 2,440
Re: AI War state of the game
« Reply #113 on: November 09, 2012, 10:32:01 pm »
Clearly, we need pvp in aiwar. Keith, if you would..
Ideas? Suggestions? Concerns? Bugs to be squashed? Report them on the Mantis Bugtracker!

Author of the Dyson Project and the Spire Gambit

Offline Gallant Dragon

  • Full Member
  • ***
  • Posts: 144
  • An RTS gamer in an FPS world
Re: AI War state of the game
« Reply #114 on: November 09, 2012, 10:33:56 pm »
To me, there's not much strategy left.  If I can make a plan from the very beginning, and not hit any road bumps or obstacles along the way, then I'm no longer playing a strategy game, I'm just rehearsing a build order basically.  Why are counters in the game?  What difference does it even make?  You take your blob, you move it against their blob.  The counters become so pointless and superfluous - if you have a little bit of everything you beat everything.  What's even the point of it all? 


Lurker here, just popping in to say:

THIS.

The original idea behind AI War was to have a game that could be unpredictable and which forces you to strategize whenever the AI throws a spanner into the works, right?

Maybe we need to take another look at the difference between all the iterations of AI war and not only look at what's changed but make an analysis of how that's changed how the game is played.
It's just carriers all the way down!

Offline chemical_art

  • Core Member Mark IV
  • *****
  • Posts: 3,952
  • Fabulous
Re: AI War state of the game
« Reply #115 on: November 09, 2012, 10:45:46 pm »
@ Chemical_Art - I haven't seen that the higher difficulties change the strategies or make the decisions any harder.

Have you won a 10/10 or 9.6/9.6 game with 80 planets with CSG with your current formulaic tactics? High difficulty is relative to the player, so I went to the extreme and went 10/10, but I am really curious if you can manage that with your current tactics. Until you manage to win with the game at its highest difficulties, saying you don't notice a difference in difficulty is like saying in a game with easy, normal, hard, and nightmare that you don't notice a change in strategy from normal and hard and thus nightmare is the same strategy as well.


I disagree with you that just because this game isn't PvP, it can't force you to make unique and important decisions every game, more than once or twice a game - maybe several times every hour.

I am curious at what strategic games games go 10+ hours and have several game changing decisions every hour. The longer a game goes, the less strategic decisions over time (due to a game, any game, only able to have so much depth). A game of 30 minutes (which you mentioned in your previous post) is going to have more strategic decisions over time because there is less time to begin with.

Honestly, I don't believe that what AI War has turned into is what Chris originally designed it to be.  I don't think he wanted the only real danger of the game to come during huge waves and exos, and the rest to be a formulaic grind to the finish.  It used to be much different than this.  AI War has improved in a lot of ways, but it has also gotten worse in others.

The post I gave that you  (AI tempo) was from Chris himself from several expansions ago. Before exo-waves. Where the only threat was CPA's and AI waves. The original AI Wars from what I understood did have the AI and player be equal, but Chris made the game changing decision that they would not, and in some period after that decision he made that tempo post. You have to be A LOT more specific on what you think Chris wants given just how much this game has changed.

To me, there's not much strategy left.  If I can make a plan from the very beginning, and not hit any road bumps or obstacles along the way, then I'm no longer playing a strategy game, I'm just rehearsing a build order basically.  Why are counters in the game?  What difference does it even make?  You take your blob, you move it against their blob.  The counters become so pointless and superfluous - if you have a little bit of everything you beat everything.  What's even the point of it all? 

You play the game the same way, and get the same results, and complain about the game being consistent? I don't see how it is different then if in a PvE strategy game someone rushes and wins, or turtles and wins. If you can demonstrate to me that using two random seeds, where you play  one game with 10/10 mad bombers and another game 10/10 tech turtles and using the same exact research techs and manage to win, then I'll take my hat off to you. Because if you are using different tech trees in reaction to different AI's, you are demonstrating different strategic decisions.

Maybe you're right, maybe AI War just isn't the game for me.  I just see a game that could be so much better.  It could actually be about strategy instead of what it is now, which is rehearsing build orders and blobbing to victory.  If the community is fine with that, I guess I just need to move on.  I'm not fine with that.  It wasn't always this way for sure.  There were times when blobbing from planet to planet was the worst decision imaginable, back when it was an actual strategy game.


You keep assuming that the game has no strategy, despite evidence otherwise. Just because strategy is in choices in planets and techs instead of tactical choices doesn't make it any less strategy. I'm REALLY curious what version of AI wars didn't encouraging blobbing, since it certainly wasn't before 3.5ish era when I joined up.
Life is short. Have fun.

Offline Gallant Dragon

  • Full Member
  • ***
  • Posts: 144
  • An RTS gamer in an FPS world
Re: AI War state of the game
« Reply #116 on: November 09, 2012, 10:55:35 pm »
Dammit, stop making such good counterarguments.   :-\
It's just carriers all the way down!

Offline chemical_art

  • Core Member Mark IV
  • *****
  • Posts: 3,952
  • Fabulous
Re: AI War state of the game
« Reply #117 on: November 09, 2012, 11:05:14 pm »
Dammit, stop making such good counterarguments.   :-\

Wingflier has taught me well  :P
Life is short. Have fun.

Offline TechSY730

  • Core Member Mark V
  • *****
  • Posts: 4,570
Re: AI War state of the game
« Reply #118 on: November 09, 2012, 11:07:19 pm »
I do agree that we need to look at the 2.0 days and see why the waves were more threatening and the AI seemed to deal with "blobbing" and deep strikes better.

Was it because the AI got turrets back then?
Was it because the player economy was so much weaker? (thus, the AI was competitively stronger to the player at most points in the game as a result)
Was it because the waves started out stronger in the early to mid game?
Was it because the humans had far less tools from the start back then?
Was it because a different balance model between ships, and different "balance targets" for each ship class?
Was it because of the different armor/shield system?
Was it something else?

I honestly don't know.

Someone needs to look at what it was back then and see if any of the concepts, balance goals, and limitations (on the player especially) will fix some of the issues the game has now, while trying to integrate them with the solutions to the issues the game had back then (especially when it came to AI ferocity, even if it is found that it was artificial due to player limitations)

Again, complicated stuff. ;)

Offline Lancefighter

  • Core Member
  • *****
  • Posts: 2,440
Re: AI War state of the game
« Reply #119 on: November 09, 2012, 11:08:45 pm »
20 minutes into an 8/8 game on 2.0, i got a 1.6k unit wave warning. I ragequit.
Ideas? Suggestions? Concerns? Bugs to be squashed? Report them on the Mantis Bugtracker!

Author of the Dyson Project and the Spire Gambit