We need a new game model without AI

[SpencerH]

I've been thinking about this for a while so I thought I'd post it here for discussion.

In the old days (pre-PC) we had games and wargames that were played human v human. One problem was that you needed other people who were also interested in the same game. That's OK for monopoly, but it was especially problematic if you played complex wargames. Then came the computer revolution that allowed us to play games without another human player. Great, but while fun, the games were not as mentally challenging as the previous boardgames. Now we are getting human v AI games where the AI's are getting better. Unfortunately, the best AI's today still only have the thought processes of an ant. Therefore, the complexity of these games are 'dulled down' in order to make the AI competitive without too many 'cheats'. The 'game' becomes learning a set of moves that virtually beat the AI every time. Once players understand how to beat the game, many look for ways to make the game more challenging such as the civ 1 city game etc or stop playing until the next version is published.

What I'd like to see developed are human v human games that are only playable MP. NO AI. Without the restrictions of the AI, I believe that civ-type games and wargames could be created that eliminated the necessary over-simplifications that are so irritating to some of us. Broadband internet connections already make it possible to interact with people from around the world so it's no longer a problem to find other game-players. Imagine playing the old SPI board wargame "The War In The East" with someone (or a group) somewhere across the world without the tedium and problems of casting die, checking rules, and moving precariously balanced stacks of armored divisions. I envision the computer and game, without AI, as a facilitator for complex human v human competition built around the game rules, rather than a competitor itself. For me, that would be a great improvement over the current (repetitive) game choices.

[pg]

what's so wrong with games like civ, or eu2 in multiplayer? imo strategy games have been and always will be primarily single player games so even if you did design a game like you described i doubt you'd find many players(but i may be wrong). while the games we have today might not be as complex as what you described as your ideal they are plenty competitive and very intense when played versus other humans. i don't see what you are complaining about frankly.

[SpencerH]

Is that your expert opinion?

[Asmodean]

spencerH: You should give Galactic Civilizations a try. It has some of what you request, although it's AI is just that: An AI.

Furthermore, I happen to disagree with you. I always thought single player was more fun. I never liked tha hassle of trying to get schedules to match, finding substitute players, if one couldn't make it that day, or just rustling up enough players to get a decent game going.

But that's just me. If you have a steady supply of people waiting that are always ready to play whenever YOU feel like it, I totally understand your opinion. But personally, I like to play more than once or twice a week. MP never seems to provide more than that.

Asmodean

[SpencerH]

I'm not saying that they should do away with SP games. Its just that MP versions of SP games are often inadequate for human v human because of the AI restrictions in the SP versions.

I'm gonna give Galciv a try as soon as I get another PC (I'm not sure when, money is tight). For now I'm learning to beat RoN with the demo and an old, slow, PC that craps out about half way through the game.

[Sore Loser]

Spencer, look at my thread named "Massive Multiplayer Civ?" and tell me if that's getting closer to what you're thinking of. One of the key ideas there is to do replace the AI with human opposition.

[SpencerH]

Yes, it appears to be an example of a MP game-type that is possible once you exclude AI. If we go back to my "war in the east" example it might be possible for different players to be army group commanders playing MP, just as we did while playing the board game.

[Inverse Icarus]

i've pondered that myself, but then your entire game rides on the availibility of a community.

if it's too complex for a lot of people to like, you have less sales.

if people can't / don't play multiplayer (connection, time, etc) you have less sales.

if you were going out to make a profit, making a MP only game wouldn't be in your best interests. a small dev house that wants to build a community might be able to do it though.

[SpencerH]

I dont play much (any) doom type games but it seems to me that they are built around a similar if not the same model as I'm suggesting. AFAIK the most popular use of those games is as a vehicle for online human v human play without the limitations of the AI (which is what I'm suggesting for strategy games).

You're right though that the gaming industry might not see a profit but I disagee with them.

[pg]

fps type games are extremely different from civ. the biggest difference is probably the length of the game(20mins per duel or teamplay game in most fps) and the fact fps games are much more social. fps games also have dedicated servers so people have a place to meet up(the most popular servers are always busy 24/7). observing or sitting on the side lines of these games is also quite easy. most servers usually have a capacity for 16(but many have more now) and usually only 8-12 play max per game. many servers have 8 or more spectators waiting to join the game quite a bit. also i'm pretty sure the people who play fps games are much different than civ players and much more inclined to be competitive and try multiplayer gaming.

civ could have a much stronger multiplayer community than it does now if it took a clue from fps games or battle.net. setting up a civ multiplayer game is a pain in the ass and patience seems to be the most important skill a civ player needs.

[Inverse Icarus]

exactly. not everyone can (or wants) to devote 4-72 hours of gametime.

[SpencerH]

What does time of play have to do with removing AI-based game limitations?

[Inverse Icarus]

Originally posted by SpencerH
What does time of play have to do with removing AI-based game limitations?

removing AI means you need 2 (or optimally, 4-8) people to sit down and play at the same time.

take civ. we had a lan, and in 15 hours we were still in the ancient age.

not a lot fo people want to do that, and without an AI, thats the only way to do that.

unless you love PBEM, which i don't.

[SpencerH]

Originally posted by Uber KruX


removing AI means you need 2 (or optimally, 4-8) people to sit down and play at the same time.

take civ. we had a lan, and in 15 hours we were still in the ancient age.

not a lot fo people want to do that, and without an AI, thats the only way to do that.

unless you love PBEM, which i don't.

The answer is to be able to save MP games that can be re-started later.

My ideas about removing the AI came from CIV3 PBEM. The AI civs were squashed quickly by the humans. Then all that was left was game played human v human with AI-based limitations about how the game functioned.

[Jac de Molay]

I wouldn't necessarily rule out an AI, but have a completely programmable one where you could set the parameters of what you want it to build, more exact strategy paths, etc.

This would work great in SMAC.

[Spiffor]

I would never buy a Civ game that has no SP. Well, they can do whatever they want with MP, I don't care because I don't play MP at all. But if they half-assed the quality of the AI to improve the MP game, I'd be very unhappy with it.

What I enjoy in Civ, unlike other games, is that I can take as much time (or as little time) as I want. For example, I have dropped a captivating game as the Koreans a few months ago, and I'm sure I'll resume it in a few months in the future, when I feel like it. This is my way to play Civ, and I sure don't want to sacrifice this schedule freedom.

[Garth Vader]

A number of years ago I played VGA Planets which was essentially what you suggest. It was meant as an 11 player PBEM game, and when you actually got players who would stick with the game it was great, one of the best gaming experiences I have ever had.

The problem, and the main reason why I stopped playing is that real life time constraints mean I want to play a game when I want and not worry that my empire will have problems if I miss a turn. I am not sure the best way to get around that.

[Maquiladora]

I can see what Spencer is saying but its not that much of a problem, besides, no one wants to play a hugely complicated game online. The best thing is a perfectly balanced game, AI or no AI, thats where the longevity in MP is.

By the way I dont think FPS are more social, in fact TBS are more social since you have more time to chat between turns and you usually get along with the people youre playing in some way if youre in a long interesting game. But in FPS, omg.... OMFG BS H4X n LIES TBH, and you quit, **** that.

[Trifna]

Would it be possible to get around the problem with a U-turn by making computer games closer to ancient board games?

With board games, we could all be together and laugh, and drink, and eat pizza... all around one piece of cardboard. Would it be possible to get this around a PC, or a few laptops connected to each other with a central one?

This is actually what I've been thinking about recently. This old cardboard Risk still has advantages over this computerized no-dices no-hassle no-trouble PC version. Could it be changed?

[Panzeh]

War in the Pacific is very detailed, but it's nearly unplayable and the AI sucks. We don't need details down to the squad level.

Basically, its AI will play one basic strategy, and if that's not possible anymore it just 'shuts down'

[Quezacotl06]

Maybe the AI can constantly undergo improvements? It watches the player play the game over and over, and it just emulates strategies. Maybe the player data can just be sent to Firaxis over the internet, where they can quickly find imbalences in the AI and fix them (like an automatic update or something)

You'd just need a pattern-recognition filter built into the game, and the AI "learns" how you win, and why you win, and those counterstance areas are what they focus on the most.

For example, in SMAC, the SP "The Virtual World" is invaluable to the University Faction. So if you're an AI playing the University, then the AI concentrates on making sure they never get it (Either by nabbing it themselves, trading the tech to friends more urgently to encourage others to try and get it, or conquering the base itself). Eventualy the AI could even learn to mind-control the base 1 turn before the project reaches completion (if you're a smart player and you try doing that yourself)

[Brent]

But only if and when you want it to. Maybe you will always be able to play without it taking advantage of its analysis of you if you want to even after it has gotten to know you.

[Quezacotl06]

Yeah, you could have the difficulty settings, and choose between the standard AI and the "adapted" AI, both with their own difficulty levels. An automatic update can improve them both, too.

[Epistax]

-Replying to the initial statements-

I both agree and disagree. AI should be completely modular, that is, an AI sees and knows nothing that a player doesn't. The first impact of this is your desire: the AI has absolutely no effect on the game engine.

My rules of an AI:
1) No AI shall know if a given player is Human or AI.
2) An AI shall be given the same amount of information as any Human who would be controlling the same player, with the exception of rule #1.
3) An AI shall have no pre-programming aside from the interpretation of the rules of the game (eg "This is good", "This is bad"), all further technique is extrapolated, or assisted based upon the difficulty setting.
(obvious 4: no engine parameters can depend on the type of controller (Human or AI) for a given player)

This does make for a slow AI, however the alternative is a cheating AI. If you want to play multiplayer, the AI simply isn't used. At any time a Human can be subbed for an AI, and vice versa.

Anyway that was all a tangent. The real deal is:
If you just follow basic principals of software design, AI will never affect the game. Either the AI and game engine are separate units, or you are doing it incorrectly.

[Urban Ranger]

I completely agree. I totally hate computer players that cheat.

[Shogun Gunner]

Back, again, to the OP, I agree with SpencerH. The strategy games are dumbed down because of computer logic limitations. This still limits the human players -- even when no AI players are active. The mechanics of the game are modified to accomodate the limited AI intelligence. The type of intelligence needed in the AI programming is just not there yet -- at any price.

When I was younger, I also played the SPI, Avalon Hill, et al wargames. A lot of fun, but tough to find players. I ended up playing Squad Leader solo because no one I knew played that game. Thinking back, the amount micromanagement needed to play that game was astronomical.

If they would just take these older wargames, computerize them, I would be the first to purchase and play these games against online opponents. I wouldn't even demand state of the art graphics. The fun was in the strategy and competitive nature of the wargames. SPI must have made 200+ board wargames. Some of them were awesome!

/me 's thoughts fade back to an era which is forever gone

[Urban Ranger]

Originally posted by Shogun Gunner
When I was younger, I also played the SPI, Avalon Hill, et al wargames. A lot of fun, but tough to find players. I ended up playing Squad Leader solo because no one I knew played that game. Thinking back, the amount micromanagement needed to play that game was astronomical.

I used to play Star Fleet Battles, beginning with nothing less than plotted movement.

What fun

[Addled Platypus]

has any other ASL player used : http://www.vasl.org/



1/4 turn plotting by e mail....oh yeah

[vondrack]

I definitely understand what SpencerH is talking about and very much agree that it would be great to have an MP-only strategy game. Over the last two years, it has become painfully obvious to me how much must be sacrificed in order to have a half-decent AI.

Surely nobody's asking for the Civilization franchise to give up SP (and consequently, AI). But a strategy game designed from the ground up with MP only in mind could be fantastic.

Having been an SP player years ago, I now find SP a bit dull compared to PBEM/Pitboss (simultaneous MP has never become my cup of tea... that's not how I like to play).

[Kuciwalker]

Originally posted by Shogun Gunner
If they would just take these older wargames, computerize them, I would be the first to purchase and play these games against online opponents. I wouldn't even demand state of the art graphics. The fun was in the strategy and competitive nature of the wargames. SPI must have made 200+ board wargames. Some of them were awesome!

They will never do that for any game that is sold today, because it would cannabalize sales of their absurdly high-priced miniatures and rulebooks.