Important vote: Single Player or Multiplayer?

[The Joker]

After a large discussion at the meeting, and some advice from Harel, we have realized that it is likely that we will need to decide on this soon.

Like Harel said at the meeting: We have to pick one. The game can not be both.

Therefore a vote may be the only sollution. Everybody are welcome to vote here, so please do. The two options for now is:

1: We make the game a single player game with some multiplayer capabilities. This, however, means that the amount of players in the game will be severely limited.

2: We make the game a multiplayer game. Have the possibility to have loads of players play the same game, but make the game multiplayer based. So no real AI.

Harel sort of convinced us that we had to pick one of these, and that a hybrid would be impossible.

So this will be the thread for dealing with this. The vote is not required right now, I would in stead like some discussion and different pros and cons of the two choises first. Then we can vote.

[The Joker]

For anyone who was at the meeting it will not be a secret that I am a huge supporter of the 2nd option.

I want this to be a massive multiplayer game. I would like to have as many players in one game as technically possible, and 100 at least. I think this would make the game world alive, and different in every single game.

I would like multiplayer games that goes on and on for months and months. Where the players come and go, as the world developed. Empires would rise and fall, things would go on as they would in history. No winner, no loser.

This means that the game design should be made so loads of players are supported. Not only does the map have to be large enough (which it should be), the focus of the game should also be made so diplomacy with loads of other players is possible and indeed an important aspect of the game.

The reason why I dont think we should spend too much focus on AI and single player games are, that the game is going to have a huge variables, statistics etc. And since an AI is one of the more dificult things to program at ALL, it is likely that we will never be able to make an AI that is good enough.

And since people will already have downloaded the game from the internet, and are likely to take part in a development discussion at our forum (at least suggest some ideas) they might as well play the game against other players online. And having loads of players will mean more fun diplomacy, and finally with real intelligence to do it with!

That is why I support #2.

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[cavebear]

Well, I will vote anyway, but more to make clear the reasons I mention afterwards...

"Multiplayer"

Ok, that's out of the way. Now, as to *why*:

1. The biggest difficulty in Civ games (and others, I'm sure) is that creating a good AI is nearly impossible. You either have to lower the game strategy to what an AI can do, or you have to make the AI cheat horribly, which really creates havoc. Besides, players will focus on how to exploit AI weaknesses rather than improve their own game-play. Eliminating an AI will eliminate a lot of programming requirements.

2. Focussing on interactive human players allows the players to be endless and cunning AI's (so to speak). What I mean is that allowing human players to do all the decision-making will bring out the best play right from the start.

3. I think that the game should *allow* for self-play (like Civ hotseat) and 1 vs 1 games. That will be important for learning the basics of the game (it *is* going to be complex). Plus, it is important to allow for people to test strategies themselves, and a hot-seat would eliminate a need for an AI.

4. People have come to expect a Multiplayer format. As a practical matter, the game will not likely succeed without it.

[heardie]

No.

I think that single player would be the best option. The game can not work, as multi-player, espically in its infancy. When the game is still new there will be not that many people, and therfore no one to play with.

There are many advantages to single player, like being able to play the game when not connected to the internet. This, IMHO, is a great advantage.

------------------

[The Joker]

Would it, Heardie?

Fact is that we are not going to have an AI any time soon. In fact it can not be constructed untill we reach alpha or beta stage, something far away. Therefore I think if we create a multiplayer client/server thing soon after making the UI, it will be possible for us to play the game against each other, and test it this way. I don't really see any other ways of testing it, honestly.

I can see the downside of having to connect to the internet to play, but if we do this well it would be possible to plan your turn (we're using preplanned turns) offline, and then just hit a submit button, which will make the PC go online, post your moves, and go offline again. At the deadline you can go online shortly agian, and get the other players moves.

Besides, the internet is evolving rapidly. The game is not going mainstream for at least 2 years, which will mean that by then many people will have ADSL lines or better, and therefore be online all the time.

But my main concern here is the amount of players. I would like as many players as we can fit in to the game. And with a 1 000 000 hex map there is room for a lot of players. I fear that if we make the game single player at first, we will limit the amount of players to 10 or so, which will mean we could never have the huge, living games that I think would be one of our greatest assets.

This does not mean that we can't create AI's later on. Or even at the same time. Just that we will make the game work with a lot of players.

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[chrispie]

Hmmm...well first we had a good meeting, I had to leave in the end, my brain had ceased to function...

This really is a difficult subject, but my opinion for what it's worth is that it should be Single Player based with good but small Multi Player support. Simply cos Big Multiplayer support is a different matter altogether.

Don't get me wrong, I'd love to make a huge online only game of sorts (I've often thought of a business game, with monopolies, world trade and players forming corporations - but that's a different matter ;-)).

My view then - is 32 players in a game, and up to 32 of them being human, but not making any promises it'll work when they do that!

[The Joker]

I see why that makes sence, Chris.

But still I think it would be idiotic to limit the game to 32 players. Why not make some huge limit, design the game as we want to design it, with diplomacy, politics, economics, units and all the other stuff that we are going to have anyway, and then find out how many players it can work with and still be fun.

I know Harel said that we have to desing a completely different game for 100 players than for 10. And not following this advice I feel like telling Einstein how to do physics. I just hope that I may be telling him about quantum mechanics, then.

I just feel that maybe it would be possible to have a game, where there is room on the map for loads of civs (and we have that) and where the player is doing the thing he is doing just like he would otherwise. But where there is more players to have fun with, to go to war against and make alliances with. In most of the design we have done so far we have been copying history. Would it be impossible to do so here?

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website

[chrispie]

So I'm an idiot joker?

It's a great idea joker, but I just can't see it being workable. Think about it, 1 million hex's makes a heck of a big game, requiring hundreds of mb memory to play. If done with a limit of say 100 players, you'd need dedicated serverS with a big S. Obviously we couldn't limit it to one game of 100 or so people, and a game with 1000 or more would really be impossible. It's just too much too soon IMO. How many of us have actually even written a full released game again?

I say we save the MMP version for ggs2...

[This message has been edited by chrispie (edited December 23, 2000).]

[cavebear]

1. Is an AI really that easy to design (given the additional complexities that you are all planning) compared to allowing MP? I'm asking, only; not trying to argue about it (I simply do not know). The AI in Civ2 is pretty wretched (I find that it functions at all to be a remarkable accomplishment, but that doesn't mean it plays very well).

It seems to me that the improvements you are planning in game play are precisely the kind that an AI would have the most difficulty with. If I am understanding the direction this game is taking then strategy, negotiation, and anticipation are going to be paramount. And, if I understand how AIs work, they are essentially tactical (as opposed to strategic) in operation.

The Civ2 AI "improves" its level of play only through massive cheating, which becomes maddening to players as they get better at the game themselves. If there is a way around that, great, but I suspect there is not.

2. Is the concern with memory requirements (100+MB) a game speed concern or a user's equipment concern? In other words, is it that the computer would take a long time processing all the moves or is it that the game would take up too much room on a drive?

Being (I think) a fairly typical gameplayer/computer user, my 6 GB hard disk is only 1/2 filled and storage demand is relatively irrelevant to me. Given the GB of space on newer hard disks these days, I can't think that storage would be a real problem.

[cavebear]

quote: Originally posted by chrispie on 12-23-2000 12:41 PM As for AI, yeah it's hard to write - but whoever said programming was easy anyway? We surely can't shy away from any AI just cos we think it might be a bit hard...?

That's good to know. I was thinking that a good AI would be separate work in addition to the game itself, and that the game should have first priority.

As a gamesplayer, I truly do value a functioning AI, but I appreciate the game itself (between humans) much more.

[ElmoTheElk]

To make things clear: I vote Single Player

Single Player with MP support. Ans i mean mp support like doen in almost every other game now on the market.

MMP games seems really great; thinking about 100s of players in a game with each other, BUT i DO think that we'll have to save this for later. I think we first have to make a game that can be played by 1 persons before this player even thinks about playing the game online. I won't play a game online if i can't play it in SP mode myself.

Also, I agree that we have to use MP to test the game ourselves before we have an AI that can be used with it, but this is still possible with SP as it has support for MP games. Not MMP, but there is a way to play a game with real-life(tm) players.

At last; before there are 100 of players that will jon 1 single GGS game, we could have 1000 of SP players. I think we should build our reputation first on SP, with all the game features showing the world what we've made, before convincing the player to join a 100players-mp-game.

Someone has said here that we're not making a civ-clone, and I agree. But we ARE making a civ-like game. Those we all like and build our enthousiastic spirit on. We need GGS to be a civ-like game, rather than an MMP game without an specific user group.

[VetLegion]

well, we had quite an extensive discussion about MP along time ago. Hundreds of messages on mailing list about it actually. I dont really remember what was the consensus (or was there any). Digression: Dan Ward, are you lurking out there? Perhaps you know someting?

I ll just give a link: http://www.planetarion.com is a web based "civ like" game in space with some 50 000 players. It is run on dedicated servers. Check it out, because that is the model if you want to go massive. Massive games include serious simplification.

Now, we may have simplification via "layers" of complexity. Meaning you can for example give order to your Minister of defence, or dive in to division level (or what ever) to command the army. If each model has few levels, we can group them. So for example you may agree to play a MP game locked at level 3, which would not allow you to micromanage lower, in effect shortening the turn time and game time too. Making game simpler.

I dont want to make a game simple. As Joker said on the meeting "to have a game with complex inner working, so things develop themselves and player only gives general orders". Our game will have _lots_ of calculation and lots of complexity. To make it massively multiplayer would not mean to make it unplayable - people play those web games, but it would take much spice from it.

So as I see it, a game with up to 8 players in MP, with an excellent AI in SP. Up to 8 not because of speed of modem, but because of speed of thought. How long do you want to wait for your turn actually?

But the design issue here is how much you want to simplify the game.

Happy holidays everyone, see you in a week or two!

[chrispie]

No, it's not a h/d space problem were discussing (I just bought a 20gb h/d for £77) It's memory...my particular concern though is moving that memory over the internet for MP games. Modern computers (with swp files) can easily handle big memory games, and ones in 3 years will be even more able.

As for AI, yeah it's hard to write - but whoever said programming was easy anyway? We surely can't shy away from any AI just cos we think it might be a bit hard...?

[Sikander]

It seems unlikely that a small group of hobbyists can create both a complex game and a revolutionary AI for it in their spare time. Perhaps you guys can prove me wrong, but I think that a complicated game will require an AI which is completely revolutionary. The more complicated the game, the more cutting edge / revolutionary the AI must be in order to give the player even the same quality we all ***** about with CIV. In order to create a quality AI in all likelihood the game will have to be designed from the bottom up with the AI in mind, which will mean that most of the basic game design decisions will need to be made by the programmers.

I think this sort of game is possible, but I also think that it should be designed to be played mostly by humans, with some AI capability to handle non-player civs / tribes and human dropouts. The game will need to be simplified whether it is for human multiplayers short on time, or the AI short on brilliant programming and processor time. The idea mentioned above about creating various control levels is a good one actually, but the best time for limiting the scope of the game is right now while very few things are settled.

Who is the player pretending to be? A sort of cultural God, who guides his people through the ages? A King or Emporer who tackles the day to day struggles of the Government? A great General who leads his armies to victory? Civ has the player fulfilling all of these roles simultaneously, and by doing so makes the game unrealistic, simplistic, and unchallenging to play against the AI. By limiting the player to a more confined role, it is possible to make the game more realistic while getting rid of a lot of complexity on the player's side (especially micromanagement). It also makes the game easier to adapt to either AI play or Massive Multiplayer play.

IMO, the player should have the same level of control as the leader of the Government would have at various tech levels. Thus he orders his armies about, but does not actually command every single movement. This saves the player time, and places him more on par with any AI used. If his Generals continually frustrate his schemes, he can increase the share of the economy devoted to the military to attract more talent, or have his wise men try to develop new weapons and tactics to improve the army's performance. The player should control Grand Strategy, Diplomacy, and Resource Allocation. The built in tactical AI should handle all of the details for human as well as AI Civs. This will give the game a real possibility to be excellent for both MP and SP. If the game still retains it's massive scope for each player, then I fear that the only possibility is for it to be played multi-player only.

[The Joker]

Chris:

quote: As for AI, yeah it's hard to write - but whoever said programming was easy anyway? We surely can't shy away from any AI just cos we think it might be a bit hard...?

But can we shy away from mp just because we might think that this might be a bit hard?

I just think we shouldn't close options at this moment.


Elmo:

I can see that this makes sence. And you are propably right that mmp has to be saved for now. But what I am concerned about is to avoid that we say no to the option to have 100 players in a game, before having any idea whether it is possible or not. So yes, we can continue to develop the game as we always have. But who knows? Maybe this game will be even more fun with loads of mp players.


Vet:

I really dislike setting a max of 8 at this point. How can you know whether 8 is a good max? I personally think the game would be much, much more fun with 10 times more people.

And the whole speed of thought thing is not really an issue with the preplanned turns that we are using. Since everybody are making their turns at the same time playing a game with 8 players will be as fast/slow as playing one with 80, or 800.

Yes, 50 000 players is completely impossible in this sort of game with our complexity. But why do you think 100 players would be?


Besides, who knows what internet speeds we will have in 2 years? I am 97% sure that I will have ADSL or better by then. This means internet mp is not really different from sp. You are online all the time, so why not log on to a large mp game and play against intelligent opponents, in stead of playing against an AI?

I don't think sp should be removed from the game. But neither do I think mmp (at least to some extend) should be so. Especcially not at this point, where we have no idea how the game will be to play.


Sikander:

We have decided pretty much the same thing as you suggested, in the question of who the player is. The player will take the role of the government of the civ. Of cause to let the player continue to play for 100s or 1000s of years he can not be a person. And reducing micromanagement is something that we are doing as well. Most things should run on their own. The player will just get his taxes and spend them.

------------------
"The future is that mountain."
- Bret Easton Ellis

GGS Website