They couldn’t really interact with each other beyond standing around and repeating the same chat scripts, they didn’t trade with each other or get into fights over petty disagreements, or hatch plots to kidnap someone’s daughter for ransom and then attempt to carry those plots out.

And I go further: If you want the feeling of an impossibly complex world where you’re largely powerless to affect your fate against the whims of uncountable other actors, that’s actually been packed in for free with your wetware.

True that what I am suggesting would be a major undertaking and it may not be a very realistic expectation of any game in the near future, but it is not anywhere near as outrageous as you make it out.

@Jackohbite – Well, I agree that it is probably not likely to happen anytime soon, and it is pretty tough to debate the ‘impossible’ argument without actually doing it. In general the most effective argument against ‘impossible’ is to just do it and then say “see, I told you so”.. hehe. But I admit it is beyond me personally as an individual at this time.

I do have to wonder though. I think that people (including most developers) accepted that it was not possible back in the 90s when we where playing doom on 386s. It was still not possible a few years later on our 300mhz pentiums with 256mb memory. But today I am running a quad core 3000 mhz cpu, 8gigs of high performance memory, and a crazy powerful video card with 2 gigs of dedicated high performance video memory.

Over the years it has just become accepted that we don’t try to go there, and once something like that is commonly accepted throughout an industry it can be a long time after the capability is present that anyone actually breaks from traditional logic and leverages those capabilities (and in this case it is most likely to succeed as a major group effort, which of course is even less likely than a few individual progressives making any such efforts)..

Look back at a game like magic carpet (huge open world 3d game, with dynamic lighting, real time deformable terrain, massively destructible environments, real time physics including gravity and fire which could spread and reduce forests to ashes. This game came out in the mid 90s and ran on 100mhz CPUs with 32mb memory. Of course it didn’t use hardware rendering, and characters where actually 2d sprites, but then again this kind of fire was in a game well over a decade ago, yet today is lauded as a revolutionary leap forward when seen in far cry 2.

Fire that spreads and after a short time causes a leafy tree to be replaced with a charred leafless one – it is a simple thing to make work, and the truth is that this sort of thing should have been the norm a decade ago, but is a standout feature today instead. And even now, other games don’t raise the bar to that level. NOLF had respectably well done reflections of your character in mirrors, AvP (original) also had this feature only it looked better. AvP 2 which used the NOLF engine had mirrors – but no reflections in them, this was a successor to 2 great games, but rather than adding to the pallete – it subtracted from it.

We probably won’t see many other games come out with spreadable fire for a long time yet to come, not because it’s impossible to calculate the physics of it, or even very difficult. But rather because they can sell us games just fine without going to the negligible trouble of making spreadable fire.

Then if you check out the Men of War games, and consider that most of what is happening on screen in some of the major battles is not scripted at all (there is some scripting but even then most of the scripting is just to send units into the fray – 99% of what is happening on the battlefields in those games are not scripted and purely dynamic.

Also look at what nVidia has been doing with Tesla – we are talking about exponential leaps forward in processing power. To bad they are pushing this tech further and further away from consumer level – they started out marketing to a level which was tantalizingly close to consumer affordable. But at the same time, even as enterprise level tech it opens up new possibilities. A game like this could be packaged with a piece of dedicated hardware which would handle a great deal of the processing overhead. That would raise the cost of the game but with the right licensing deals and decreasing cost of processing power, it might not need to raise the cost all that much.

The other possibility could be the massive online multiplayer approach. You could calculate the AIs in a distributed manner on a powerful server farm and clients would just see the results of those calculations…

So it may not be probable in the near future but I don’t think it is all that far fetched, and while it would be a major undertaking, I think the main reason it won’t be done anytime soon is not because it can’t be, but rather because it would be risky to try…

So really, what I’m saying is I think you’re right, that would be an amazing, open game world. But it’s not possible, at all, and I’m much more interested in what is possible, because I get to play it.

Tangentially, do you really want to shoot something that can generate interesting human behaviour on its own? Would ‘it’ still be ‘something’?

Even if Oblivion had tried and it didn’t work out doesn’t necessarily mean it can’t work – just need to do a better job. In some ways what I describe may be even easier to make work and process then what Oblivion was attempting. Oblivion was trying to run hundreds of scripts, mostly unique, whereas what I described would just be random reactions but the percentage chance of given reactions would be influenced heavily by their attribute trees on the fly.

Of course, as you mentioned, the play mechanics are of great importance, but one of the things that almost always bothers me most in almost every open-world game is the shallowness of the world, and the sense that I am the only one doing anything. NPCs that sleep, eat, and walk around is not the answer. They are still not doing anything of any consequence in the world, unless it is something specially pre-scripted for my benefit and triggered by me.

When the people you are shooting are not just cardboard cutouts then it would give shooting people in a MiG so much more weight. More I think even then shooting other players in multiplayer, who may not be cardboard cutout NPCs, but who will just respawn and be back for more.

Some mini/meta territory games which to me seem like they would only further distance me from deep immersion.

I think the key to the most awesome open world game ever is to make it truly open world. Make randomly generated world, with randomly generated NPCs whose random attributes go beyond simple measures of physical and mental capabilities. So NPCs get a randomly generated, fairly deep attribute tree outlining their likes and dislikes, their temperaments in different moods, their minor issues and hang ups, as well as the occasional full blown neurosis. Give them jobs and families. Make their interactions with the world be determined in large part based upon their unique attribute trees. And make those attributes trees also be effected by the world and their continuing experiences. So different NPCs who are hungry will take very different approaches to handle the situation. One may start stealing, another may look for a job, another might start scrounging for goods which he can trade – and each of these courses will have repercussions. The mugger may kill a victim who fights back, and now he is a murderer and a family’s father won’t come home from work because he is dead and the family will get worried and call police, they will miss him and be sad which will effect how they react to situations (or maybe they won’t miss him – one might be angry that he’s dead, another might be happy depending on what he was like and what they are like), and eventually they’ll get hungry without his paychecks. The scrounger/trader may end up a dumpster diving hobo, or he may only start out that way and end up a successful merchant with his own shop depending in part on his own attribute tree but also in part on luck as the situations he encounters could have major impacts on the direction of his life.

The biggest problems with this are the processing power required to calculate all those NPCs lives and communications both between NPCs and between the player and NPCs. It would be great if it was hard to tell who was another real player and who was just an NPC. With a fantastic communication interface this blurring the line between NPC and Player could be possible.

4d if i remember correctly is time, and if a game mechanic could be worked out for this, it could show a number of possible future action paths which are relevant to the “present” or whenever you traversed to 4d. or you could subtly influence things from 4d, mild telekinesis, telepathy, at higher levels pyrokinesis and the ability to spontaneously create organised patterns of matter.. using these you could encourage certain action paths, create new action paths, or just torture the crap out of all the ncp’s :D.

each dimension would give you various powers and attacks would be few, but would work differently in different dimensions

1d, there would be a line on your screen and thats it.

this would be an awesome game if you were a hyper-dimensional entity, who could unlock powers allowing access reality more concretely (this being the objective, ultimately to manifest yourself within that reality (negatively of positively or powerfully etc). imagine being able to incite a riot in a crowd by projecting image( eg their president) with connotations (negative or positive to varying degrees) directly into their brains. then you use pyrokinesis to set a few cars on fire, the riot police start to overreact with you projecting emotion (anger, fear, duteous). next you use your newly unlocked ability to manipulate matter to start a fight between two coworkers in the office, its escalates and ends up with the 68 floor office building demolishing like 10 blocks of city.

more ideas plz!

The idea would be that you have up to four agencies, each with 1 or 2 players(too many, and it becomes deathmatch with extreme lag) vying for control of the island, all against both each other and the AI president, kinda like story mode. The first team to hold power for at least five minutes without the capitol and surrounding territories getting captured by the opposing teams wins.

It plays almost the same as story mode, just simplified and accelerated. Collectibles are out, missions are all opportunistic(which allows for multiple ways to start it(like ERglabs mentioned), but with greater rewards, and even the regular opportunities hold great prizes. Everything happens in real time, but fast travel is removed to prevent people from using it when they rightfully deserve to be killed.

If you want, you can just stay in the safety of your tent, upgrade your troops to take all the territories/find weapon caches/etc., and pray to god they dont find your camp and kill you. Or, you could be the hands on guy. Go in guns blazing, upgrade everything you’ve got, and team up with local revolutionaries to hunt down the other players.

However, when something like the president fleeing, or an enemy officer patrolling occurs, everyone is informed and gets a chance to do the deed themselves. For example, in mid chase, one team could ttake out two of the convoys, but another would snipe the president from far off, only for another to install theirs. This can create near-creamblast feelings when you underhandedly win at such a situation, when the other teams are too far away from you to stop the timer before it runs down.

And yes, this is near impossible to balance properly, but would sell millions if it were.

One thing I think that would seriously benefit open-world games, yet is rare, is designing missions that have multiple start points. ie. ways to stumble into a mission on your own without being given a missive from some slack-jaw in the nearest town. Open worlds have an inherent call to explore, and few things render exploration tedious like “Oh, wow, I found this totally cool… oh wait. this must be a mission I don’t have yet.” (I’m looking at you, Bethesda) Then you have to traipse back to town to find the person that will lead you by the nose to spot you felt so awesome for finding earlier.

Obviously, it couldn’t work for every mission, but Fallout 1 & 2 did a fantastic job of creating dynamic missions that could be triggered a dozen ways other than finding that key NPC who wants you to go do x.

The distraught mother who wants you to go find her kidnapped daughter, obvious starting point.BUT, if you had stumbled upon the bandits shack first it shouldn’t be empty, it should have the daughter regardless, who would then ask you to help her find her mother. Or a discarded piece of clothing on a hidden trail would add the mission, etc…

telling little stories of other players and their failures.

I think one thing that game developers have trouble with is a conflict between:
1. Letting the player whittle down the opposition (whether it be stages of a boss battle or the game in general)
2. Wanting to constantly ramp up the challenge.
I really love the idea of sabotage progressively weakening the army, reducing the amount of weapons, etc. that they can use against you, but how to do this without making the game get easier as you go?

Great stuff, Tom. This is the best game I’ve played in my head since I read your alternate ending to Bioshock.

Plumberduck, approved.

CheeseLord: “Imagine a game such as GTA4 with it’s missions divided as you described, the choice between missions would no longer be governed by interest in the story of specific characters but instead be governed by a preference for certain chunks of the gameplay.”

Brilliant! The choice would no longer be between four characters I hate equally, but between four mission types – some of which I like. Hmm, do I want to do another job for the Painful Escort Mission guy, the Tedious Third Person Shooter guy, Bullshit Scripted Chase Man, or the Exciting Getaway Driving guy?

The themes for these mission threads are broader, obviously, and they’re not supposed to be side-mission type jobs: each objective would have something to do with the last, and a complication that makes it more than a simple hit. The Assassination missions could potentially be as varied as Hitman: Blood Money’s, more so with vehicles. The Heroics ones could chain together, so that three jobs are all preparation for a spectacular fourth, like the final few missions of Oblivion’s Thieves Guild. And the Sabotage mission thread could basically be the entire game of Mercenaries 2. I’m not saying it’s easy or cheap to do all that, just that the structure itself doesn’t have to limit variety or story.

Can’t really respond to your other point because I don’t see what it’s referring to – there’s no mini-map icon to save a ‘random character’ in this. I do want gangs fighting: drug cartels, government and rebels. But I don’t want the survivors to become your bestest friends forever if you help out.

I think a suitable compromise is a the mission structure of Red Faction: Guerrilla; it describes your objectives and gives you the option of accepting or rejecting your mission, rather than just dumping you at the start of the mission.

Overall I think the game you describe is too linear in it’s approach; when I have the option to save a random character, for example, I don’t want to see a little icon appear on my mini-map giving me the choice, as it is in Assasin’s Creed. It should be a more fluid occurence, such as in STALKER: Shadow of Chernobyl where if you find two AI’s in a fight and kill one of them, the survivor, if of a neutral faction, will show gratitude and become an ally.
True “Opportunity Missions” should be a randomly generated occurence, not some bullshit side-mission.

WHY?

One simple thing is the option to un-liberate an area. An example of what this would be like is when you reach that point in Sim City where it’s almost self-sustaining and good. BORING. So you nuke it, send a tornado its way, or just demolish large portions. Undoing your progress gives you more content, provided that it has a unique enough experience to make it not repetitive.

None of this Hot Coffee, Simon Says idiocy. We’re talking full mouse or analog stick stimulation. (Uh… stimulation via analog stick, not analog stimulation of any sticks. We don’t want to ruin anybody’s values. This isn’t Bioware).

Admittedly, developing the kind of system that would really be a justified reward for players is going to mean having to cut back on some other areas in the game. You probably won’t need vehicles; who would want to drive away from all these sweet honeys anyway, right?

And guns? Come on, if you’re playing Just Cause 2: Sex Wars, you’re more of a lover than a fighter. On that same note, maybe start all areas as permanently Liberated? Or, you know what, way simpler to just have one area. This might get in the way of the whole “open world” thing, but really, isn’t a woman a world? And these ladies are certainly open. To love!

I like the idea of collectibles, though. One thought: instead of dead people, how about sexy underwear? I guess you could put the sexy underwear on the dead people. It’s your game, Tom, I’m just throwing out ideas to make it even better.

Brit, an early draft of this post spent a lot of words on how to make different regions of the world distinct both visually and functionally. Then I actually played Just Cause 2 and thought ‘Oh, they’ve pretty much got that covered.’

Verde, couldn’t agree more about the horrible defense alerts in Spore, and the same thing damn nearly ruined Dark Crusade’s otherwise awesome campaign map. In both games it was irritating because your territories couldn’t defend themselves, but you could, so it felt like babysitting. I think the solution is to divorce the player’s in-person power from his strategic power, so that even if he shows up, he can’t do it all himself. Otherwise the whole strategic game is reduced to how much you can be bothered to rush around tending to your territories.

I think I’d have it so when the government invade one of your territories with enough Officers to win, it’s gone right away – you’re not even told until you next check the strategic map or go there. And you can’t move a new Revolutionary there to take it back until you next get one. So getting a Revolutionary is what gives you a new ‘turn’ on the strategic map: you make all the moves you want to, then you’re powerless to react to the government’s response until you get another. Perhaps when you go to the strategic map to move some Revolutionaries, you’re warned that when you’re done, the ones you haven’t moved will be told to entrench themselves, and won’t be in a position to invade for a while.

I’m getting increasingly tempted to try making a mock-up of this strategic map in Game Maker or something, to see if it actually works.

I think if I explained this design to a real game designer, like our close personal friend Clint Hocking, the first objection they’d have is that Liberation empties the world of content: those areas no longer have combat in them, so they’ll feel vacant. I think the challenge there is not to find an excuse to have lots of combat in Liberated areas, because then they’d feel the same as Martial Law and regular ones, but instead to give them a different kind of play. Fighting is never my favourite bit of an open world game anyway. Searching for collectibles and convoys would work, but it probably needs more.

Bret, yeah, I wouldn’t want the default ending to feel like a smack in the face – it should be handled as a mega victory, and the hints at the new leader’s faults could almost be read as nothing more than black humour. If the player has doubts about the virtue of what he’s done, I’d want them to be his own and not the game saying “YOU DONE WRONG!”. Did you ever do the Dark Side ending in Jedi Knight? The game basically punched you in the face for it.

The reason there are two ways to get the secret epilogue ending, by the way, is that I want ‘free will’ type players to be able to avoid doing a bad thing in the first place, but I also want regular gamers – who’ve learned to just do what the game tells them even if it seems to suck – to have a path to redemption.

Oddly, the one thing that sounds trickiest? The endings. I mean, if the default feels too much “Meet the new boss. Same as the old boss.” then it’d feel more like a sequel ad, and the worst kind, than an actual conclusion. On the other hand, ignore the whole the new guy is an asshole thread, and it’ll feel like something interesting is being ignored entirely. I suppose it says a lot of good things that there’s a whole game described with all sorts of wild ambitions, and the closest thing to a flaw I picked up is a minor worry about a half mentioned extra ending.

Man, if the real game is half as nifty as this design, I’ll need to get it eventually.

While I completely agree with about everything that you say, there are a few caveats:

For the idea of Liberation, if there are resistant areas fighting back, it shouldn’t be calling you back to your areas repeatedly to save it. A good example of a bad example is Spore, how it would CONSTANTLY call you back to save one of your territories from Eco-Disaster, or invaders, or pirates. It was a HUGE deterrent from exploring the open universe.

And on the subject of Spore, which despite my rants I thoroughly enjoyed, it constantly forced you to refuel. I suppose that it did force you to meet new civilizations, but it really got in the way.

Still trying to digest the mission threads and their consequences. I think there’s some room for improvement, but I can’t put my finger on it yet.

Whole article is great food for thought. Pentadact is credit to team.

Honestly, I can’t think of much else that I’d want in an open world game, this is so well done. I also really like your idea for the plot, with the ability to turn against the dark past you’re being forced down at the end, I really hate it when games like GTA don’t let me be anything but a bastard. Being good should be harder, but with a rewarding sense of achievement, exactly like the epilogue mission you describe here.

Anyway, it was a great read and I’m sure it would work well for a game like Just Cause. It’s a good mix between being game like and having something that adds to the atmosphere. I have to say though, Fallout 3 came pretty close to my idea of an ideal open world game, mods helped it along, but it was definitely going along the right tracks. Even though it stumbled at every hurdle it still did a really good job at it overall.

I just prefer smaller-scale stuff (such as GTA) where you aren’t necessarily the super-top agent fighting against hordes of enemies. That’s my subjective problem, of course – otherwise, from what I learned by skimming through the topics, it could be a fun game. But I’d like you to try improving GTAesque games as well.

I’d be a much better word boffin if i could actually spell :D

Great read :)