Dynamic world framework

[ConsumerHans 1]

I know the title sounds awful - I'm after some real world feeback about an idea I have had in my head for about a decade. I have to say I'm not a game dev, I'm a consumer.

There are game engines - Frostbite, Unreal, to name some that I have heard of myself. They add the physics, AI, rendering, collision and some other aspects of the game. What they don't seem do is to simulate the world that the game actually plays in. Therefore more often than not, this world seems to be simulated by scripted events (missions, waypoints, etc.) or in a more "Open World" kind of approach that let's you do whatever you like, but often without any real consequences.

I was thinking about a framework that simulates, on an abstract level, an artificial dynamic world, and that could be connected to a game engine by choice via API/interface.

I'll try an example:

There's a table with a suitcase full of guns. It is rendered by the game engine. I can kick the table and the suitcase will fall down. But if I pick up the suitcase and go out on the street and happen to run into a police control, I'll either get arrested or, in case of resistance, shot at. So on an abstract level, there's an item A that triggers a response from a group B based on my action C. I know that this can be done by scripting, but that's not what I'm interested in. I'm looking at a dynamic framework layer that a game dev could connect to a game/physics engine in order to create a game.

Is this understandable? If so, is this in any way doable or marketable, from a professional's point of view?

Thanks, Michael

 

[Kylotan 10351]

Understandable? Yes.

Doable? Definitely, if and only if you were a lot more specific about what exactly it would be doing. "Item A that triggers a response from a group B" is so vague as to encompass pretty much everything, so really you'd be looking at codifying a set of rules and ways to express the operation of those rules. Even that itself is still far too vague, so there's a lot of detail lacking here.

Marketable? Unlikely, because game design is largely about choosing which rules your world will operate under, and whatever rules you chose may not be the rules other games will want.

[ConsumerHans 1]

Well it needs to be kind of abstract I suppose, because it is an abstract layer that comes in *before* anything is being rendered.

3 minutes ago, Kylotan said:

Marketable? Unlikely, because game design is largely about choosing which rules your world will operate under, and whatever rules you chose may not be the rules other games will want.

Isn't that what a framework does? At least in my own experience. I'm a web dev, but because we all mostly use just a few different popular frameworks, doesn't mean that all the websites or web apps are the same.

So the liberty would be to create and control rules and interactions in the virtual world with the help of this framework, which would be something extremely time consuming if done by every game for itself. I mean, that's the same basic thought behind selling/licensing game engines - to create different games.

[Kylotan 10351]

Web frameworks are quite clearly defined. HTTP requests come in on URIs, they get mapped to handlers, auth gets checked, a bit of custom logic is run, then data typically gets piped into templates which are then sent as HTTP responses. Because most of these stages are highly constrained it's very clear what the input and output will be, with variations mostly being down to taste (e.g. different templating languages, different ways of routing URIs, different ORMs you might plug in).

A world-simulating framework is far more abstract than that. It has no clear concept of a request or a response, just "something happens" and "something else therefore happens". On this trivial level, game engines already give you this functionality - providing you can give it a concrete definition of the first condition, it can emit an event, and you can attach the concrete definition of the consequences. For example, Unreal can tell you when an actor takes damage, and you can hook your "alert the guards" code there. But you have to write that code.

You say you don't want this - you want it to be more than 'just scripting'. That presumes you want some sort of layer to do this sort of thinking for you, but there is no standard way to do that, nor any clear definition of what it would involve.

The reason we have graphics and physics engines is the same reason web developers have frameworks - there are standard inputs and outputs and clear expectations for how you get from one end to the other, with some scope for custom logic in the middle. To do the same for in-world events, you'd need to devise a mental framework for what form that should take, and only then could you consider a code-based one.

 

[Dramolion 19]

I suspect the popularity of the engines you mentioned has led to a flood of unfinished games as well.

Are you looking to combine other people's unfinished work?

[ConsumerHans 1]

On 16.10.2017 at 6:28 PM, Kylotan said:

To do the same for in-world events, you'd need to devise a mental framework for what form that should take, and only then could you consider a code-based one.

 

Thinking about it, the term "Dynamic World" might be aiming at something too complex. So let me call it "Dynamic Campaign Framework". I think my initial example was a bad one and therefore I was maybe misunderstood in what I am looking for. Let me try it again, I'll keep it as short as possible.

In a traditional RPG, the player get's thrown into a world under some pretense (story), the goal usually being of eliminating the ultimate Orc overlord or something similar. The simplicity of this logic is obfuscated by a more or less complex story. You then collect weapons/armor, level up, and ultimately become strong enough to face the different challenges all by yourself. When you have ultimately triggered all the necessary events, the game is won - hurrah. This is basically where game development is today and has been for 20 years.

In a dynamic campaign, Orcs and Humans will just do what they do - battle each other. They won't just stand around, waiting to be triggered. Whether the player engages in the campaign himself, doesn't really matter. One could just stand on a hill and watch the war going on. The outcome at this point would be pure RNG. So the player's goal would be to support his faction in winning the war.

This could be done in several, semi-realistic ways:

• He could sabotage the enemie's supply lines
• He could join the next assault, hoping to slay an Orc or two
• He could help with the manufacturing of (advanced) weapons for this faction
• He could try to find and pay a traitor for information
• etc.

So the challenge for a dynamic framework would be to simulate two faction at war. There are certain parameters I could think of:

• Supply. An army needs a constant flow of goods, which is also one of it's weak points.
• Assault and defense. When a certain amount of supply or troop strength is reached, the faction will carry out an assault. If it is successful, that part of the map is taken. If it fails, the faction will retreat, waiting to re-supply.
• There could be parameters like loyalty or morale to influence the combat strength of a faction.

The immersion in such a scenario would come from the fact that a player cannot win the war all by himself as usual, which is silly and played out anyway. Instead, all of his actions would have a real effect on the dynamic campaign.

There was only one game I could think of that did this, Falcon 4.

Because it was so difficult to learn and rather sterile in it's looks, it might not have appealed to many gamers. But those who played it know that the level of immersion was hardly ever reached by another game. User InSight nails it pretty well in his comment on GOG: "After all these years, no other sim has come close to the immersion of being a fighter pilot in an ever-changing dynamic campaign, where you see this coordination of different flights and ground forces trying to achieve victory." What it established should have served as a template for all other types of games - be it a military simulator like ARMA, an FPS or an RPG. Unfortunately, that hasn't happended. Instead, we play the same stuff as always, but with better graphics.

The complexity of such a dynamic campaign led me to believe that it might be better to create a framework, so no single company would have to do the job all over again each time they want to create a dynamic campaign game. Orcs could be replaced by T-80's or Reapers and could be rendered by the CryEngine or Unreal, it doesn't really matter.

That's all.

 

 

 

Edited November 7, 2017 by ConsumerHans

[ConsumerHans 1]

On 16.10.2017 at 6:32 PM, Dramolion said:

I suspect the popularity of the engines you mentioned has led to a flood of unfinished games as well.

Are you looking to combine other people's unfinished work?

Not sure what you mean by that. But I think I'm looking for something new here that couldn't be established by combining existing work.

[Dramolion 19]

I think  i understand what you mean.

Games like the Witcher have (attempted to?) make something like that, dunno how it worked out.

There's two problems with your idea:

1) Players don't like too-strong luck-factor, and if the luck factor becomes smaller the game becomes more scripted again.

(err, solution, make events random where they happen, diminish luck in the grand scheme of things of which faction is winning)

2)The most dynamic worlds are multiplayer, and your idea seems single player orientated while trying to give the same experience that already exists on multi-player servers.

[ConsumerHans 1]

I think games like Witcher 1-3 have tried their best to come up with something that feels alive and moving, other than let's say the RPGs by Bethesda, which are loved by many and described as kind of empty and generic by at least as many. But as far as I know, they don't have a dynamic campaign, it's all just a bunch of scripted events.

1) Not sure if this can be generalized. XCOM has pure RNG and is still very popular. Also, it's only pure RNG if you decide not to do anything, which is not the point of a dynamic campaign.

2) That's a good point. I haven't had too many multiplayer experiences besides Battlefield, so I can't really say how static or dynamic their worlds/campaigns are. I tried Warframe recently and was pretty disappointed in that regard.

 

 

[Dramolion 19]

On 16-10-2017 at 5:22 PM, ConsumerHans said:

If so, is this in any way doable or marketable, from a professional's point of view?

Anyway, i think i kinda understand what you're trying to make, and, err;

Doable, sure, anything is, but it'd be more way more work because you need more content(since the player might miss some) and make that content interact with each other.

you also have the restriction that you can't have a too big world being alive the whole time on a single computer.

Balancing luck and difficulty is going to be difficult as well.

Marketability, well, i suspect there 's plenty of players waiting for a dynamic single player world, but they 're playing other games which, together, fill in this niche enough to keep them satisfied.(and enough broken promises to make them suspicious)

Companies looking to use such an engine(?) would probably not be interested until a game using that engine is successfull.