08.14.08
The Central Concept
Yesterday I had a spontaneous impulse to write down one of my generic game concepts, thematically based on a long-defunct Deus Ex mod name of The Preachers - or rather, based on the impression I got of it from reading its website 7 years ago. Reading over the concept, it was starkly clear that in many ways it resembles half of my other game ideas. I mentioned this to Gelo, and he confirmed that he’s experienced the same thing: Before deciding to write the book he’s currently working on, Rebel Cell, he had half a dozen very similar concepts under consideration.
A few more people have since confirmed this, and it does make a lot of sense: When you have loose plans to make something creative, you’ll have a pretty good idea of the basic sort of work you want to do, but it’s easy to think of many different variations of this central concept. With that in mind, here are the three components that seem to appear in most of my game concepts:
An open world
Not necessarily a completely open streaming one-level world like eg. Oblivion, but almost all of my game ideas entail freedom of movement on the micro- as well as the macro-level. If it’s not a streaming world (Oblivion, Boiling Point), it’s a linear succession of wide-open levels (STALKER, Crysis) or a selection of multilinear levels accessed from a map screen between missions (Jedi Academy - sort of).
Problems with multiple solutions
If the previous element meant nonlinearity on a macro level, this is the nonlinearity on a micro level. I have a vague idea that it must be possible to design a game from the bottom up in a way that allows the game to react sensibly to pretty much anything the player can throw at it - to design missions in a way that permits any solution the player can think of within the fictional world of the game.
Unfortunately whenever I try to actually define this idea more specifically, to make it more concrete, I always draw a blank. Until I manage to wrap my head around the actual implementation of such a design ethos, multilinearity will have to do: specifically implementing as many different solutions to a given problem as you can think of. It’s a lot of work, but it makes for some impressively lasting player experiences.
Non-lethal gameplay
Whether through sneaking, diversion, negotiation, or social engineering, I am hot for non-combat gameplay. The question often comes up “Why would a game be more fun if you took out the combat and replaced it with dialogue trees or stealth?” and the answer is that it might not be, but it’d certainly be different, and I’m utterly convinced it could be at least as much fun as blasting your way through problems. And don’t we have enough combat-based games already?
So basically, I want to make games that I like. No surprises there, I guess 
Milton said,
August 15, 2008 at 02:36
“to design missions in a way that permits any solution the player can think of within the fictional world of the game.”
The number of solutions to a problem is dependent on the number of tools and abilities your avatar has… so huh… for example.. in deus ex did you ever encounter a situation were you though a certain solution would be plausible, but you couldn’t do it due to the lack of freedom the game provided?
I give Deus ex as an example because it has almost (there isn’t a diplomatic guy approach ) all the possible approaches to the problems presented.
Something that I think would be very interesting would be a much more reactive world/plot.. but unfortunately until the development of certain technologies (perhaps strong A.I.) that can’t be achieved. (and maybe this is what you are referring to…)
Jonas said,
August 15, 2008 at 14:53
Yeah you’re definitely right, the player’s toolset limits his options, and now that open-world design is becoming easier and more common, the main limitation for game designers is the lack of convincing character interaction - AI, as you say, but real AI is so far off it’s almost depressing. Just getting really convincing speech synthesis would be a huge step forwards as the cost of voice talent is a significant limitation.
But the problem is that right now, you’d have to design 5 different games just to allow the player the freedom I’d like: Action, stealth, negotiation, platforming, hacking… that’s a huge amount of work. I’ve been wondering if there’s a way to do for general gameplay what Havoc did for world interaction: Creating and implementing a more generalized simulation that uses some very powerful and consistent rules that can be applied to all sorts of objects and situations that would previously have to be specially coded.
I’ll be the first to admit this is both very very vague and currently completely useless and blue sky, but I’m sort of wrestling with the concept, and I’m not quite ready to give up on it yet. In fact Havok is a really good example: It’s such a powerful system (or set of systems) that it’s completely changed our expectations to world interaction - just think about Half-Life 1 compared to Half-Life 2, or compare the physics in DX with those in Invisible War. And Havok is so general, it can be applied to pretty much anything.
Lawrence said,
August 15, 2008 at 14:55
“to design missions in a way that permits any solution the player can think of within the fictional world of the game.”
I don’t know if that idea is possible, at least, not yet. It seems to me that if you wanted *every* possible solution to be viable you’d really need to have all the voice acting generated on the fly, as you could never account for all the possible choices. Of course, I guess that doesn’t really work either, as you’d have to at least write the dialog choices for the player. Unless your player never speaks, ala Gordon Freeman.
I don’t believe you can create a truly open game play environment of the type you’re talking about until someone comes up with true AI. And by the time it’s available to apply to games, we’ll probably all be slaves to our new robot masters.
The multilinearity of course is doable to whichever exent a team might choose to inflict such pain upon themselves. While it definitely makes for an excellent gameplay experience, it makes testing a nightmare. In many ways what we’ve done with TNM far surpasses the freedom afforded the player in Deus Ex, and I think you’d agree that we’re now seeing during testing just how many small logic gremlins creep into such a complex system. Getting more complicated than TNM would necessitate some sort of FAR more efficient testing process. Granted, our “process” basically consists of chaining an invinite number of monkeys (no offence to our testers, who are very intelligent monkeys!) to computers and having them play TNM until the bugs mostly disappear.
Some sort of system that works out all the permutations and outputs a flowchart of events which have to be tested might work. Of course, you’d likely end up with 78 000 different testing paths, or something ridiculous. Especially if the entire game was one cohesive construct and not divided into distinct missions. I can’t imagine the chaos that would result from TNM being one mission.
Jonas said,
August 15, 2008 at 15:31
There’s so much carry-over between the missions in TNM that it’s still pretty chaotic, but we could’ve planned it better. We could’ve made flow charts from the beginning, keeping track of the different possible permutations of the game and then checking them off one by one as they were tested. I’m sure there are ways to automate parts of the testing too, if we had uninhibited source code access.
Anyway, I’m not saying it’s possible to allow everything the player can come up with right now, but steps have already been taken to generalize certain parts of games. I’ve already described what Havok did for object interaction above, but the same can be said for open world design and player movement - instead of designing particular paths through each level as they did in DX, with an open world you can just design a whole city and let the player pick any path he wants.
Doubtlessly, the final bastion for this generalization will be realistic character interaction. It’s already possible to generalize it by simplifying it immensely - Fable and The Sims, for example, both employ iconic character interaction towards this end. Currently you just have to simplify it quite a bit more than I’d like.
Of course another approach is to simply eliminate character interaction completely, like in the Shock games.
Lawrence said,
August 15, 2008 at 16:16
Ooo, yes, let’s have only crazy monsters in our next game! They still get excellent voice acting, but it can be completely random
Jonas said,
August 15, 2008 at 16:22
We have a term for that though. To “chicken out”.
Lawrence said,
August 15, 2008 at 16:48
Are you saying BioShock chickened out?
I think not!
Jonas said,
August 15, 2008 at 17:00
“We can’t do 100% realistic character interaction, so let’s not do it at all.”
I think so
Milton said,
August 16, 2008 at 02:04
What concerns me is that even with the development of technology (since Deus ex/fallout debut) the industry seems to be stagnated in therms of interaction complexity.
While every year graphics get better, the same cannot be said about characters reactions and freedom of choice in the storyline. Fallout multiple ways of solving problems and karma system are pretty much the limit and even games like KOTOR and Mass effect failed to make any significant advance (the games are great anyway) the problem is that this type of game requires a enormous team and thus the percentage of games with these features is much lower than what I would like.
Some kind of automated system would be nice, these types of games would be easier to make and consequently improve the quality. The good/bad/neutral guy system is kind of becoming uninteresting, but to find a more multifaceted viable alternative… hmm… well I guess that would be a million dollar idea
Jonas said,
August 16, 2008 at 02:55
Well, look at Oblivion: In my opinion it’s not better than games in the Black Isle / Bioware vein (Fallout, Torment, KOTOR, Jade Empire, Mass Effect), but Oblivion is at least an honest attempt at implementing a different type of nonlinearity - one that’s based around “spheres” of plot that don’t influence each other.
How you handle one mission in Oblivion doesn’t really affect how other missions can play out, and when missions are strung together, the strings are pretty much entirely linear. That allows you to make each mission as complex as you want because each mission is self-contained rather than part of a huge plot system that can get out of hand very very quickly.
I don’t like that solution though. It makes your achievements feel worthless and your choices seem inconsequential (because, well, they ARE inconsequential). I prefer the complexity and the codependence because it makes for deeper and more satisfying experiences.
Again, for an actual “automated” system (sort of), look at Fable and keep a really close eye on Fable 2. Since all the character interaction was iconic a’la The Sims, every choice you made could have a consequence because the consequences were so simple: If you’re evil, you get Evil points and people fear you - if you’re good, you get Good points and people love you. It’s an uncompromising generalization of character interaction, but it has a similar significant drawback to Oblivion’s approach: You’ve simplified it so much, it becomes shallow and predictable. The consequences of your choices in Fable will never surprise you. Maybe they’ve improved that in Fable 2. We’ll see some time around Christmas, probably
Jonas said,
August 16, 2008 at 03:03
Oh and a bit of clarification: The best solution right now is to string different generalized systems together. Combine an open-world game with a robust physics engine, and you already have quite a compelling game, as Mercenaries 2 and Far Cry 2 will both be keen to demonstrate this fall. That’s a LOT of freedom for the player right there. Character interaction is where the approach currently stalls a bit, because there’s no way to generalize that - it all has to be prescripted.
I don’t think that’s likely to change soon. Milton, you hit the nail right on the head when you called it “a million dollar idea” - that’s exactly what it is, and that makes it worth keeping in mind and mulling over when I have nothing better to direct my thoughts at. I’m not Will Wright, so I doubt I’ll ever actually figure it out, but maybe one day somebody else will have part of the idea and cause something to click into place in my head, or maybe I will have part of the idea and somebody else will grab hold of it and add the missing pieces.
Until then, this blog is the best place to discuss it, because nobody here will roll their eyes at the uselessness of such a vague and unfinished idea.