Sunday, October 11, 2009

Design Reboot: Invasive Species (Part II)

Sunday, October 11, 2009

Homelife - Scrivener's Waltz

Part II on Invasive Species is an overview of the major players and variables in the world:

The first major variable is the roughly eight or so escaped convicts: they're unpredictable, amoral bastards (just like an FPS player). Each has a certain profile, habits and goals, though most want to escape the planet as well. Every playthrough a couple of them die at random in the crash and their starting order (time at which they escape from prison ship wreckage) also randomizes, which increases replayability.

This is partly inspired by the uncompleted Black Isle Fallout 3 design (Van Buren). Chris Avellone posited that the most dangerous force in the world as other player characters and their parties. The player is always the prime mover in games--what if NPC characters were given the same volition?

The second variable is the deathgrass: the crash will scatter the spores somewhere different for each playthrough, which may hurt or hinder its chances of growing, and alter the local geopolitics considerably when it does. While the crashed ship is always in the same location, the deathgrass is not. As the playthrough progresses, the deathgrass is more and more a factor, moreso if it's not dealt with early on.
Deathgrass, contrary to it's name, is not strictly lethal. But it hurts and will impede movement. To a healthy player it's scarcely an issue, but to find oneself in a firefight deep within a field of deathgrass is asking for trouble.

Third variable is the local population. Unlike in Unreal the Nali are not a stock alien slave race/noble savage stand-in. They can and do fight the Skaarj regularly.
The Skaarj on the other hand are feral, bestial race of hunter gathers who rule the majority of the considerable jungle/forest areas in the game. They are far more difficult to gain trust/alliance for the player, but it is possible.
An alliance with the Skaarj will net basically no new technology or supplies other than basic food, but will allow safe travel through considerable portions of Skaarj back country as well as some helpful friends for certain fights.
Depending on your character's diplomacy and/or xenolinguistics, you may ingratiate yourself with either side, or be the bane of both by raiding food and supplies. Other convicts coming into contact with either side before you do first will also change things.

The final variables are the other crashed spaceships on planet aside from your prison ship, including the pursuit ship. Skiptracers are out for what they think is easy money when their ship crashes; other ships crashed and have been marooned for centuries. In them might be dormant aliens or powerful weapons, but like other factors in the game their location and presence are not reliable.

If it sounds like a too-busy free for all, good. Nothing worse than a neat game environment with nothing to do in it, or replayability reduced to nil because of strict linearity.

If you think it vaguely resembles S.T.A.L.K.E.R. if you cock your head at an angle and purge enough radiation from your system with GSC vodka, you're right. That game reaffirmed my believe that you don't sweat player choice if it's not really that tough to let them have it, all else being equal. And you can let them have a lot more choice than most, if you plan from the start to allow for it.

Invasive Species becomes about what the player wants it to be: the stated goal is to escape, but there are no forced failures if they decide to stick around. Myriad 360/Steam achievements are wired in to recognize various player-set goals: capture all other convicts to bargain for your own release, kill all other convicts, King of the Nali, King of the Skaarj, I Survived 200 Game-Days of Deathgrass spreading And All I Got Was This Achievement, and so on.

On release, Invasive Species receives a Metacritic rating of 83/100. It is roundly praised for the extensive replay value and the variability of one person's playthrough as compared to another. Gaming forums have multi-page discussion threads where players recount specific playthroughs and their attendant unexpected gameplay moments. Other players dive into their own new playthroughs in an attempt to reproduce such scenarios.
And so by request a patch is produced that allows players, upon a single successful playthrough, to start a "New Game +" that exposes a number of otherwise hidden variables for seeding the gameworld.


Davie said...

This is brilliant. Developers should build on your ideas. The whole setup seems to allow for the sort of random events and emergent gameplay that various open world games have taken a stab at, but this sounds far more replayable than most.

Johnnyburn said...

Three things remind me of the world-randomization in Sid Meier's Alpha Centauri:
1) Unknown set of NPC spawned on world
2) in random locations
3) with xenoform plant spreading.

With those Civ games in general, you'd be like, "Aw man, I got spawned next to Genghis Khan, what am I gonna do?"

I like the idea of appying this randomization to the FPS. The Civ games also had a high degree of NPC 'intentionality' that drove the gameplay. It would be interesting to see how some personality appied to each convict would or could affect the player's actions in the FPS world.

Also, starting as the last man out of the wreckage gives it some "Most Dangerous Game" feel. Which, as I recall from freshman English, is a good thing.

Another achievement: Kill 100 squirrelbats to unlock the squirrelbat patchwork fur suit skin, like a space-Daniel Boone. Complete with hat.

gauss said...

Davie: thank you for the kind words. There are a number of design issues that crop up regularly enough with open world games that we seem to have developed a suite of approaches to deal with them, unfortunately most of them boil down to making the game into a more linear, conventional narrative. Still, I think wildly configurable variables--such as the many, many multiplayer game modes possible with Rare's Goldeneye for the N64, to throw out an example--are an uniquely, intrinsically "game-y" element we ought to see more of.

John: yeah I think achievements could be used to artfully direct gameplay for some of the less than intrepid player base. If people want to experiment and do as they please, they may; but for those that require more structure the achievements could be there to provide various short term/long terms goals for where you can find the fun.

And yes, good comparison to Civilization. The variability is a big part of that game's replayability--I can configure a game size/play length/enemies etc. to my hearts content and then run it again. Outside of competitive multiplayer modes, FPS and the like seem stuck to linear progressions and are pretty stingy with these kind of options. I suspect it's because we're still more or less stuck in the authorial paradigm of other media.

(Most Dangerous Game is such a great story to teach at that age, as is Lord of the Flies. We have a lot of failures in education, but messing with kids' heads with provocative literature isn't one of them.)

Kerc said...

I don't know if I pulled it out of my arse or not but I seem to remember the whole 'NPCs being a prime mover' thing was originally one of the ideas for the game Overlord, where constantly heroes would be trying to thwart your evil deeds, I never actually played the game so I'm not sure if this was actually put into the game or if I just dreamed the idea.

It always irritated me with more recent games that there is a constant focus on the player. RPG's suffer this the most with the whole 'everyone's incapable of doing the most simplest of tasks except you' syndrome. It would be interesting to see a game as you say, where technicality there are other 'players' in the field, doing the jobs you could of done, creating the relationships you could of created. If they didn't have a head start on you.

I think a good example would be Stalker's A-Life system, and to a lesser extent Oblivion's 'radiant ai'(Much lesser) where Stalkers would travel around the zone independently, hunt, search for artifacts etc. While they do form groups it seems to be more on the basis of 'who is around that isn't hostile' rather than seeing the same group of stalkers wandering around the zone because they're used to working together. But it gets to the point where the AI complexity goes through the roof, instead of calculating simple movements and decisions, we're talking a full fledged list of variables varying from Personality traits, habits and even likes and dislikes towards such traits when regarding other Stalkers.

It would be interesting to say the least but more than likely overlooked.

gauss said...

Kerc: Thanks for the comment--I haven't played Overlord either so I couldn't tell you, but it's interesting to consider examples. This train of thought crops up every now and again.

RPGs certainly suffer from Hero of the Whole Universe syndrome and when every game is about saving the world it's hardly all that spectacular anymore--and in fact it's so much harder to get that scale of threat to seem real at all to the player I think you're almost always better off going with a smaller and more intimately scaled plot.

Also--it's 2009 and we still have RPG quest characters just standing around? Really? we can't let them sit down?

Stalker has a lot of good things going for it. Entirely convincing AI isn't one of them but then, is there any such thing? The AI are fun to fight against, and the world is lively enough to maintain immersion handily. Once while playing, the would be leader of the stalkers in the garbage area was thanking me in a dialogue sequence only to get brutally shot in the head mid-sentence. That was not scripted and was genuinely surprising.

Ironically, I think the more a game invests in "cinematic" type assets and presentation, often the less real and believable the game world becomes. It's a game and so the more things are expressed in those terms, rather than that of another medium, I think the more exciting/immersive it becomes.

Also on the subject of AI, it's difficult to tease out what we really mean when we say AI. Three fourths of the time when casual players or reviewers speak to "good AI," they're almost always talking about secondary assets that make the illusion happen. Vocal cues, animation, that sort of thing. It's an awfully smoke-and-mirrors type deal.

Andreas said...

One thing that comes to mind is how to ensure that the randomizing of the game world has some kind of personality to it: Let the player come up with a name for the Planet they want to crash on and use that name as a seed for the RNG!

Players could talk about it in the games forum like "You think Arth was hard? Play on Beteigeuze next time for a REAL challenge, N00B!"

Also nice potential for easter eggs: let "Kashyyyk" be a jungle world with massive trees, let "Dune" be a desert world etc.

Kerc said...

Gauss: It's probably one of the strengths of the Stalker series in terms of its RPG side of things, while the story is significant to you 'as the character' it could very well just be yet another tale to the other inhabitants of the zone, and the way Call of Pripyat seems to be, that's all it is. In the end (well, with the good ending) you think you've finally made it and 'saved the world' but it didn't really happen (apparently).

I don't see why RPG's feel the need to always make the main character a hero, it's even worse when you have tagged on 'moral' systems like in Mass Effect or KOTOR where regardless of how much of a complete cunt you are you're still going to save the day. Although it wasn't as bad in KOTOR because you didn't really 'save' the day, just made it so you were the one dooming it. Moral systems are another thing that just needs to fuck off for good, Karma from Fallout 2(I can't remember if the first one had it) is the only system that really worked 'well'.

Well I don't know if you remember but back when HL2 was first being shown at E3 they had a scene in Ravenholm where the player pushed a table up against the door and the Metro cop kicked the door in and was claimed that this was in fact an AI reaction to the door being stuck, and it turned out to be scripted. This is a major problem for reviewers and the like because a lot of the time they play the game through a single time and see some fancy scripting work and go 'oh wow look at the AI in this game' when the AI is dumb as a rock and only knows to run at the player and possibly shoot him. Stalker's combat AI wasn't anything to post home about but I wouldn't really go as far as to call it dumb due to the current industry standard for AI is far lower than what Stalker's is, it's only real failing I would say is more towards glitches than anything else. Although, it does tend to have a problem with navigating through buildings

gauss said...

Andreas: absolutely. The only shooter I remember having anything of the sort would be the random mission generation in Soldier of Fortune 2. It was a little undercooked, which is a pity, because that kind of thing is something other genres have done well for a very long time and yet shooters have never really explored the same territory. Fits in well with the end-user swappable content thing that seems to be so en vogue.

Oh! And there was also the game "Nosferatu", which was a shooter that had you rushing to save family members held captive by a vampire in a castle, whose layout was semi-randomized a la Diablo everytime you played. I only played the demo... terrifying.

Great points, Kerc. Stalker is a design exemplar in so many different ways it's hard not to bring it up in most any discussion about enriching shooter mechanics... and yes, the whole question of authentic AI, if what they're doing is "real" can make a big difference. In Stalker a lot of things happen that aren't necessarily all that novel or exciting in and of themselves, but their unexpected admixtures through "real" AI decisions can make for an intoxicating gameplay experience.

Charles said...

Hi everyone!

I just wanted to leave a quick post saying that your ideas are incredible gauss. Your creativity and attention to detail are amazing.

Also, everyone that's posted here has really interesting ideas. These posts, fused with the reboots themselves, make for concepts that I would really like to see in more of today's games.

The evolution from original game, to reboot, to reboot with comments is just awesome.

Anyways, great job everyone. Keep the creative juices flowin!

gauss said...

Thanks Charles--I feel the same way. I'm very fortunate to have attracted a group of likeminded, thoughtful commenters, here's hoping they stick around.

In fact, the number of interesting suggestions makes me think I might start rolling iterative design concept posts--this would be your "v.1" of the concept, and then the "v.2" post would incorporate ideas from the commenters, maybe with additional art. We'll see though, still trying to iron things out.

Valaistus said...

Got to say, I'm really interested, and I'm looking forward to it.
I've got some really important question though.

Will yuo make something for the game to work normally with multi-core processors? (I've got Core2Quad, and a lot of games are really troublesome or do not work at all)

Anonymous said...

I hope you redesign Doom 2 -- it's way too fun and simple as is :)

Dread Lord CyberSkull said...

Sounds like a lot of fun. Great point about designing from the beginning to make choices. Procedurally generated thinks like the death grass would make for an interesting change. Would you be able to watch the two sides fight it out to random outcomes or would most of that happen off screen?

Great food for thought.

gauss said...

Thanks a lot DLCS, I appreciate you going through older posts and commenting; I like to think all of the material on the site is still active.

And to answer your question, absolutely--"monster infighting," to use the Doom terminology, is one of my favorite elements in games. Even when it's mostly about you pointing a gun and shooting people in or about the cranial region, areas of gray (friend or foe ID) and factional dynamics absolutely enrich the experience.

It's tough for games to avoid the "one man army" feeling, but having enemies attack someone other than you once in a while can go surprisingly far.

Copperkat said...

I was wondering about the reward system? It's easy to scoff at lesser gamers and their need for a carrot on a string, but a lot of the time that's the driving force of a game. That's what produces the fun chemicals in our brains, and at the end of the day, that's why we play games. For fun.

In the past I've found achievements less than satisfying. Will they be the only form of progression marker/reward? Also, other than a lack of stats, how do you avoid the open world rpg grind of taking random missions and carrying them out for increasingly meaningless rewards? Granted the ultimate goal is to get off the planet, but how do you keep a player motivated without linear paths?

Copperkat said...

Oh and I hope you don't think I'm just poking holes, I'm just really interested. I've spent the better part of my free time today reading all of your posts, including the comments (which have great discussions of their own). I even added a few comments to some of your older posts. I hope you do read this, so I can formally tell you how much I'm enjoying what you're doing here. Your ideas and reboots are so refreshing.

gauss said...

Thanks Copper, and I'm glad you have been commenting on the older pieces. I keep email notification turned on so I can see if new comments I try to respond as soon as I can.

In fact, this newest set of comments has prompted me to change the site layout like I've been meaning to for a while--a "recent comments" sidebar and a notice to the effect of what I just wrote has been added. Thank you for the motivation.

Andrew Stiltman said...

I'm going to steal some of your ideas from here. (Mechanical ideas, not narrative ideas)
If that's okay with you.

Post a Comment

gausswerks: design reboot. Design by Pocket