Page 1 of 4

### Describe a game you think would be awesome

Posted: Thu Dec 11, 2008 6:33 pm UTC
Imagine you had millions of dollars and a team of developers and artists. What would you tell them to develop?

My first idea is a space-based MMO Flight Sim/FPS/Econ Sim. Spoiler'd for wall of nerdy text.
Spoiler:
I want to break physics and astronomy as little as possible in this game. The whole thing will take place in a single star system (possibly Sol, possibly not) because there is no faster than light travel, though coming close is possible. Travel works the way (i think) space travel would actually work where using the engines is only necessary to accelerate and decelerate. The energy required to accelerate increases the faster one goes (because mass increases), until eventually you reach a point where your propulsion systems cannot put out enough force to accelerate any faster. For craft with large engines and little mass this could be 99.99% of light speed, for larger craft that could be 50%.

Obviously accelerating fast enough to reach 50% of the speed of light in less than a year is going to generate enough G-force to kill anyone in the spacecraft. To overcome this i have created a fictional device called the Inertia Controller. With the IC running anyone within the spacecraft experiences a sense that the spacecraft is accelerating straight up at approximately 10m/s2. IC disruptors, similar to EVE's warp scramblers, can damage the field produced by the IC, reducing it's power output. It can still produce artificial gravity, but the spacecraft is limited in acceleration speed.

I mentioned above that this is a flight sim. It doesn't use menu-driven commands like EVE, but instead spacecraft are actually flown by the pilot. Spacecraft are not single objects the way they are in many games, but are instead an enclosed map on which players walk, sit, and operate the spacecraft and its equipment. Not only does this give you something to do on long trips (some objects in this Solar System are over 100 AU from sol, meaning a trip could conceivable be as far as some 200 AU. At the speed of light that trip would take over a day.) but it adds elements of wreck survival, engineering, locational damage, and boarding teams. For example, a missile impact may breach the hull in a certain area. Anyone in that area would be sucked out into the black and die, and the craft's commander must seal that area off or risk a life-support failure.

Unlike EVE, a single person cannot operate an entire spacecraft, at least not a big one. Effectively using all turrets, remote drones, remote missiles, electronic warfare tools, and lord knows what else will require additional personnel to man them. Think ship-to-ship combat in "Pirates of the Carribean Online". Because smaller spacecraft may not have the fuel storage to travel long distances carriers may become popular, too. Also, in order to keep everything working on a large craft engineers (also players) will be needed. They can fix equipment damaged during combat, and optimize equipment for tasks at hand on the fly. They can also help equipment work better than normally possible.

Operators are players who don environmental battle-suits and small arms and force their way onto another spacecraft to take it. They can do this through open hangar bays or by cutting their way through the hull (which might take a while). They will have to battle remote controlled drones, any turrets on the craft that can hit them, and other operators on the victim spacecraft. If successful, however, they will have a free craft. They are able to reach other spacecraft either by jetpacks or by their boarding craft ramming it. They stay attached to the hull with electro-magnetic boots. This is the FPS section i mentioned earlier, though i'm toying with making it a third-person shooter.

The basics of combat and travel having been covered, it's time to get into the economics. There are a lot of asteroids out there. They are mined by landing on them and using mining equipment mounted on the spacecraft or hand carried by the miner. Or both. This creates an opportunity for piracy without even having your own ship, just several operators hiding amongst the asteroids. Players can also land on planets and moons (the map size probably won't be as big as actual planets and moons) to harvest resources or set up processing/production facilities (as compared to space stations, which are also a possibility). The metallic cores in gas giants are also available, though the environment there requires the highest quality environmental suits.

This seems like a good time to break and mention monsters. Obviously if you're on a planet or moon in the goldilocks zone you're going to have to deal with lions, tigers, bears, and xenomorphs. If you're on a planet outside of that you just have the xenomorphs to contend with. Dangerous atmospheric beasts will occupy gas giants, and even in space monsters can be found floating about Zerg style. In addition to simply attacking players (especially miners) the young may crawl into spacecraft and hide out in maintenance shafts, potentially resulting in the first "Alien" movie.

I have utterly failed to think of a skill system that i think would complement this well, and am thinking about not having one. You can use anything you can afford to buy, basically.

Idea the second: comedy RPG
Spoiler:
Opening cinematic: A shot of a beautiful castle in some manner of magic kingdom. Cut to a man lying facedown in bed wearing nothing but torn and stained tightey-whiteys. After a moment we cut to a magnificent golden dragon glittering in the sun and flying above the streets in a bustling city. Skyscrapers reach out above it and below cars move about. Cut back to the man sitting on the side of his bed now. After a moment he gets up and groggily shuffles off camera. We now see a marketplace where two men are competing for business. One is wearing sorceror's robes and causing puffs of smoke and conjurations, the other wears an old-tyme lab-coat with goggles around his neck as he directs various tic-toc machines. Two people in medieval armor direct the crowd to keep the peace. Back to the man, he is in the kitchen and has poured milk and egg yolks into a mixing bowl. He reaches into a cupboard and produces a box of instant pancake mix. As he begins pouring a puff of dust leaves, then nothing. The man attempts to force more out with increasing urgency but it quickly becomes clear that the box is exhausted. He casts it to the ground and reaches to the sky yelling "NOOOOOO". As he does the screen fades out as the title fades in

Pancake Quest: The Most Important Meal of the Day

After this we are given control of our hero and directions to get him dressed and march him to the general store for more pancake mix. As you walk down the street you are periodically passed by patrol cars with their sirens screaming and knights on horseback, also with sirens. After a short ways you pass them again, now fleeing in the other direction. When you round a corner you see the general store just in time to see a man with cartoon-style TNT strapped to him. He immediately detonates it, taking the general store with him. Our hero rushes forward, hoping to find some pancake mix that survived the blast. Instead he finds a man who had been facing sideways from the explosion so that one half of his body is severely burned and the other is normal. Our hero digs a bandaid up from the wreckage and applies it to his new friend's cheek. The man somberly thanks our hero and pledges vengeance upon the people responsible for attacking the general store, asking if he may follow our hero on his quest. Also being exploded has given the guy powers over fire for some reason. Our hero cares nothing for this man's vengeance quest, but he doesn't want to go home and he doesn't know how to get to the grocery store across town, so he agrees to help him.

That's as much of the story as my friend and i worked out specifically. You can beat the game by going straight to the grocery store across town, however you will meet many people along the way and have many opportunities for side quests which will involve organized crime, political intrigue, totalitarian city governments, and corporate espionage. Our hero will play a pivotal role in all of these things, and will give a shit about none of it. He never voices his apathy, however, so he is lauded as a hero by the end of the game when in reality he only wants pancakes. Two plot points we have worked out is that at the end of the bomb victim's side quest he will be faced with the guy who ordered the suicide bombing of the general store. He will then demand that the man "look at the damage he has caused!" at which point he will yank off the bandaid on his cheak, revealing charred skin that looks just like all of the other charred skin clearly visible on his body. The villain will be shocked at this sight, but will still fight to defend himself. Also, at the end of the game, after you have defeated whoever the final boss will be and gotten to the grocery store to get pancake mix. The closing cinematic will show the hero arriving home and starting the pancake process again. He pours out two pancakes in a skillet when his phone rings. What proceeds is a few minutes of akward conversation with his mother. After he finally gets her off the phone he returns to his pancakes and flips them to show that they burned while he was on the phone. Our hero casts his spatula to the ground, raises his arms to the air, and screams "NOOOOOOO" as the screen fades to black, leaving us open for a sequel.

As for the actual mechanics of the game, we're thinking something similar to the open system they used in FF12, except not retarded. Each character will have weapon and armor styles they specialize in (we want to introduce a hippy anarchist chick who fights with molotov cocktails and a gun-nut who fights with various home firearms. Virtually all of the characters will be some manner of caricature) as well as various skills. Haven't decided on a skill system yet. One thing we definitely want, however, is what we call the e"quip"ment system. In this game the characters will launch various quips at friends, enemies, and eachother during play. For example, cutting off someone's head and saying "Not a good time to lose one's head!" Basically the kind of thing you'd hear in a Bond movie. The characters get access to new quips depending on what equipment they're wearing. Cooler looking equipment, like fedoras and heel-boots, will allow for more and better quips, but are less functional as armor and weapons.

Those are my retarded ideas, what do you think? And what are yours?

### Re: Describe a game you think would be awesome

Posted: Thu Dec 11, 2008 10:54 pm UTC
I want more multi-genre games.

Example: Imagine a modern RTS similar to Command and Conquer, with a kind of hero-unit (or squad) function. In single-player games, the AI controls them and you give them rough commands which they generally follow.

In multi-player games, you have the option to let other humans play them, as your RTS game is compatible with the maps for a corresponding modern FPS game. You can match up and talk with players of the FPS and put them on a favored player list, so that the server favors matching you guys up whenever possible. Then, you and your hero-unit players go fight other teams consisting of player (or AI) controlled leaders with player (or AI) controlled military units.

Edit: Ha, I just noticed Rippy has a thread dedicated to just this sort of example! I'll make another example:

Puzzle Quest (though ideally, the monster AI wouldn't cheat so blatantly to ramp up the difficulty).

### Re: Describe a game you think would be awesome

Posted: Fri Dec 12, 2008 12:26 am UTC
Be more specific. Describe the game, not the genre!

### Re: Describe a game you think would be awesome

Posted: Tue Dec 16, 2008 8:35 am UTC
I want an MMO based on George RR Martin's Song of Ice and Fire Series. It would have to be done kind of like Matrix Online, where the major characters are played by GM's and can be interacted with.

Either that, or I want to Warhammer 40K MMO to finally come out.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 16, 2008 8:38 am UTC
PS- One thing I just remembered.
Indon wrote:I want more multi-genre games.

Example: Imagine a modern RTS similar to Command and Conquer, with a kind of hero-unit (or squad) function. In single-player games, the AI controls them and you give them rough commands which they generally follow.

In multi-player games, you have the option to let other humans play them, as your RTS game is compatible with the maps for a corresponding modern FPS game. You can match up and talk with players of the FPS and put them on a favored player list, so that the server favors matching you guys up whenever possible. Then, you and your hero-unit players go fight other teams consisting of player (or AI) controlled leaders with player (or AI) controlled military units.

Edit: Ha, I just noticed Rippy has a thread dedicated to just this sort of example! I'll make another example:

Puzzle Quest (though ideally, the monster AI wouldn't cheat so blatantly to ramp up the difficulty).

There is a game like this. It is called Savage, and it is available for free online somewhere. In it, one player on a team is the commander and played it like an RTS controlling the worker units. All other units are controlled by other players who have an over the shoulder view and play it like an RTS, where they can spawn as any researched unit and buy any researched upgrade as
a weapon or item.

### Re: Describe a game you think would be awesome

Posted: Wed Dec 17, 2008 2:11 am UTC
I would like a nomially multiplayer RPG. Just you and a few friends. The combat system would sort of be like star ocean, except the other three chars are controlled by people instead of being raped by lizardmen. And the role playing portion would respond differently for the different characters.

I would like game where you cast spells by drawing glyphs, the actual shape of what you draw determines if it's fire/ice exploding/flowing ect. Definitely with an analog input like a mouse or the DS touch screen. Maybe as it's own puzzle game or embedded as the magic system into another game.

### Re: Describe a game you think would be awesome

Posted: Wed Dec 17, 2008 3:49 pm UTC
XCOM in full 3D with fully destructible terrain, but maintaining the same basic atmosphere and gameplay of the original. These Russians don't seem to be able to pull it off.

### Re: Describe a game you think would be awesome

Posted: Wed Dec 17, 2008 4:08 pm UTC
Two Four ideas I'd love to see:

1: I want a Silver Surfer game for the Wii, using the balance board in conjunction with the wiimote and nunchuck....and for it to NOT be a rail shooter, but more of a full starfigher-esque game. I imagine you use the balance board to dive, pull up, and bank left and right, and use the control stick on the nunchuck to rotate the camera, and the wiimote to aim. Use remaining buttons for stuff like selecting a kind of power to shoot folks with, forward thrust, etc.. Probably move the nunchuck sharply in a direction to dodge/drift in that direction.

Also, environments need to be HUGE, like you're almost going around an entire planet in space, fighting off giant spaceships, etc.. Probably almost never go down TO a planet, focus on space battles. Probably go to different planets through hyperspace, maybe have a worm tunnel section for that that IS like a rail shooter.

2: I know they're already doing a sequel to FFIV that happens a LONG time in the future, but I've always wanted one that was just a year or two down the line, in which you play exclusively as Kain. He said he was going to go off and train to be a better Dragoon, so where is he going, and what does he do? I want a direct combat action-RPG that lets you take advantage of the absolutely insane agility a Dragoon should be capable of. I imagine a kind of hybrid of Kingdom Hearts (action-RPG, leveling, etc.), Ninja Gaiden (hack'n'slash), and Spider-Man 2 (crazy agility). I want a char who can run at incredible speeds, leap for miles, run on walls and ricochet off them, and then put a spear through a dragon's throat at 800 feet in the air.

And since Kain doesn't really use much magic himself, I figure maybe let him get a wyvern companion partway through, similar to in FFXI. Let the wyvern be used for the "casting," let it have elemental breaths and also be able to cure Kain. Maybe a bit like the Dog in Secret of Evermore, except here it's the support pet that does all the casting. Probably Kain could use items to revive it if it was killed, though.

Also, in terms of equipment, there'd probably be shoes/leggings you got that increased jumping ability as you went through the game, increased running speed, let you run on walls a bit, etc.. Also, weapons would probably be split between things like spears, lances, polearms, javelins, etc.. Some would be better for swinging, some for thrusting, and some would be for throwing. Main attack buttons would be one for stabbing, one for swinging. With just those two, you could probably give a char some pretty involved combos.'

EDIT 1:

3: I'd also toyed with the idea of this being the first in a series of FFIV sequels, with each one focusing on a different char, and the gameplay being focused on that char being totally in their element. I.E., Edge would be typical ninja sneaking and backstabbing, etc., Rosa would probably have sniping and healing innnocents, Rydia getting to summon giant scary things and blow stuff up with black magic. The problem would be Cecil, as what a Paladin does best is defend folks....and if you make that the whole focus of a game, it becomes a giant escort mission, which sucks.

EDIT 2:

I've also been waiting for YEARS for them to make a Bionic Commando 3D remake, and now they finally are, and I'm ecstatic

4: Also, I think they should do a 3D remake of Mega Man 1, with heavy emphasis on exploring the boss's areas. I picture the game as being a bit more metroid-like: Just one large world, but it's sorta split up between different territories. I figure that you have a functioning teleportation unit, but with two restrictions on it. Firstly, it can teleport FROM anywhere (usually), but can only go TO certain locations that have a receiving unit set up. Second, part of the plot can probably be that each territory has a number of jamming devices that prevent teleportation within little areas....go and blow them up, and you can teleport out from that region. There's also probably certain receiving units that have been turned off....power them up again, and you can quickly teleport to that location from anywhere that isn't currently jammed.

Another twist would be that there's actually multiple locations where the bosses could be hiding, so instead of just making a bee-line for them, you have to actually hunt them down. Of course, they probably pop up now and then to battle you briefly or make your life a living hell, then after that drop the teleporter jammer just long enough to flee back to their sanctum.

Also, when you defeat a boss, there's probably a BIG explosion....and parts of them get sent all over the game world. Only their basic weapon function is left where they were, but if you can find the other pieces, then you can access other abilities of theirs. Maybe Fire Man's armor lets you survive extra-hot areas, or Elec Man's feet let you walk accross electrified floors without getting zapped. Combine Bomb Man's armor with his weapon, and you can probably pull off something like a rocket jump. Naturally, some pieces will be required in order to access others.

Oh, and I'm not sure, but it might be cool to also have NPCs around, who you sometimes have to protect, or who give info, items, etc.. This city is supposed to be a war zone, so where are all the freakin' humans?!

### Re: Describe a game you think would be awesome

Posted: Thu Dec 18, 2008 11:14 pm UTC
There need to be a game based on the battle "simulator" in Ender's Game. I would become religious.

### Re: Describe a game you think would be awesome

Posted: Fri Dec 19, 2008 11:10 am UTC
My friend and I have put together a prototype for a game idea I had while I was at the bird sanctuary. I'd like to complete it.

It's a top down survival horror called bird sanctuary. You carry a bag of bird seed. As you get close to a group of ducks, they start waddling towards you, and you must distract them by throwing bird seed off in strategic directions while you navigate the level. If they reach you, they begin tearing you apart in their ravenous quest for bird seed. I think the controls would be similar to snow craft, if any of you played that. You essentially move using the direction buttons, and point to where you'd like to throw bird seed with the cursor, a meter would fill up as you held down the mouse button, with a marker indicating the accuracy of your shot.

It sounds simplistic, but I think it has a lot of untapped gameplay potential.

### Re: Describe a game you think would be awesome

Posted: Fri Dec 19, 2008 1:44 pm UTC
@ roundedge
That's trippy. I think that would be one of the survival horror games I'd hate to play simply because it would freak me out too much (which to me means it's probably a success. I couldn't stand to play the early Resident Evil games and I don't want to play Dead Space for the same reason).

I have a concept for a spaceflight MMO too, Gunfingers. Long description is long.

Spoiler:
Essentially it's more a large scale further in the future one that what you've described. I like the newtonian physics aspect (inertia and engine power vs. mass determining acceleration), but instead of distracting the user during long trips, I propose trade lanes. It's already been done before in games, essentially like a giant mag rail that pulls ships within it forwards. Thousands of large rings form the lane between two planets or starbases or what have you. Shooting the rings disables them, so pirates can ambush traders in the lanes.

It's very heavy on economy, though. Agricultural planets produce foods and buy luxury goods and such, mining bases built on large asteroids buy food and tools, and sell raw and/or processed minerals (processing stations could provide another step in the chain if it isn't feasible to have the processing station on the asteroid with the mining station), while other planets and bases can demand the minerals and other materials to manufacture goods, tools, ship components, what have you.

Since there's an existing supply and demand mechanic, the servers can keep track of the prices and update them so everything runs smoothly. Players naturally chase profits and perform the trading, but CPU controlled traders can exist just to keep the balance running. As much as possible I'd want to avoid direct intervention in the universe by admin/moderator types to rebalance things, so basically there's be automated in game traders to prevent scams. That doesn't stop powerful players from blockading planets, but this is where the server run stuff really starts to kick in.

The game keeps tabs on trade routes, pirate activity, basically everything players are doing. If a trade route is getting a lot of pirates, the game throws up patrol missions to players for cash, so players can essentially become "police" on top of being traders or pirates. If a particular pirate is becoming very aggressive, they can be targeted for a bounty by the system, forcing them to move on or fight off players who fancy themselves bounty hunters. You can place bounties on players yourself, too.

Throw in contraband goods and you have smugglers. Players who do patrol missions can be given cargo scanning equipment (A big part of creating genuine classes out of these jobs is in the ship components and ship types that are specialised to them). Smugglers might actually want to dock with capital ships to get through trade lanes and the like (You only have to be more wily than the hired cop players, after all) or take less patrolled routes to get through. Contraband naturally fetches a higher price than legit goods, and police are rewarded for busting smugglers moreso than just shooting down pirates.

To minimise greifing and hammering lower players, you separate the system in concentric circles, all centred in on Sol. The inner systems are the newbie areas. Any player can enter or exit, but you can't buy space stations or planets there, it's really just the low level class stuff. There's a higher percentage of computer controlled traffic responding to any incidents involving larger craft, making it painful for any big players to really cause damage to the little fish there, not that they'd get much profit from it anyway.

Out in the 'patrolled sectors' players can have the opportunity to make more money from trade lanes, building their own production facilities on planets and generally populating more of the galaxy for profit. Player owned buildings and trades over a certain value are taxed to fund the patrols (both automated and player based) and while there is less police presence, there is more profit.

Finally, the great unknown houses mostly uninhabited planets and unmined asteroids, maybe alien artifacts too. You're on your own out there and empires can rise and fall quickly, but players can always return to patrolled space if they get knocked around too much. Patrolled space can actually increase in size; if you colonise the outskirts you can make a one off payment to turn it into a patrolled zone, which makes it taxed also, but lets players have the best of both worlds.

Just like your concept, the player is one person. While you might start in single man vessels, eventually you'll be buying huge trading vessels and capital ships, whole fleets if you're big enough. You can make money by offering smaller players passage aboard your ship, or hire players to man your vessel. Only I'd be a little less strict about it and let any unmanned stations be handled automatically, though not as effectively as a good human player (turrets fire slower and less accurately, etc). The cut off would be pilots. You can have a capital ship with a fleet of 2 dozen fighters and bombers, but if you're on your own, you can only launch the one. No computer controlled fighters.

I haven't played Eve so I don't know how similar the whole economy and planet owning and such is, but my idea is the game just changes over time. You start out on your own, learning the ropes, and if you're good, you'll naturally move up to bigger things. I'd expect clans might dominate the outer sectors, but it's not like that would be where all the fun was. It'd just be for people who were into that sort of thing. If the system worked well, I'd just want to be adding in content that let players weave their own stories. Stuff like planting an artifact in some unpopulated place and picking some random guy who didn't have a ship capable of using it, telling him where it was and seeing what he did. Maybe he'd sell the info, maybe he'd sit on it. Maybe I'd tell people who'd want to know that he knows the location of a powerful artifact and see what happens.

That's enough from me for now.

### Re: Describe a game you think would be awesome

Posted: Sat Dec 20, 2008 2:07 pm UTC
I want an Anime Space Strategy RPG w/ Massive ships and massive fleets that doesn't take 10 days to complete one game.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 5:57 am UTC
As you can tell from my avatar/sig, I'm a big fan of the "action-puzzle" games. I wish there was a more steady supply of mind-bending goodness like Portal and Braid.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 6:37 am UTC
I want a damn "Dune" game in the style of Legend of Zelda. You've defeated the boss, and thereby acquired more prescience! Also, YOU GOT THE MAKER HOOKS! Use them to ride Shai-Hulud!

OK, that was a joke, but I wonder... what would prescience look like as a game mechanic? Just to start, let's ignore the sheer amount of computation necessary and just consider how a game would work in which players could, in some limited way, see the future game state.

Also, in the spirit of "things that should be in sequels of games I like", there should be an espionage video game, or at least espionage missions/gameplay in games with intrigue-based plots.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 7:36 am UTC
A sort of casual (but neat) Wii shooter with the following characteristics:

Spoiler:
You control a multi-legged robot. While running or rolling the crablike machine either uses all eight of its legs for stability and speed or rolls them up into itself, but when walking or stationary four of its legs rise up and can be used as guns. One player controls the machine and a single gun, while the others each man one of the guns. Though the game has stealth elements it's a pretty standard sort of action shooter (i.e. aim at the screen to shoot). However, charge attacks are possible and if multiple players target nearly the same spot a sort of super attack can be done (let's say a flurry of missiles or some kind of megashot). This would be useful for bosses.

By swinging the remote a player can use their arm as a claw for smashing or attacking, or can dig or drag it into the ground for added stability and traction.

Alternately, each player could be in charge of a different function of the robot: one could have a gun targeted with the remote, one a sort of mace that can be swung out either to attack enemies or work as a grappling hook, one a rocket and one in charge of walking, shields and an energy blade. It's difficult to realize a game that pretty much requires four players, but this one would be pretty awesome. Added flexibility through giving each player the option of transferring their ability to any other player at a given time or maybe even having two of the same weapon (two maces for suspending the robot like a hammock).

I would prefer to see this as a gory sort of game with a robot sporting beastlike characteristics (it runs like a sort of crazy squirrel and makes breathing sounds rather than being cool and metallic like most robots).

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 8:30 am UTC
Another space game here. It's one I have actually started designing properly and sort of know how to implement.

Space Action RTS RPG game in space, described in just a few words. Most of the action happens in "Action mode", which is a top-down 2d view of your ship in space, sort of like asteroids. You fly around and shoot stuff, basically. Of course, you can visit space stations, planets and all that fun stuff, where you get new missions, upgrade your ship, do trade etc. Eventually you'll also take on wingmen and form a squad of sorts. You'll be able to upgrade to larger ships of various types, build space stations of your own, have ships under your command who's not directly following you but off doing their own thing on the other side of the galaxy. I'd also like to put in character-based action scenes, like boarding action, and possibly even planet invasion, though I'm not sure exactly how this is to be done yet.
All characters will have stats affecting their abilities to perform in various areas. The game will be all 2d, low-resolution, like all the best games in similar genre have been.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 3:47 pm UTC
aleflamedyud wrote:OK, that was a joke, but I wonder... what would prescience look like as a game mechanic? Just to start, let's ignore the sheer amount of computation necessary and just consider how a game would work in which players could, in some limited way, see the future game state.

The game would probably feed you cut scenes or images that gave hints or directions on how to continue. In combat, maybe your super-fast mind would be shown by temporary slow motion.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 5:17 pm UTC
Prince of Persia: Sands of Time does that, though it isn't prescience, just regular time manipulation.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 5:57 pm UTC
A sim city that lets you conquer other cities. I mean you have a military base in the game but you don't get to do anything with it.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 9:01 pm UTC
Gunfingers: Check out Kingdom of Loathing.

For me it would probably be a combination of Engine of War, Spore, Guild Wars and an FPS. Preferably following in Guild Wars' foot-steps with no monthly fee.

The game begins with you working in some scrap-shop type thing with this rusted mech. Possibly add an Iron-man-esque back story. Repairing the mech would be tutorial of the game. Wiring would have to be replaced, stuff re-programmed, etc. This aspect of the game would be very similar to Engine of War however the Spore creature creator would also be incorporated and expanded as well. Unlike Spore, physics does matter, things like balance need to be considered. Also since combat would be done in an FPS style, armor positioning and general shape would control the hitboxes overall making a huge difference on how it runs. Cooling and Power also need to be addressed. With only a limited amount of points you can only do so much.

Now not everyone would want to have to worry about engineering every individual piston in their legs so something like a sporepedia would be integrated, however only things that you wish to be shared would be shared. This would allow the top players to retain atleast some of their trade-secrets. PvP would be the focal point of the game, a ladder, observer mode (for top games) and clan battles would be a must.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 9:23 pm UTC
Berengal, your game sounds like Escape Velocity (except for the building of space stations, the purpose of which I can't imagine. Then again, if you could create a game like Escape Velocity but with a more persistent world you'll succeed anyway.

### Re: Describe a game you think would be awesome

Posted: Sun Dec 21, 2008 10:50 pm UTC
Let me first give you a metaphor for my 'awesome game'.

You take Darkfall and Erepublik (a game that calls itself a 'social strategy'), and marry them. At the same time, you take Civilization and Pharaoh/Caesar III and breed them. Then you take the offspring of each pair (D/E and C/P) and breed them together (giving D/E/C/P). Then you give that a healthy injection of Dwarf Fortress and send it on its way.
You'd either get a game of magnificent proportions or a total flop.

The far longer explanation that actually makes (a bit of) sense:

Spoiler:
Essentially, it would start with the creation of an entire persistent virtual continent (or preferably more than one) in the manner of Darkfall. Within this, when the game begins, the players are dropped into some nondescript corner of the world and they have to build. They have to build a whole civilization, that is. From the ground up.The game world would be 'realistic' as much as possible, with little to no outside interference.

So on this first level it would be a social MMORPG: You would have a character or possibly a family of characters, and you would be left to your own devices. You would choose what you do each day - as your characters perform tasks they become more skilled in that field in a Darkfall-ish manner, that is - no levels. But be warned, you can't be skilled at everything; whenever you haven't used a skill for more than, say, a week, of ingame time it starts to decay. And besides, each character has a 'lifespan' which would last several months at least in real life time, but that would mean that you generally wouldn't be able to master more than one or two skills before you conk out. You know how the saying goes, jack of all trades, master of none.

So from this you would have a social interdependency - everyone would need food, but people would need wood and stone for their buildings too, not to mention people to work them - and of course, you can't do it all yourself. The individual roles in the society would naturally diverge, as people move from being solely hunters living in the tents they are given at the start, to slowly discovering the bounty of the resources in the area - finding that hey, we can probably build things with the trees all around us. People work out how to cultivate crops, Bam, you've got yourself Agriculture.

The flip side of this would be, as I implied before, that it would be completely unenforced - no silly game rules to stop anyone jumping you, killing you, stealing your house, raping your children and wearing your hat. Or worse, scamming you. No restrictions whatsoever. If you want to cheat, go ahead and do it as long as it is due to other people being stupid and not due to exploiting a bug in the game. Chaos, right? Wrong. At the early stages, I'm sure people would try it, but you try living easily when you've been exiled by the rest of society. Or when you've been stoned to death by people who enjoy keeping their possessions. Once a large population of 'law-abiding' citizens has been reached, it becomes much more feasible to have a dark underbelly to a city or town, as it becomes possible to escape anonymously into the crowds, if necessary. But at this point, anyway, it's pretty hard to debalance society into a chaotic crimefest once there is a big lawful population.

Which brings me to another aspect of the game; you would have to work it out for yourself, with a touch of intelligence. All of it, beyond the basic interface; the game 'rules' would be hidden. So in the same way has having to, say, keep your wits about you while trading so you're not scammed, you also have to work out where is best to build your house, what to build it from, which types of land are more fertile, which of the fruit can be eaten, whether your band of pickaxe weilding thugs can beat that shady looking bloke with the rapier if you attack while his back is turned. Ultimately, your success or faliure in the game would be due to your intelligence and decision making rather than your character's stats.

Life - You'd only have one. If you die, shock-horror, you are not about to get up and go for a ghostly jog, you are dead. Your options would be continuing as one of your late character's family members (meaning you keep any possessions you didn't lose when you died) or as an entirely new guy. This would really enforce the social aspect of the game - being a general dick will mean that you will end up with a dagger in your back and you're back to square one. Yep, that house that took you three weeks to build is now the property of the guy that you attempted to kill for teh lulz. Oops, perhaps you should have noticed that guard standing nearby.
Similarly, injuries are pretty permanent. If you get an arm lopped off in a war, it isn't gonna grow back. You're armless for the rest of your life. Some small injuries may be cured if someone nearby has discovered the appropriate herbs to make a poultice, but most likely you will be out of luck.

Sandboxishness - this is where the dwarf fortress-ish aspect comes in to play. As I said before, everything would have to be worked out yourself, or taught by a wiser player. So you could design your house however you like, build it out of whatever you like. Of course, 95% of the things you try won't work, but with a bit of rational thinking you should be able to build something that works reasonably well - and improve on your ideas. Think every aspect of game mechanics horrendously detailed in a dwarf fortress kinda way.

On the whole it would work as an RPG, in a similar sort of way perhaps to being one of the dwarves in Dwarf Fortress. But once you are in a position of power, it may act more like a strategy game in some situations, or a citybuilder if you manage to become the ruler of a town, or even if you're lucky it would be a game of Civ with real people once you become the ruler of a nation...

blech brainmelt, I may add more later...

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 2:16 am UTC
For the prescience, I'd thought of maybe doing something kinda like the water tendrils in "Donnie Darko,", where there'd be colored streamers coming off of enemies at all times, showing the path they were likely to take. Right next to the enemy, they'd be strongly colored, but they'd fade away the further from the enemy they got, showing you were less sure of what actions they'd take after that.

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 3:03 am UTC
That would require the AI to be deterministic, in a sense. For the pathways to be accurate and therefore useful, the enemy would have to follow them regardless of unexpected changes in the environment. Like a guard on patrol; it'd be fine while he was on patrol, but wouldn't he react to you shooting at/near him? Would the path suddenly change, making its predictions less prescient and more extrapolation? Or would he continue on the preset path while the path changed at its end to reflect the change, making all enemies react a second or two later to anything?

The only way I can see it implemented is in reverse, ala Sands of Time, where you could attempt various possibilities and rewind back to a point until you achieved the end you were after, which is really what prescience would allow you to do. For true prescience this would have to be an unlimited ability.

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 8:18 am UTC
InstinctSage wrote:That would require the AI to be deterministic, in a sense. For the pathways to be accurate and therefore useful, the enemy would have to follow them regardless of unexpected changes in the environment. Like a guard on patrol; it'd be fine while he was on patrol, but wouldn't he react to you shooting at/near him? Would the path suddenly change, making its predictions less prescient and more extrapolation? Or would he continue on the preset path while the path changed at its end to reflect the change, making all enemies react a second or two later to anything?

The only way I can see it implemented is in reverse, ala Sands of Time, where you could attempt various possibilities and rewind back to a point until you achieved the end you were after, which is really what prescience would allow you to do. For true prescience this would have to be an unlimited ability.

As I understand it, the paths would be more extrapolations... what the AI plans to do in the next t ticks of time rather than what it actually will do. Even Paul Atreides, after all, couldn't see past a decision made by a prescient or a consequence of such unless its probability was massive.

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 9:19 am UTC
Berengal wrote:Another space game here. It's one I have actually started designing properly and sort of know how to implement.

Space Action RTS RPG game in space, described in just a few words. Most of the action happens in "Action mode", which is a top-down 2d view of your ship in space, sort of like asteroids. You fly around and shoot stuff, basically. Of course, you can visit space stations, planets and all that fun stuff, where you get new missions, upgrade your ship, do trade etc. Eventually you'll also take on wingmen and form a squad of sorts. You'll be able to upgrade to larger ships of various types, build space stations of your own, have ships under your command who's not directly following you but off doing their own thing on the other side of the galaxy. I'd also like to put in character-based action scenes, like boarding action, and possibly even planet invasion, though I'm not sure exactly how this is to be done yet.
All characters will have stats affecting their abilities to perform in various areas. The game will be all 2d, low-resolution, like all the best games in similar genre have been.

If you want to develop this, I'd be interested in helping out. I do a lot of game developing, and I looove 2d low resolution games (and I can actually draw somewhat decent in that format).

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 9:32 am UTC
I can do pixel arts lawl.

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 6:01 pm UTC
notch, I'd heard of you before but I hadn't gone to your site. logic hex is fun I covered the field in not gates and one + pretty patterns resulted. Miners4k is fun and it gave me an idea for a game that makes this post on topic.

You have a whole bunch of little critters like in miners that have rules for when they jump turn around and when they dig, and when they place signals and what those signals mean. (you can edit the code and the whole result would be something like an L fractal.)

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 7:31 pm UTC
Clumpy wrote:Berengal, your game sounds like Escape Velocity (except for the building of space stations, the purpose of which I can't imagine. Then again, if you could create a game like Escape Velocity but with a more persistent world you'll succeed anyway.

Never played that, but it sounds familiar (in that I've played similar games). If you'd played X3 you'd understand the space stations. They're also in because all my game ideas eventually converge on the "everything game" game, and last I checked "everything" includes space stations. If given enough time and resources any game of mine would turn into a fully fledged OS, and would eventually even be able to make you real life coffee and beat the turing test.
Notch wrote:If you want to develop this, I'd be interested in helping out. I do a lot of game developing, and I looove 2d low resolution games (and I can actually draw somewhat decent in that format).

I'd love some help, but at the moment I'm still just messing around with the design. I also have the bad habit of a short attention span when it comes to fledgeling ideas, so until this one's matured enough and I've found a proper course of development I consider it "some crazy idea that would be really cool if became real" and I don't feel like involving others in something that's probably not going to happen (probability calculated based on my previous track record. My guts tell me this'll really happen, but I'm a man of facts). When/if I feel like I'm on to something real I'll definitely cry for help.

### Re: Describe a game you think would be awesome

Posted: Mon Dec 22, 2008 11:59 pm UTC
I wish there could be a Harry Potter RPG instead of just a movie tie-in game. I'd love a sprite rpg with all the HP characters that was a stand-alone story. Kinda like Sailor Moon: Another Story...but I'm probably the only person who has played that

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 1:41 am UTC
aleflamedyud wrote:As I understand it, the paths would be more extrapolations... what the AI plans to do in the next t ticks of time rather than what it actually will do. Even Paul Atreides, after all, couldn't see past a decision made by a prescient or a consequence of such unless its probability was massive.

It sounds like a neat idea, but difficult to implement in a useful fashion for a shooter. What I think would be awesome is a hand to hand fighter using this mechanism. Since the Bene Gesserit are trained in the Weirding ways, along with the petit perception, instead of having Donny Darko ghost tubes sticking out of people you could have an entire ghost of the model running half a second to a second ahead. The AI is lagged in this case and reacts late to your movements, but that's the whole point.

You could have a 1st or 3rd person hand to hand fighter where prescience gives you the ability to know what move the enemy was going to make before they made it. Thus, you could parry, dodge or block just about any attack and completely dominate in a one on one situation. You could start throwing in 1 vs 5 or even more, so the player is outnumbered by enemies who are actually fairly lethal, but the player will be able to beat them all without taking a hit without resorting to overdone slow-mo effects.

I think it'd be pretty awesome to be having matrix style standoffs against a dozen enemies who are attacking viciously, yet being on top of it all while it's running in real time.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 6:42 am UTC
InstinctSage wrote:
aleflamedyud wrote:As I understand it, the paths would be more extrapolations... what the AI plans to do in the next t ticks of time rather than what it actually will do. Even Paul Atreides, after all, couldn't see past a decision made by a prescient or a consequence of such unless its probability was massive.

It sounds like a neat idea, but difficult to implement in a useful fashion for a shooter. What I think would be awesome is a hand to hand fighter using this mechanism. Since the Bene Gesserit are trained in the Weirding ways, along with the petit perception, instead of having Donny Darko ghost tubes sticking out of people you could have an entire ghost of the model running half a second to a second ahead. The AI is lagged in this case and reacts late to your movements, but that's the whole point.

You could have a 1st or 3rd person hand to hand fighter where prescience gives you the ability to know what move the enemy was going to make before they made it. Thus, you could parry, dodge or block just about any attack and completely dominate in a one on one situation. You could start throwing in 1 vs 5 or even more, so the player is outnumbered by enemies who are actually fairly lethal, but the player will be able to beat them all without taking a hit without resorting to overdone slow-mo effects.

I think it'd be pretty awesome to be having matrix style standoffs against a dozen enemies who are attacking viciously, yet being on top of it all while it's running in real time.

Who said a shooter game? Prescience would make shooters far too easy, IMHO. Your action/adventure/fighter scenario sounds much closer to what I had in mind. Now as to implementation, AFAIK most enemy AI already plan ahead by a few seconds in current games of the appropriate genres. That's how they do things like execute combos, yes? So we really would have to just keep a queue of moves the AI is planning in time order from execute-this-step to execute-next-step to execute-next-step-after, and so on. The "actual" enemy would pop the next move off this queue and execute that, and afterward (in programming terms, but during the same frame) the ghost would peek at the move at the head of the queue and execute that. When the AI reevaluates its strategy and/or changes its plans entirely, it would clear this queue and the position and orientation of the ghost would reset to those of the actual enemy.

Now as to Weirding Way, I always thought of that as being like automatic counter-attacks, timed combos, and the "hidden skills" from Twilight Princess mixed up in a blender. They let you execute a powerful (or even lethal) move quickly and without the need for showy magic or complex button combos. Say we have a single button combination WW (for Weirding Way, duh). To use the Weirding Way properly, we'd have to first lock onto an enemy, then use some footwork to get into proper position (this could include dodging or rolling to get a better angle on our enemy or counter), and then hit WW. If the enemy doesn't react in time to defend, the Weirding Way skills come out and look cool. If they do defend or attempt a counter, we'd have to hit WW again at the correct moment to counter-counter-attack. As noted above, our own counter-attacks would work by dodging/defending and then executing a timed WW.

Fuck fantasies of a Dune game, I think we've got the beginnings of a good game of our own here.

EDIT: A few more ideas are coming. Ideally, a player in such a system shouldn't even need counter-kill moves done automatically. Instead, they should line their enemy up such that the enemy's ghost begins performing a combo, and then dodge/defend. Thus, the player (having seen it coming) is in position to do a simple, ruthless attack as their enemy actually begins their combo or even does a one-step attack (though one-step attacks would of course be harder for the player to work with). It's like Cohen the Barbarian in "Interesting Times": a good player shouldn't have to do some elaborate dance over the battlefield, they should just always be where they want with an opportunity to strike because they reacted to what was going to happen instead of what was happening. Of course, some enemies and bosses could even turn this to their own advantage by changing attacks/combos/strategies frequently, forcing the player to rely on real-time fighting skills.

For setting, I think replace Duniverse with roughly the modern day or the near future in a largely urban environment. This gives our setting gun/arms-control laws, which provide a rational explanation for why the player can't just shoot their way through the game, and instead has to use martial-arts moves to fight and knives or other small blades for their kill moves.

'nother EDIT: For puzzle or platforming situations, the game would allow the player to use prescience for a Braid-like mechanic. The player would have to hold themselves in "prescience mode", scanning the possible future time-lines, which would disable them from moving or doing anything, but this would let the player create a ghost-self that could perform any action the player desires with no actual consequences, since all effects the self-ghost's actions had would be ghostly themselves and undone when the player ends "prescience mode". To put pressure on the player not to just take a "see what happens" approach to all gameplay, the world outside the player-ghost and its actions keeps running while the player scans the future in prescient mode.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 7:20 am UTC
I have no idea where the shooter thing came from. For some reason I thought someone else had brought it up, but it was me all along, ho ho ho.

In my limited experience looking at AI, it's usually a 'state based' system. AI states or modes are triggered by various actions or events. This can be as simple as the player making a noise and AI changing from a 'rest' state to a 'searching' state, and can become as complex as the player starting to run towards the AI and the AI changing state from a 'neutral attack' to a 'charge defense' state. The AI remains in the state it's in until a trigger moves it to a new one, and each state naturally has actions that make sense. This is efficient since the system only has to pay attention to a small block of instructions for guiding current actions and is alerted by the trigger events whenever the situation changes.

In this way, combo attacks are actually pre-programmed in, which is why (particularly in versus fighting games) you'll see the opponent react the same way to a situation that keeps arising, or fall for the same trick over and over. It's the main reason practicing against the CPU on versus fighting games can't take you very far. Over time they can be very predictable, and any flaws in their defense are never adjusted for, so you can punish them over and over, even on the hardest difficulty.

I thought about it more and it would probably be a bit clunky to implement a ghost image constantly though, so all in all we've come to the same conclusion. The ghost can telegraph any attacks or other moves and the player gets a chance to either pre-empt, or counter while the AI opponent is committed to their attack.

I hadn't considered an all purpose awesome/special attack/counter button. Personally they're not my thing, and I considered the weirding ways would just be the fight style in general, not a super move. If I had to come up with a control mechanism I'd say we'd need a dodge/roll button, a block, an attack and a parry/counter. Counters can usually be put with block as a "hit just on strike" thing, which works well. Depending on how developed you'd want the fighting, you might need to extend the range of attacks to more than one button though, like high/low, punch/kick, normal/heavy, etc.

It could very quickly get unwieldy for a controller.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 7:29 am UTC
Okay, mine's a bit of a different idea than the rest in that I have no idea what the plot or characters would be about.
A FPS/3rd person shooter/Sandbox platform with complete freedom, destructible environments and buildings and the like with really good physics but average graphics (For framerate). Instead of making a proper story and a full game, the creators would instead bundle an easy to use (Think FC2 map editor) editor with it, and keep up a communal map exchange server for people to trade and host their creations on, as well as the parent company creating gameplay as the game gets more popular, possibly with true episodic content released weekly to the populace. Essentially, an engine and a map creator with very little non-user-created game, creating a true sandbox experience (Like Gmod, but with better creation tools and physics). Failing that, I want Rockstar to make San Andreas again on the GTAIV physics engine, with the classic graphics.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 8:03 am UTC
InstinctSage wrote:In my limited experience looking at AI, it's usually a 'state based' system. AI states or modes are triggered by various actions or events. This can be as simple as the player making a noise and AI changing from a 'rest' state to a 'searching' state, and can become as complex as the player starting to run towards the AI and the AI changing state from a 'neutral attack' to a 'charge defense' state. The AI remains in the state it's in until a trigger moves it to a new one, and each state naturally has actions that make sense. This is efficient since the system only has to pay attention to a small block of instructions for guiding current actions and is alerted by the trigger events whenever the situation changes.

In this way, combo attacks are actually pre-programmed in, which is why (particularly in versus fighting games) you'll see the opponent react the same way to a situation that keeps arising, or fall for the same trick over and over. It's the main reason practicing against the CPU on versus fighting games can't take you very far. Over time they can be very predictable, and any flaws in their defense are never adjusted for, so you can punish them over and over, even on the hardest difficulty.

Ah, state machines! I like it! To put the planning system into an FSA-based AI, we simply need to add an input symbol into the language the FSA interprets: null. The null symbol means that the player (or other AI) has done nothing that would make the AI change its plan, even though the player may have actually done something. Or do we need even that?

Anyway, now we have an FSA representing the AI, and we of course have a queue/stream/string of input symbols to that FSA. Again, the actual enemy pops the next input symbol off the queue and feeds that to the FSA to get its actual new state. Then the ghost peeks at the next symbol and feeds that to the new FSA state to get the ghost's FSA state. The actual enemy then acts according to its state, and the ghost acts according to its state, which is one step of computation ahead of the "actual" enemy. Depending on the level or skill rating of the player's prescience, we could even have the ghost progress through the input stream at a different rate than the actual enemy to give players a nearer or further look into the future.

This is even more elegant than my stupid little planning queue, because FSA transitions are generally quite cheap to compute.

I thought about it more and it would probably be a bit clunky to implement a ghost image constantly though, so all in all we've come to the same conclusion. The ghost can telegraph any attacks or other moves and the player gets a chance to either pre-empt, or counter while the AI opponent is committed to their attack.

I hadn't considered an all purpose awesome/special attack/counter button. Personally they're not my thing, and I considered the weirding ways would just be the fight style in general, not a super move. If I had to come up with a control mechanism I'd say we'd need a dodge/roll button, a block, an attack and a parry/counter. Counters can usually be put with block as a "hit just on strike" thing, which works well. Depending on how developed you'd want the fighting, you might need to extend the range of attacks to more than one button though, like high/low, punch/kick, normal/heavy, etc.

It could very quickly get unwieldy for a controller.

You're right about auto-counters getting unwieldy. For the moment, let's consider an X-Box 360 or Playstation-layout controller, since these are the most commonly used on consoles and PCs. We do want some mechanism to simply execute automatic combos or counters, but we want to minimize the amount of controls used to activate it. Why do we want it? Because it commits the player to a course of action that another player could see as a ghost in multiplayer mode. Or do we not want multiplayer mode?

Anyway, we can pretty quickly assign one trigger or shoulder button to enemy lock-on. Versus fighter games lock-on automatically, so we only need this for a beat-em-up scenario. Then we might put fighting controls basically the way you described: hand, leg, block, weapon; thus we've assigned the four basic buttons. We're going to assume that the right control stick is not used in combat, since switching between the basic buttons and the right stick is damned awkward when the player needs speed. So now we have 3 trigger/shoulder buttons left to assign, and possibly the D-Pad. Since the D-Pad is also awkward to reach, we may as well go and assign its four directions to combo/counter moves: hand, leg, block, weapon. These could be performed with one tap of the pad, but set the player on a course-of-action they can't alter instantly. I figure one pair of the trigger/shoulder buttons should do left-right dodging. And since we have one button left, we can assign it to "jump" to perform jumping moves (jump + kick = jump kick, jump + kick combo = Trinity-Matrix-Kick, jump + dodge = roll, etc).

Now we've pretty much assigned the controller's functions for combat. Such a scheme may well be inconvenient or even annoying in practice, but I think it's a good start from which to improve. And if the game is not a pure-fighting versus-fighter (meaning that it has some world to move through), every single controller function could behave differently when the player has not locked onto an enemy, though in practice only the fighting buttons really need changing. The "dodge" buttons can just become "strafe" buttons, and the jump button can still make the player jump.

Yet Another Goddamned Edit: We could just give the player an inventory of moves and let the player equip them to fighting-move buttons in the same way that Zelda players equip items. This would have the benefit of letting the player build their own custom automatic combos out of the simpler moves (for example, a character could have a stat dictating how many moves they can chain into an autocombo) that could then be equipped to D-Pad directions, as well as letting the player customize the arsenal of moves they bring into a fight based upon what sort of enemy they would face. If the player faces an enemy with strong protection from kicks and sweeps, they could equip both the punch and the karate chop instead of being forced by the control scheme to leave a button for doing leg moves mostly useless in their current situation.

Oh, and a neat thing is that I think a fairly standard game-engine could do all this stuff. I'm not sure there are standard game engines for martial-arts beat-em-ups or versus-fighters, but if there is one that can do such a game, it could be used instead of having to write a custom game engine.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 2:53 pm UTC
May I make one other small suggestion regarding the prescience? I agree with the ghost-image effect, but the reason I originally suggested "streamers" is because the ghost-image only shows ONE moment in time, perhaps a second in the future. A prescient would be able to see the next FEW things coming. A streamer can show the general path an AI is thinking of taking...and it can also branch off in case they're thinking of more than one path. Using a shooter such as maybe Unreal Tournament as an example (just to make it easier to illustrate), imagine you're chasing the AI down a hallway, and the hallway branches off. there might be a streamer leading off the AI down the hall, getting fainter as it gets further away (signifying uncertainty), and it may have smaller, fainter streamers leading down each of the branching halls it's considering. As it gets closer and firms up its decision, the streamer for one branch grows fainter and dies off, while the other brightens and becomes more solid.

Now, the idea of using it for hand-to-hand is interesting, but then there's the problem of seeing the next FEW moves ahead. Basically, if an experienced player were to look at a frozen screenshot in the game, they shouldn't just be able to say, "Oh, that enemy is going to do an uppercut in about one second," they should be able to say, "Oh, that enemy is planning a body blow, followed by an uppercut, and then they're considering a trip after that while the player is stunned."

Perhaps a few ghosts shown at a time, with different color coding? Say the enemy holds 3 moves in its queue, at max. If it takes a moment to plan (though some may be more reactive than others), it may show a red outline for the next move it's planning, an orange, fainter one for the move after that, and a very faint yellow one for the third. This could get tricky, but if the entire FOCUS of the gameplay is on the prescience, a player would probably have to learn to deal with something like it...and if it's mostly just outlines of the AI, not an entire filled-in image, it won't overwrite too much. Also, a player could probably have the option of dialing down the prescience to not see so far ahead, simplifying things.

Also, two other things. Firstly, similar to how usually such games have a meter for using bullet-time, it may be there should be some sort of focus meter needed to use Prescience, that recharges over time, for two reasons. Firstly, it's less interesting if you're in God-Mode all the time. Second, in order for it to be prescience to be properly cool, there needs to be a point of reference of what it's like WITHOUT it. There's a difference in games between "you GET to do X, cool!" and, "you HAVE to use X, blegh." Think of the spider-ball in Metroid Prime games....and now think of if it was never NECESSARY to use it, but if there were just places where you had the OPTION to make things easier. How much more would you have enjoyed those tracks if they were all just shortcuts?

Another thing is that, with regards to meditating for platforming sections, I agree that you could have time actually continue while the player experiments with possible actions they could take, but in that case, you probably want real-time slowed down, with prescient-time continuing at the "normal" pace. Makes it easier to experiment, and otherwise, the prescient-actions would just be overlapping with the real-time movements of the platforms. You're not seeing "What I COULD do in the future," you're seeing, "What I COULD be doing right this moment."

EDIT:

BAH! More stuff keeps occurring to me! Two more thoughts for your consideration: Firstly, you may want to make it medium-range combat (such as God of War), because that would make it easiest to display the images showing what the enemy AI is planning, without obscuring what they're currently DOING. Hand-to-hand, there might be too much overlap. Second, just a thought, you may want to occasionally have bosses who are themselves prescient, and a (admittedly cheap) way to do this would be that, if you both have prescience activated at the same time, then they don't show any ghost images. If THEY have it on while YOURS is off, the cop-out way would be to reduce their own reaction time to practically nil (hey, so far as YOU can tell, it's almost like they knew what you were about to do). Obviously you either need to be good enough to beat them on a level playing field, or wait for an opportunity to strike while your prescience is on and theirs is off. Or possibly, there would be environmental factors you could take advantage of, particularly while you have prescience up ("So I see where that explosive fuel drum is GOING to go, and so do they....but they don't see where I'm about to SEND it").

You don't want to make it seem like the enemies are just morons and you're human and can predict AI, you want to make it seem like the enemies are themselves badass....it's just you're even moreso.

### Re: Describe a game you think would be awesome

Posted: Tue Dec 23, 2008 9:15 pm UTC
MMORPG, but with a little SIM city mixed in.
Spoiler:
A massive war between the gods has broken out, and it is being fought by humans. Players fill the role of clergy, wielding the power of their deity to destroy their enemies, aid their allies, and convert unbelievers.
Basics of your available skillset will depend on what deity you choose. One deity might focus in fire and destruction, while another will allow you to be a crazy good swordsman. And that's just in relation ot the combat skills. Skills in the game will be generally broken down between combat, craftsmanship, and agriculture. A player may train in whichever of these he wants, depending on his goals, or train all three. You get experience by killing enemies, creating goods, or harvesting crops, and all related activities. Or maybe just from combat, i'd need to decide based on balancing.
There isn't much in the way of MOBs in the game. You'll pretty much never go out and grind levels. What exists instead are "mortals", people who could not channel the power of the gods and are therefore just regular schmucks. Some of them will be part of your congregation and will follow your orders, others will work for another player who serves the same deity, and others will worship another deity, making them your enemy to be killed or captured and possibly converted. They will have names, will fall into one of the general skill areas, will live, grow, marry, have children, and die. A year will be compressed into an actual day, and your people will probably live a total of 30 days. You can train them, based on whatever skills you have. The stronger you are the stronger they will become under your tutelage. They can be made into mighty warriors, master craftsmen, expert farmers, whatever. But no matter how well trained eventually the mortal will die. On the plus side, their offspring will have retained some of their knowledge and skill because genetics are kind of Lamarckian in this game. A lot of this was inspired by the way the Dunedain were affected by living under the influence of the Valar. Thus if you take a guy and specially train him to be your mightiest warrior and he produces a young 'un while at the peak of his skill he will produce a stronger son who will have a natural skill in combat (might start off in level 3 instead of level 1 or something, which is significant because he only gets so much time to level before he dies) despite the fact that genetics doesn't work that way at all.
You get experience by killing them, or the enemy players who lead them. In combat players would function something like heroes in Warcraft III. That is to say, particularly powerful but most valuable for leadership and buffs rather than for individual combat capabilities. This will change grinding monsters for experience into leading raids and attacks against other players as a way to level up.
An incredibly complex AI would be necessary, because the mortal settlements would be perpetual and would have to be able to function and even defend themselves when players aren't online, but would do so differently based on how they were trained. They would basically go about their lives farming, making weapons and tools, pulling guard duty, fucking, and whatever.

### Re: Describe a game you think would be awesome

Posted: Wed Dec 24, 2008 2:32 am UTC
Aikanaro wrote:May I make one other small suggestion regarding the prescience? I agree with the ghost-image effect, but the reason I originally suggested "streamers" is because the ghost-image only shows ONE moment in time, perhaps a second in the future. A prescient would be able to see the next FEW things coming. A streamer can show the general path an AI is thinking of taking...and it can also branch off in case they're thinking of more than one path. Using a shooter such as maybe Unreal Tournament as an example (just to make it easier to illustrate), imagine you're chasing the AI down a hallway, and the hallway branches off. there might be a streamer leading off the AI down the hall, getting fainter as it gets further away (signifying uncertainty), and it may have smaller, fainter streamers leading down each of the branching halls it's considering. As it gets closer and firms up its decision, the streamer for one branch grows fainter and dies off, while the other brightens and becomes more solid.

Now, the idea of using it for hand-to-hand is interesting, but then there's the problem of seeing the next FEW moves ahead. Basically, if an experienced player were to look at a frozen screenshot in the game, they shouldn't just be able to say, "Oh, that enemy is going to do an uppercut in about one second," they should be able to say, "Oh, that enemy is planning a body blow, followed by an uppercut, and then they're considering a trip after that while the player is stunned."

Perhaps a few ghosts shown at a time, with different color coding? Say the enemy holds 3 moves in its queue, at max. If it takes a moment to plan (though some may be more reactive than others), it may show a red outline for the next move it's planning, an orange, fainter one for the move after that, and a very faint yellow one for the third. This could get tricky, but if the entire FOCUS of the gameplay is on the prescience, a player would probably have to learn to deal with something like it...and if it's mostly just outlines of the AI, not an entire filled-in image, it won't overwrite too much. Also, a player could probably have the option of dialing down the prescience to not see so far ahead, simplifying things.

Well the thing is, I thought (from your last post, in fact) that the AI's decision-making process was represented by a deterministic finite-state automaton. Now, "uncertainty" can certainly be represented by a nondeterministic finite-state automaton in the shape of a decision tree (and getting the set of possible states reachable from each state gives us states to render ghosts in), but then how does the AI refine its decisions with time and actually commit to a course of action? Randomly choose a branch of the NFA? I'm sure there are answers to this stuff, I just haven't had a proper AI class.

Color-coding and fading of ghosts also sound like great ideas, especially since alpha-blending the enemy's model and rendering it an extra time is the most obvious way to produce ghost images. As to outlines, I think an outline would be too hard to see and a full image would probably camouflage too well with the real, physically "there" enemy. Do you think a wire-frame or untextured rendering of the model would balance those two display factors?

Also, two other things. Firstly, similar to how usually such games have a meter for using bullet-time, it may be there should be some sort of focus meter needed to use Prescience, that recharges over time, for two reasons. Firstly, it's less interesting if you're in God-Mode all the time. Second, in order for it to be prescience to be properly cool, there needs to be a point of reference of what it's like WITHOUT it. There's a difference in games between "you GET to do X, cool!" and, "you HAVE to use X, blegh." Think of the spider-ball in Metroid Prime games....and now think of if it was never NECESSARY to use it, but if there were just places where you had the OPTION to make things easier. How much more would you have enjoyed those tracks if they were all just shortcuts?

Yeah, definitely. Well, actually, I like the idea of limiting the player's use of Prescience, but not so much the idea of limiting it in terms of time. 30 seconds (or however many seconds you got in the various games) of bullet-time worked well because the player only needed/wanted bullet time for 10 seconds at once for most of the game. Unless under heavy, heavy attack players could easily stop and let their focus meter recharge without taking too much damage. So I'm not really sure what the limiting resource should be. Space, perhaps? That may sound weird, but it makes sense mathematically and philosophically to suppose an inverse relationship between the accuracy of a prescient vision and the volume of space envisioned. More space means more stuff in the space means more variables interacting in the different ways.

Something like that actually sounds like a good possible limiter. If the player used prescience on too many things and looked too far into the future, their view would become practically fogged with ghosts of prescient vision. This would mean that any sufficiently crowded scene would require the player to, to some degree, turn off prescience and play in real time.

So we could have prescience be a limited-use power limited by some sort of resource (I admit time is the simplest to program and the simplest for hypothetical players to understand), or we could also have prescience become self-limiting.

Another thing is that, with regards to meditating for platforming sections, I agree that you could have time actually continue while the player experiments with possible actions they could take, but in that case, you probably want real-time slowed down, with prescient-time continuing at the "normal" pace. Makes it easier to experiment, and otherwise, the prescient-actions would just be overlapping with the real-time movements of the platforms. You're not seeing "What I COULD do in the future," you're seeing, "What I COULD be doing right this moment."

Sorry, my descriptions of the meditating should have gone into more detail about snap-backs. When a player stops meditating, everything that "happened" while meditating is reversed, like rewinding time in Braid. The idea is that if the ghost player walks forwards for 3 seconds than takes 7 seconds to jump across a series of moving floating platforms, that experience serves as an accurate representation for the player of what would happen if they attempted that with their character's real body, and so when the player stops meditating they snap back to the original place and time at which they meditated. This would enable players to try something out in meditation, see it work, snap back, and immediately do just what they rehearsed for real. Players could also, of course, just try things without using prescience to test them, though this would probably result in dying more often.

BAH! More stuff keeps occurring to me! Two more thoughts for your consideration: Firstly, you may want to make it medium-range combat (such as God of War), because that would make it easiest to display the images showing what the enemy AI is planning, without obscuring what they're currently DOING. Hand-to-hand, there might be too much overlap. Second, just a thought, you may want to occasionally have bosses who are themselves prescient, and a (admittedly cheap) way to do this would be that, if you both have prescience activated at the same time, then they don't show any ghost images. If THEY have it on while YOURS is off, the cop-out way would be to reduce their own reaction time to practically nil (hey, so far as YOU can tell, it's almost like they knew what you were about to do). Obviously you either need to be good enough to beat them on a level playing field, or wait for an opportunity to strike while your prescience is on and theirs is off. Or possibly, there would be environmental factors you could take advantage of, particularly while you have prescience up ("So I see where that explosive fuel drum is GOING to go, and so do they....but they don't see where I'm about to SEND it").

All excellent ideas. Having the player fight with bladed or blunt weapons, particularly knives and short swords, would actually make a neat little homage to Dune's knife-fighting while also making prescience a bit easier to use. It also lets us use practically any setting we please. Everyone everywhere everywhen has something with a blade or heavy, blunt end that can hurt people.

You don't want to make it seem like the enemies are just morons and you're human and can predict AI, you want to make it seem like the enemies are themselves badass....it's just you're even moreso.

Damn straight.

### Re: Describe a game you think would be awesome

Posted: Wed Dec 24, 2008 3:37 am UTC
aleflamedyud wrote:Ah, state machines! I like it! To put the planning system into an FSA-based AI, we simply need to add an input symbol into the language the FSA interprets: null. The null symbol means that the player (or other AI) has done nothing that would make the AI change its plan, even though the player may have actually done something. Or do we need even that?

Well, not really. If you think about it in terms of state loops, essentially a number of states that naturally go to one another until you interrupt the stream with a new input to push it to a new path. We don't have to do that specifically, because we can design a single state to loop and it's the same thing, but you don't need a null input and it does let you see several steps ahead, at least until the new input is thrown in and everything resets.

aleflamedyud wrote:You're right about auto-counters getting unwieldy. For the moment, let's consider an X-Box 360 or Playstation-layout controller, since these are the most commonly used on consoles and PCs. We do want some mechanism to simply execute automatic combos or counters, but we want to minimize the amount of controls used to activate it. Why do we want it? Because it commits the player to a course of action that another player could see as a ghost in multiplayer mode. Or do we not want multiplayer mode?

I think limiting the player to pre-defined combo sequences is a bad design choice. In the long term this eliminates the need for prescience altogether, as experienced players would be able to predict the moves to follow given the initial action(they've seen it a thousand times before). While they might not be able to react to the initial input in time to act correctly, prescience would never help with that as it can't predict what input the player is going to make either. It's one of those things that helps beginners but destroys any long term depth to the system.

If you wanted multiplayer, the best thing I can think of is an entourage. PVP where each player has CPU controlled buddies. That way each player still feels super powerful as they have their prescience in play, but just like the Bene Gesserit powers, if you are aware of their nature you can be trained to resist them. hence the players cancel each other's powers out. Creating an intelligent and deep fighting system can make the PVP meaningful and fun, and the prescience is like a tool you can use against 'weaker' opponents. As i said, in this way the enemies are actually just as lethal as the players, but thanks to prescience telegraphing enemy moves, you can outplay them easily, which is why they have strength in numbers.

aleflamedyud wrote:A lot about controller config/fighting system

I'd envisage using the left stick for movement, right stick for camera, with a "center behind player" button, ditching lock on. I know that's a bit unwieldy as it means the face buttons can't be used with the right stick for anything common, but the advantage is the left stick can be used to direct attacks and the player feels more in control of the character. A lock on makes it difficult to work with fighting multiple enemies at once and makes for a 'duel' mentality, and I'd like to see a more freeform system, players surrounded by 4 or 5 enemies who are flanking and moving in to attack, yet being able to deftly deflect, block dodge and counter to manage each of them in real time all on their own.

If you ever played any of the Kengo series, they had an interesting system that revolved around 'stances'. Your talk of having an inventory of moves to equip reminded me of it. Different attack combinations are available from different stances, as well as different counters. No single stance will dominate against everything, but some do better against others, and worse against others. You can pick any 3 stances and load in any moves and counters you've "discovered".

We could apply this using the D-Pad or face buttons to switch between stances, since it's a function the user can afford to let go of either of the sticks to do. That frees up the shoulder buttons since you can have an all purpose attack which is modified based on which stance you're in, as well as a block which can serve as a counter on perfect timed instances. One shoulder button could be tap for dodge, hold for run, with the final one being free for something special. Maybe a modifier like duck, which could make for a slide attack with run, or low attacks. Or jump, like you suggested.

Face buttons could be used in the entourage mode for simple group commands, I guess. That'd be about how I'd do it.

Aikanaro wrote:May I make one other small suggestion regarding the prescience?

While in reality you're quite right about prescience and how it might be implemented, in a gameplay sense your suggestions tend to obfuscate the usefulness of the ability from a player's perspective. If you have an indication of what a character is about to do, and they do that consistently, it's something the player can rely on and use instantly. If you have multiple possibilities and some of them are false or misleading, it's initially going to be confusing and ultimately going to boil down to the player focusing on the most likely or consistent ghost. You did say the player would have the option of dialing it down if they find it too busy, but I think that resigns the point of having the multiple ghosts in the first place. I'm against stretching out the prescience for too long, because players can extrapolate future possibilities for themselves without having them shown on screen, and that isn't really a power. Having the short term definitive answer is useful and has some real benefits for a player, hence why I focus on that.

This mechanic is our hook, we want to exploit it.

Aikanaro wrote:Also, two other things. Firstly, similar to how usually such games have a meter for using bullet-time, it may be there should be some sort of focus meter needed to use Prescience, that recharges over time, for two reasons. Firstly, it's less interesting if you're in God-Mode all the time. Second, in order for it to be prescience to be properly cool, there needs to be a point of reference of what it's like WITHOUT it. There's a difference in games between "you GET to do X, cool!" and, "you HAVE to use X, blegh." Think of the spider-ball in Metroid Prime games....and now think of if it was never NECESSARY to use it, but if there were just places where you had the OPTION to make things easier. How much more would you have enjoyed those tracks if they were all just shortcuts?

That's an important design decision. I'm leaning towards making it a constant power in the title we're describing, but I've also thought of a bunch of other mechanics based on the same theme that would work better as a charged special.
The thing is, this power seperates the game from any number of 3rd person melee combat games. I'd want to design the power to be the kind of thing people would want to just keep using whenever it's available. Maybe you could have enemies that it doesn't work on, but if it's a passive thing that the player just has and doesn't need to worry about, it's not "you HAVE to use x". It's just there. By making it an active power the player has to engage, it really does become a 'have to use' power, because if it isn't, why is it even there?

Aikanaro wrote:Another thing is that, with regards to meditating for platforming sections, I agree that you could have time actually continue while the player experiments with possible actions they could take, but in that case, you probably want real-time slowed down, with prescient-time continuing at the "normal" pace. Makes it easier to experiment, and otherwise, the prescient-actions would just be overlapping with the real-time movements of the platforms. You're not seeing "What I COULD do in the future," you're seeing, "What I COULD be doing right this moment."

Here's a good spot where you could have a charged special. But all I can think of is the dagger of time stuff. Well, no, I could actually think of more, but it wouldn't be prescience so much as different ways of implementing time control.
I don't like slowing game time down as it's been done over and over, and it's probably my own personal bias that I'm tired of it. But I can't think of anything better right now.

Aikanaro wrote:Range, Bosses, Prescient battles, environmental stuff...

You don't want to make it seem like the enemies are just morons and you're human and can predict AI, you want to make it seem like the enemies are themselves badass....it's just you're even moreso.

I'd want the enemies to be pretty lethal. Everyone would be fragile, taking only a few hits to go down, including the player. This emphasises the power of prescience because it's what lets you defend and attack with such proficiency.

Once the player knows what attacks look like, the ghost really only needs to display about 1/3 of a second of the animation before the player is going to identify what's about to happen and react. I like your idea of them fading out as they go into the future and I think that's how it would work, because the player is only going to pay attention to the ghost until they recognise what its doing, and then they'll focus on timing their response to when the enemy actually performs their move. This would still be tricky for ultra close quarters, but when I say hand to hand picturing people being at least a few feet away from each other and moving in to attack, not standing toe to toe and trading blows. So I think we might have a bit of wiggle room to get the ghosts in there without making it too busy.

Bosses and environmental stuff I hadn't thought of yet.

### Re: Describe a game you think would be awesome

Posted: Wed Dec 24, 2008 10:28 am UTC
InstinctSage wrote:Well, not really. If you think about it in terms of state loops, essentially a number of states that naturally go to one another until you interrupt the stream with a new input to push it to a new path. We don't have to do that specifically, because we can design a single state to loop and it's the same thing, but you don't need a null input and it does let you see several steps ahead, at least until the new input is thrown in and everything resets.

Excuse my ignorance. I just got out of Intro to Discrete Mathematics and Computation at school, so my experience of FSAs isn't exactly a highly practical one.

I think limiting the player to pre-defined combo sequences is a bad design choice. In the long term this eliminates the need for prescience altogether, as experienced players would be able to predict the moves to follow given the initial action(they've seen it a thousand times before). While they might not be able to react to the initial input in time to act correctly, prescience would never help with that as it can't predict what input the player is going to make either. It's one of those things that helps beginners but destroys any long term depth to the system.

I think so too, which is why I thought that auto-combos should be available, but preferably built by the player and with definite tradeoffs. As I imagined the whole thing, it went like this: "combos" work as they do in martial arts, whether done automatically or manually. Chaining move X and move Y one after another ("X -> Y") in a combo means doing move X, then immediately doing move Y without the character ever withdrawing into a neutral position or stance. The idea is that to manually combo, players would have to hit the button(s) for their second attack fairly precisely at the end of the first attack; acting too slowly would lead oneself open, and button-mashing too quickly would result in inputs getting "dropped", ex: never executing an Oyizuki Jodan (face punch) if the player presses its button while their character does a Mayengeti (front snap-kick). So for the convenience of beginning players (and some experienced players with favorite combos), the game would let the player make auto-combos and equip them to a button or d-pad direction. The obvious downsides of autocombos are that they lock the character into attacking an unmoving target (ie: if the enemy dodges an autocombo, the character keeps attacking the same spot like a moron), and they make the character vulnerable to the prescience of others. Manual combo'ing would have neither of these problems, but would require plenty more skill from the player.

If you wanted multiplayer, the best thing I can think of is an entourage. PVP where each player has CPU controlled buddies. That way each player still feels super powerful as they have their prescience in play, but just like the Bene Gesserit powers, if you are aware of their nature you can be trained to resist them. hence the players cancel each other's powers out. Creating an intelligent and deep fighting system can make the PVP meaningful and fun, and the prescience is like a tool you can use against 'weaker' opponents. As i said, in this way the enemies are actually just as lethal as the players, but thanks to prescience telegraphing enemy moves, you can outplay them easily, which is why they have strength in numbers.

Sounds cool.

I'd envisage using the left stick for movement, right stick for camera, with a "center behind player" button, ditching lock on. I know that's a bit unwieldy as it means the face buttons can't be used with the right stick for anything common, but the advantage is the left stick can be used to direct attacks and the player feels more in control of the character. A lock on makes it difficult to work with fighting multiple enemies at once and makes for a 'duel' mentality, and I'd like to see a more freeform system, players surrounded by 4 or 5 enemies who are flanking and moving in to attack, yet being able to deftly deflect, block dodge and counter to manage each of them in real time all on their own.

Is there any gamepad control system nowadays that doesn't use the left stick for movement? But that scheme sounds basically good. My only quibble would be whether the right stick should control the camera "look/POV" fashion - so that pushing left makes the camera look left - or "camera" fashion - such that pushing left makes the camera orbit left around the player's character. I personally prefer the latter in third-person games and the former in first-person, but I understand that some people have vastly different preferences.

Anyway, what you said about ditching lock-on gave me a nice idea. We could just use a button to turn prescient vision on and off. When talking to NPCs in a non-combat area, the player could turn it off to avoid seeing superfluous legions of ghosts that do nothing but walk and talk (also lets these areas be a bit more crowded due to the reduced render/computation load without prescience). Got in a fight? Hit a shoulder/trigger button to turn on your vision and you're in business against any number of foes. Mind you, one obvious purpose of the lock-on button was to recontextualize controls from civilian mode to combat mode, but I suppose a switch can take care of that.

If you ever played any of the Kengo series, they had an interesting system that revolved around 'stances'. Your talk of having an inventory of moves to equip reminded me of it. Different attack combinations are available from different stances, as well as different counters. No single stance will dominate against everything, but some do better against others, and worse against others. You can pick any 3 stances and load in any moves and counters you've "discovered".

I've not played Kengo. I probably should the next time I can pick up a PS2.

We could apply this using the D-Pad or face buttons to switch between stances, since it's a function the user can afford to let go of either of the sticks to do. That frees up the shoulder buttons since you can have an all purpose attack which is modified based on which stance you're in, as well as a block which can serve as a counter on perfect timed instances. One shoulder button could be tap for dodge, hold for run, with the final one being free for something special. Maybe a modifier like duck, which could make for a slide attack with run, or low attacks. Or jump, like you suggested.

Face buttons could be used in the entourage mode for simple group commands, I guess. That'd be about how I'd do it.

I don't personally like the idea of an all-purpose "attack" and "block" control that adapts based on (circum)stance, but then again I tend to like action adventure > RPGs > versus-mode fighters (like Soul Caliber). Mind you, brawlers are great fun and fall into some weird hybrid between versus-mode fighters and action/adventure when you look at their control systems. I'd imagined this game as very much a brawler.

Honestly, the exact manner of allocating face buttons and triggers/shoulder buttons should probably be determined in playtesting. Mostly we just have to make sure that we don't make up more things for the player to do than exist conveniently-accessible controller buttons in any one control context.

Here's a wacky idea, why not switch contexts (meaning civilian->combat and vice versa) with the button inside the left control stick? Most games I've played seem to ignore its existence, but using it to trigger combat controls would at least theoretically let the player move into combat with less hand movement than using another actual button. And it saves buttons!
While in reality you're quite right about prescience and how it might be implemented, in a gameplay sense your suggestions tend to obfuscate the usefulness of the ability from a player's perspective. If you have an indication of what a character is about to do, and they do that consistently, it's something the player can rely on and use instantly. If you have multiple possibilities and some of them are false or misleading, it's initially going to be confusing and ultimately going to boil down to the player focusing on the most likely or consistent ghost. You did say the player would have the option of dialing it down if they find it too busy, but I think that resigns the point of having the multiple ghosts in the first place. I'm against stretching out the prescience for too long, because players can extrapolate future possibilities for themselves without having them shown on screen, and that isn't really a power. Having the short term definitive answer is useful and has some real benefits for a player, hence why I focus on that.

This mechanic is our hook, we want to exploit it.

I rather liked the idea of multiple ghosts, some of which were more definite than others. A scant future possibility could be an easily-ignorable ghostly outline, while an impending fact requiring the player's attention could look almost solid. The player wouldn't notice several outlines running around unless they wanted, and as their vision of the ghosts got closer to real time the ghosts would both look more solid and be far fewer in number. There might be 8 moves a character could make 10 seconds from now, but only 1 they're planning for the next instant, yes?

That's an important design decision. I'm leaning towards making it a constant power in the title we're describing, but I've also thought of a bunch of other mechanics based on the same theme that would work better as a charged special.
The thing is, this power seperates the game from any number of 3rd person melee combat games. I'd want to design the power to be the kind of thing people would want to just keep using whenever it's available. Maybe you could have enemies that it doesn't work on, but if it's a passive thing that the player just has and doesn't need to worry about, it's not "you HAVE to use x". It's just there. By making it an active power the player has to engage, it really does become a 'have to use' power, because if it isn't, why is it even there?

Seconded. Putting Chekhov's Gun on the mantelpiece is pointless if the player doesn't really have any want or need to take the gun down and shoot someone with it.

Once the player knows what attacks look like, the ghost really only needs to display about 1/3 of a second of the animation before the player is going to identify what's about to happen and react. I like your idea of them fading out as they go into the future and I think that's how it would work, because the player is only going to pay attention to the ghost until they recognise what its doing, and then they'll focus on timing their response to when the enemy actually performs their move. This would still be tricky for ultra close quarters, but when I say hand to hand picturing people being at least a few feet away from each other and moving in to attack, not standing toe to toe and trading blows. So I think we might have a bit of wiggle room to get the ghosts in there without making it too busy.

Bosses and environmental stuff I hadn't thought of yet.

I thought knife-fighting sounded cool and made an awesome homage to Dune. Just MHO, and it would give us an excuse to put the characters just far enough back from each other (out of arm's reach, roughly) that we'd have space for ghosts.