Tuesday, July 01, 2008

No More Hit Points!

I'm trying very hard to make this as short and simple as possible, but it's kind a complicated idea. Bear with me.

Imagine a game where a barbarian warrior rushes into a village of orcs. What we would play through would be a large number of combats with small numbers of orcs. First we would attack, then they would counter-attack, then we would attack again.

If it was an action game, there would be some maneuvering involved: stay out of arrow-fire, try to find nice ambush points, fight in some place where they can't come around to your back...

But compare this to a movie where a barbarian warrior rushes into the village of orcs. You get a level of variety and tension that you don't find in a video game.


The barbarian's sword is stuck in an orc, so he resorts to punching them in the face for a bit until he can pull one of their rotten swords from their hands. He grabs one by the leg and throws it into a pack of them. There is a moment of quiet while he is ringed by orcs, all of whom are armed but none of whom are willing to attack the scary human. He is partially entrapped by a thrown bola, he is injured in the arm, but he keeps fighting. He leaps from orcish rooftop to rooftop to stay ahead of the snarling pack, and throws one of them into a firepit. He staggers out and lights the whole village on fire...

Not only is the actual combat more interesting, but the context is more interesting. If the barbarian does go into a swing-swing-swing repetitive killing mode, the camera slowly zooms back and we see the landscape crawling with orcs, all slithering and leaping towards our hero. The whole hillside is on fire. There are a pack of child and woman orcs you accidentally run across. An orc shaman attempts to parlay... one of your party members is taken down by nets and carried into a hole in the ground...

In a video game, this is basically unheard of. In order for any of those things to happen, you'd have to cut out to a cutscene or, at best, have some kind of scripted situation.

Well, perhaps not.

...

In a game, the primary play loop (usually combat) has to be extended so it lasts a while. If the primary play loop is over too quickly, the player will be able to explore it too quickly and get bored too quickly.

So the game designers stretch it out by making it iterative. And they do that by giving everything hit points.

Hit points fundamentally make absolutely no sense. They're just an arbitrary way to make fights last longer, to give the player something to worry over. A resource for them to control.

In terms of gameplay, it works fine. But... does anyone else find it stupid and boring that the werewolf takes exactly thirteen sword hits and the soldier takes exactly three full-auto bursts from a machine gun to kill?

That's pretty... immersion breaking! The only reason we put up with it is because we haven't really seen many alternatives worth talking about.

...

Let's take it away.

Let's say... the player's primary play loop is not the moment-to-moment combat. Let's make that just visual fluff.

The primary play loop would be... quicktime events?

All of those moments in an action game where you're in the middle of a combo and they flash a button up on the screen and you're like, "X A Y, super-skull-cleave-combo!" and it's like "NO! You were supposed to press B to go to a cut scene and stab him in the head! Since you didn't, you lose half your life and he gains half his back!" and you're like, "what? I didn't even see the goddamn cue!"

Yeah, those. Except, you know, not shitty.

As the years progress, game designers are putting more and more "non-combat" stuff into fights. Quicktime finishing moves are basically standard fare these days, the equivalent to the "big glowing spot" that we had in the old days. Even when the player doesn't get to actually do something to cause that interesting event, it still happens: Kratos stabs the sea dragon through the tongue, nailing it to the deck so he can wade in and pull its eyeballs out. Really, that's not exactly tied to the "B" key, but with the number of times he does it, it might as well be.

What if we made those meta-combat events the real gameplay?

What if we collected them like cards in a CCG? What if the player actually has a "Stab tongue to floor" card, which can be played so long as you have a weapon and it's got a big freakin' tongue?

Anything - sea serpent, dragon, giant llama, mutant crab - anything that has a big tongue, you can stake it to the floor by using that card. What effect it has depends on the current level of tension in the fight. If the fight's been going on a while, it's a good finishing card, because it will automatically finish up with some gory Kratos-Rips-Something-Squishy-Out moment. If it's played in the beginning, it will very much upset the enemy as he rips his tongue in half trying to get it free.

It's more complex than that of course, on both ends. On the play side, you need to have some way to make it balanced and interesting and iterative. On the development side, you have to come up with some kind of meta-animation that can be applied to anything with a tongue...

But, on the other hand, think of what you save. No need for an advanced, pathfinding AI. No need for combat balance. No need for combat-worthy physics.

All of that is replaced by a large number of meta-events that make for an interesting set of fights.

Yes, you need to map out a "jump on his back and try to stab him through the neck" animation for everything larger than a breadbox. Yes, you need to make any given fighting arena know how to display "both of us run sideways until we clash together at the last moment". And, yes, you have to know how to do more interesting camera work. Since we're no longer bound to the moment-to-moment character actions, we can do zooms, pans, establishing shots, slow motion... any kind of fun, movie-like effect we please.

So, if you take this path (and assuming it's a visual game), you would need a kickin' animation engine. But you WOULDN'T need kickin' physics or kickin' AI. Is it more expensive? Less? Harder? Easier?

I guess that depends on how kickin' your systems are/would have been... I do know that you can't use most middleware - Torque cannot handle mixing animation, let alone animating things on the fly.

But... the real key here is that your meta-combat game has to be very interesting. You are creating an interesting, dramatic fight, it's true... but more than that, it has to be just as crunchy as the fight itself would have been. It has to have resource management, attrition, luck, counterattacks...

But it doesn't have to be boring.

Hooray!

17 comments:

Patrick said...

I'd like to see a game where the barbarian plays this meta-combat modules as flashbacks, and the gameplay is him reflecting on these as an old man, trying to come to terms with the act of genocide and innocent slaughter of a racial minority. There's probably need to be another play loop, the combat is precluded as a win, so the real win/lose is how you deal with these moments.

Craig Perko said...

It might be interesting, but it's not something I would build.

Marc Majcher said...

That is actually exactly what hit points were originally intended to represent. (See 1st ed. AD&D DMG, p. 61) But the bare numeracy of hit points, as implemented in pretty much every other game afterwards, yeah. Dead on. There are plenty of tabletop games that address this, but it's pretty difficult to handle in code.

Craig Perko said...

I don't think their intentions mattered at all: they may have intended HP to represent a complex, shifting battleground, but it never worked out that way.

While it is difficult to handle in code, so are many things we handle quite regularly... it is straightforward, although tedious to program.

Darius Kazemi said...

Your zoom-out-during-tedious-combat scenario was implemented by Sean Barrett in his Indie Game Jam game Very Serious RoboDOOM. (Don't bother downloading it, unless you want to track down a copy of DOOM 2 and rip the sprites from the WAD files.)

Tom Hudson said...

Using the current scheme of things, Epic games has something like a 70 artist : 14 programmer ratio. My impression from the industry in general is that character rigging/animation time is one of the biggest bottlenecks we've got - see AoC's problem with female combat animations right now for an example of this. I think your proposal would require more animations with far more complexity than AoC?

So it's a cool idea, but doesn't seem straightforwardly feasible. Could you do something semi-procedural, ala Spore? Or maybe if you emphasized physics even further, and did most of the animation with inverse kinematics? We're getting closer to physics-in-hardware with the recent purchases of both Havok and Ageia, and so there's some hope for getting the performance we need to get those looking good.

Craig Perko said...

In truth, I'm thinking about ditching the careful 3D animation stuff entirely, at least for early generations of this.

By the time the concept is really proven, Spore-like animations should be becoming the norm rather than the exception. I don't see how an animation revolution couldn't be just around the corner...

Although I bring up the animation difficulties, I don't really think it matters. In ten years, people will look back on scripted animations the same way we look back on sprite games... suitable only for platforms that can't handle anything better.

TickledBlue said...

I couldn't agree more Craig. Great post.

We've got 3D engines, physics engines and I'm hoping at some, not too distant, time in the future we have conversation and drama engines just as readily added to the mix.

Mory said...

First off, wonderful concept.

In ten years, people will look back on scripted animations the same way we look back on sprite games... suitable only for platforms that can't handle anything better.

Not everyone looks at sprites like that. I'd love to see a collectible-move RPG with basic 2D animations, which would totally bypass the "unfeasible" criticism. This is a fun gameplay idea, and it'd be fun regardless of what it looks like. I don't care if the game is in text, I just want to be able to play this kind of strategy.

Daniel Benmergui said...

You actually gave me a few ideas on different ways to handle combat for the Single Screen RPG competition.

I always find your ideas inspiring... thanks, Craig.

Craig Perko said...

Mory: I like sprite graphics, too. But there's almost nobody releasing sprite games on powerful systems - it's just not considered marketable. People think "sprites = gameboy".

Of course, I'm not really thinking "marketable" in the first place...

Daniel: Hey, yay!

Darius Kazemi said...

Disgaea (2, I guess) is the last game I can remember on a major system that was sprite-based.

Craig Perko said...

Nah, there was that one I never got to play. Penny Arcade did a comic about it, something about plants that eat souls... Odin's something or other?

Darius Kazemi said...

Oh yeah, Odin Sphere. Still, it's a Nippon Ichi game.

Mory said...

Handheld games still use sprites. Okay, it's settled. This game needs to be made for DS.

Raine said...

Just found out your blog. Very inspiring ideas, and I love the way you write. I'll have a look at the other entries. I've actually linked another post of yours in my blog, hope you don't mind. :)

Craig Perko said...

People are always welcome to read and link. That's kind of what it's here for. :)