GDC 2009, Day 2

I'll break with tradition in my Game Developer's Conference coverage and simply say "no comment" on all the sessions I attended today, except for Hideo Kojima's, creator of Metal Gear.

Hideo Kojima

Several people told me they didn't like Kojima's presentation at all and they found it very long and boring. While he should have removed the first 10 minutes which was spent entirely on saying hi, repeating the title of the presentation several times, and apologizing and making jokes about the presentation...once he actually gave it, I thought it was worthwhile.

Kojima's talk was called "Making the 'impossible' possible." He showed a side view of a ground plane and a square obstacle, saying that this square represents any obstacle in life that we might encounter. How do we overcome it? Mario would simply jump over it. Solid Snake would tuck and roll to get over it. But what if the obstacle is taller? Mario could still get a running jump over it, but Solid Snake cannot jump that high. What can he do?

This seems like an impossible task for Snake because the obstacle is simply too high for him to overcome. But Kojima, displaying the classic traits of a visionary and of lateral thinking, devised several possible solutions. Snake could get a box which would act as a step to reach the top of the wall, then climb over. He could use large balloon for lift and hover over. He could fill up the area on one side of the wall with water, then float to the top. Kojima rattled off at least 6 solutions to the "impossible" problem. As you are about to see, this really captures the essence of what it is to be Mr. Kojima.

The First Metal Gear, on MSX2

In 1985, Kojima was given an impossible mission: create a combat game using the MSX2 hardware. First, Kojima told us what he considered the basic elements of a combat game. It has...some sort of background. It has a character the player controls. It has some enemies. How many? At least 4 seems like a good idea. Next, it has bullets from both the player's character and the enemy characters. This, to me, seems a reasonable list of things for his combat game. A bit short of a list, if anything.

The problem is that the severe limitations of the MSX2 hardware make even the few elements we've covered so far, kind of impossible. The hardware can handle only 32 sprites and a background. Furthermore, it can handle only EIGHT sprites at a time if they are arranged in one horizontal area. He showed us an MSX2 version of Gradius (called something else, I forget, but it's obviously Gradius) and showed what happens when the ship, the options ("options" are the little blobs that surround your ship and help you), the bullets, and the enemies all line up horizontally and number greater than 8. What happens is the sprites flicker on and off like a buggy mess. They made it so that all the sprites in a given horizontal line greater than 8 share the burden. As opposed to a 9th sprite being always invisible, for example, they would toggle around which one got the invisible treatment rapidly, so at least you had some idea. But in this shipped game, it was still a buggy mess.

The plot thickens though. If you want a sprite (let's say your player-character and enemy characters) to be in color, then you had to layer the sprites and use two per character. Imagine the sprites that might occur in one horizontal line, perhaps on the ground plane. The player character(1) in color(2), with enemy characterA(3) in color(4), along with enemy characterB(5) in color (6), meant you have only two left over for bullets! So if you shoot 3 bullets, one will be invisible! Not to mention the bullets of the enemies!

Mission Impossible: Build a combat game using the MSX2 hardware.

If this obstacle (this impossible mission) were a wall, Komjima said it was just too high. He would have to sneak around the side of the wall and redefine the problem to something that was possible. What if instead of hiding, you had a "hold up" game? You can't even shoot, but if the enemy gets close, your character surrenders. This is feasible, but unfortunately a terrible game design. Ok, what if it's mostly about running away from your enemies? Also feasible, but fails to capture the original intent. The movie Rambo: First Blood Part II was popular at this time, and it seemed a lot more popular of an idea to be a "hero" like that, rather than a guy who runs away. Hmm, what else? What about hiding from enemies? Seems better than only running away, but still missing something.

Infiltration. Yes that's it, Kojima thought. What if you try to avoid enemies and hide so that you can infiltrate an enemy base? That sounds more heroic than the other ideas and it is feasible given the limitations of the hardware. Kojima mentioned that the game became somewhat of a puzzle game because you were presented with a single screen (no scrolling) and had to avoid enemies (the puzzle) to progress to the next screen (the next puzzle). Anyway, he completed this game and it was called: Metal Gear. Mission complete.

Second Metal Gear

The next mission given to him was to make an even better infiltration game (we can call it stealth action by now) using new, more powerful hardware. Here we come to a key idea of Kojima's talk. Imagine this mission represented as a tall wall that he has to somehow climb. (He snuck around the last wall by creating a new mission, but this time we'll try to actually climb the wall.) Kojima said there are really main ways to climb the wall: use game design as a "ladder" or let technology raise the floor for you so it's not so hard to complete the challenge. On the first Metal Gear, notice that he used game design to make the solution even remotely possible. This time, he would rely on technology to make a better game because maybe he won't have such ridiculous limitations like only 8 sprites in a horizontal line.

Surprise change! News from his commanders states that the new hardware is taking too long to come out. His new mission is to make a better stealth action game using the SAME hardware. Oh my. He decides to rely entirely on game design as a "ladder" to climb this wall. No help from technology as hoped.

In Metal Gear 2 (also on MSX2), he added several elements. The line of sight of enemies was expanded to be a cone of sight. They could also hear footsteps now. An even bigger chance was the shift from the one-screen-at-a-time puzzle game aspect, to requiring the player to think about much larger maps. The player now had a radar minimap in the upper right corner showing the location of enemies on other screens. Finally, Kojima added a new mode to the enemy AI where after they know you're there, they go to a search mode with a much more aggressive movement algorithm. This gives the player a chance to avoid them if they can stay undetected for X time. Also, while the enemy is in this alert state, you lose the use of your radar. Together, all these changes made for a much more interesting stealth game.

3D Stealth (Metal Gear Solid)

Next mission: Make a 3D stealth action game on the MSX2 hardware.

Kojima said this wall is so high, it cannot be scaled. It is simply not possible.

Revised mission, years later: make a 3D stealth action game...on the upcoming PlayStation1 hardware.

Ah, now this is possible. Here, he said game design only had to take him a relatively short distance because the amazing advance in technology made the job so much easier. Out of everything he said in the entire talk, I thought this was one thing where really sold his own efforts short. I actually think Metal Gear Solid 1 on PS1 is an incredibly well-designed game. Anyway, the technology allowed for many things:

• Voice acting, localized in 6 languages
• 3D camera, player can zoom in and rotate using sniper rifle or binoculars
• Different camera views while crawling in duct or hiding against a wall, etc
• 3D real-time cut scenes

The game was a world-wide hit, so everyone demanded a Metal Gear Solid sequel.

Metal Gear Solid 2

Mission: make an even more realistic stealth action game on the on the SAME hardware.

Kojima didn't know how to do this and I don't blame him because MGS1 is about as far as you can push the PS1 hardware.

Years passed, and mission revised: Make an even more realistic stealth action game on the new PlayStation 2 hardware.

This time, the boost in technology was large again, though he would also use game design as a ladder to truly climb this wall. Weather effects (technology) added a feel and ambiance that was a real improvement. Higher polygon models made the whole game look much better. He added new game mechanics such as the ability to hide in lockers, for guards to go on alert when they see other dead guards, the ability of the player to hide dead guards in lockers, and a new wound system.

MGS2 was a world-wide hit and everyone wanted a sequel.

Metal Gear 3

Mission: create an even better stealth action game using the new super-powerful, next generation game system.

Revised mission: that hardware is taking too long, create an...evolution of stealth action using...the SAME hardware.

Kojima thought about what they could do differently that would not require more powerful hardware, so the game could be on PlayStation2. He would not be getting a hardware boost, but he STILL planned to get a technology boost anyway. This time the boost would come from a new game engine that was more powerful and had more features. This would get him part of the way up the wall, but he needed to get the rest of the way with game design. He said that previous Metal Gear games had been in closed environments rather than open ones, and they had been in man-made environments, rather than natural ones. This was very much on purpose, but what if they went a totally new direction and used open, natural environments? Could the game design actually make this possible? It looked like the new software engine was now capable of these environments, as long as the game design can make it work.

He added new mechanics of gathering food and eating (survival in the jungle, after all), a new cure system, and a camouflage system. Gameplay in open environments was different, but it still worked, he said. Although he didn't get the boost from new technology he hoped for, the game design was able to create the "ladder" needed to overcome the challenge of making a better (and different) stealth action game.

Metal Gear 4

Kojima wanted MGS3 to be the last. He said he always pictured it as a trilogy and he wanted to be done with Metal Gear...so he could do something else. I almost feel sorry for him. He said the problem is...MGS3 was a world-wide hit and everyone wanted a sequel. So he created a mission for himself, this time: Create the ULTIMATE stealth action game, so that he would not have to make any more stealth action games. Do this using the unlimited power of the new monster hardware that is codenamed PlayStation3.

This was his plan, but eventually his team got ahold of PS3 hardware and the super enormous boost up the wall he predicted from this magical new hardware...turned out to not be anywhere near as high as imagined.

Revised mission: Make...a better stealth action game using the actual power of the PlayStation3.

He would need a game design ladder again because he felt the wall was very high to make a worthy sequel, but the increase in technology--while big--was not big enough by itself. He created a whole new style of gameplay involving "situations," as he called it. You are now a rogue operative in the middle of a warzone, and you often encounter soldiers on BOTH side of the conflict, fighting each other. You can kill all of them. You can kill one side and befriend them. You can kill neither side and sneak by while they fight each other. There are many ways to deal with the situations you encounter. There were also other new mechanics such as octo-camoflague and your notoriety score which allowed you to more easily befriend one side or intimidate the other.

The Next Metal Gear

Kojima pointed out that he never completed that mission of ULTIMATE stealth action game. He had to back down from that when the PlayStation 3 proved less powerful than he originally thought. He is still under pressure to make sequels and still thinks his "out" is to make the ULTIMATE stealth action game. That is what he would like to do next, and he basically announced to us that he will make another Metal Gear.

Kojima said that he has noticed while all game companies overcome challenges by some combination of ladders from game design and floor-rising from software technology, that Japanese companies tend to rely more on the game design solutions while American companies rely more on raising the floor by using advanced software solutions (akin to boost he got from the new game engine on Metal Gear 3). At first I thought this was some sort of open trash talk on his part, saying that Japanese are more clever and Americans brute force our way through mediocre designs. But then I realized he didn't mean this at all. He just meant that American companies have developed some pretty damn good technology while Japanese companies have accepted lower technology with design work-arounds. He didn't mean this as a value judgment, just his personal observation.

He then said that what he would really like to see is combining the strengths of both approaches. What if he could raise the floor incredibly high by using amazingly advanced software solutions AND make a very tall ladder with a very strong game design? How high of an obstacle could he overcome then? Perhaps...the ULTIMATE stealth action game? (So he can stop making them!)

Then Kojima revealed that he and Kojima Productions are actively looking--at this very show, on the expo floor--for any technologists who would like to join his company help him create the amazing software technology the West is known for. In effect, this was all a recruitment pitch to find awesome programmers that will help him design the game he wants to make. Incidentally, I'd like some programmers like that too. ;)

Kojima's closed by zooming out on his series of diagrams to show how far his avatar had climbed in scaling wall after wall after wall of challenges. And he then pointed to that very first wall "make a combat game on the MSX2 hardware" that we labeled "impossible." Imagine if he had given up at that point, rather than finding a way to make the 'impossible' possible. Kojima said, "If I had given up then, there would be no Metal Gear series...and no Splinter Cell series either."

(That's his version of a sic burn, but Clint Hocking we love you anyway so don't take offense.)