Writings |
|||
in response to "On Splinter Cell and Other Linear 3D Games" |
 |
||
|
Tuesday, May 27, 2003
(in response to "On Splinter Cell and Other Linear 3D Games" at joystick101.org) It sounds like two issues are being addressed here: Linear level structure, and bottlenecks in gameplay. We have been training players for years that the way to get ahead is to explore the environment as much as possible. We give them bullets, powerups, invincibility, candy... there is no reason for them not to do so. If we put extraneous information into an area, the player will likely explore it all, and will expect to be rewarded. It has now become ingrained. It also makes non-linear level design very difficult. What is needed is an intense, pressing reason to stay on task. A fair criticism of the later Tomb Raider games would say that it rewarded players for lingering but provided little incentive to move on. The same can be said for FF8+, where players are best advised to linger as long as possible until their boredom compels. What we need is an old trick in new masks. Having finally experienced Starship Titanic last week, I can say the saving grace of the game (besides the excellent writing) was the use of a rather large, rather formidable bomb as a motivating factor. It made the experience much more pressing, and prevented the user from exploring the 3,000 or so empty rooms in the game. If this were Final Fantasy, every single one of those rooms would be a mess by now. Further examples of the classic timer revisited can be found throughout gaming. Rising lava tides, collapsing buildings, bombs, etc, all provide motivating factors. Crazy Taxi was only Crazy because of the pressure exerted upon the player by unruly customers. Given an entire world of bonuses and interesting crevices, the player remains focused upon the task at hand. Hence, the designer can fill the world without making every single possibility riveting. Crazy Taxi brings me to the next point, that non-linear levels require goals rather than distance. Too many designers feel that levels are completed when a player A: gets to a location or B: defeats a boss. This is necessarily lazy on the part of the designer, but it is quite overused. If the goal of the level was to "Damage enough buildings so that the power grid for the city goes down" or "Steal $2000 (from purses, wallets, bums)" or even "defeat enough of the invading hordes so they leave," then the design becomes necessarily freeform. Tenchu's level designs worked so beautifully because they had created an interesting system of goal-boss, rather than scroll-boss. This broke down in Tenchu 2 on the Tiger boss, when the player was given a non-linear level and asked to traverse it linearly. The uninteresting goal of "getting there" bogged down what otherwise was a great game. Some might even argue that the obsession with collecting items is in direct support of the "get there" goal, and without stating at the very least "complete your goal in X amount of time," the player's goal is best served by being as slow as possible. The other issue that you mention, bottlenecks, are unfortunately enough to kill the enjoyment of many games. Tomb Raider, again, was notorious for this. During Chronicles there were at least 2 places where the player was being asked to interact with a texture on the wall, a non-intuitive challenge which the engine was not up to (each took roughly 10 tries to achieve, once you knew it was there, and was as trivial as opening a sewer drain cover). The Cantina boss in Star Wars for the SNES was only killable if the player had acquired a sufficient number of weapons power-ups since the last death, a number unfortunately higher than the number of power-ups between the respawn point and the boss. Shooters solved this admirably years ago with the invention of the use-limited destructo-bombs. If a player was so inclined, they could use their get-out-of-jail-free card on a boss, swarm of enemies, or just anything that they so chose. Similar emergency cards to be played in games are the jetpack with low fuel, smoke bombs, "time freeze" devices, chaff, floating bonds, hint books, and so on. Another key to remember here is that you are not trying to kill the player, and you are not necessarily trying to challenge the player, you are trying to give the player an enjoyable experience. Many games have been successful without any way to actually kill the player. The ability for the player to get through your game without dying isn't a sign of too low a difficulty level, unless that is what you are trying for. I distinctly remember an article on the usability testing that went into the last Spiro release. Non-professionals were recruited and recorded playing the weeklys, and the points of death were noted. If there were any statistically significant bunching of deaths at any one point, that point would be smoothed out. Sadly, I have seen too many games where such usability testing is never attempted, or tried (an NPC in Starship Titanic can't respond to "bring me the bulb," "grab the bulb," "change the bulb," "take the bulb," but will release the bottleneck with "get the bulb"). Sadly, too, designers often believe that the cleverness of the player will mirror their own cleverness. As a legendary example, level 5 (6?) of X-Men for the Genesis could only be completed if the player walks up to and resets the console. No NPC tells the player to do this, and very few people figured it out. That's what happens when you are on a project for a very long time: you have an intimacy with the gameworld that no player will ever share, and which fogs your ability to understand the players' reactions. Repeat the solution to a puzzle enough times, and it becomes an obvious solution. "Of course you bake the sun talisman in the oven. Don't you get it? Sun? Oven?" As the point-and-click puzzler flashcards have fallen out of favor with most people, many games have incorporated that type of puzzle with all of the skill of a fat-fingered typist. They do so by having a one-use, specially coded move, either triggered by the action button or by another button the player is not likely to know about except when first configuring the keyboard. In the first case, the player has to accept that the "open doors and talk to people" button is really the "do what I want you to do" button, even though it hardly ever does. The second case is even less forgivable. Action puzzles in modern games are never pulled off with the same flair as Miyamoto pioneered them with. In A Link To The Past, Link's "push" action was almost as common as his sword swinging. In many other games, "push" is needed only once or twice, but the player will be required to both have mastered it and, more troubling, to remember that they can do it. No player of the aforementioned game could miss the fact that you could cut grass and bushes with your sword, and no puzzles requiring knowledge of this were presented to the player before it had been well established. In Chronicles, the player was required to make an acrobatic post spin to advance, without any prompting that the character was even able to do such a move. Movies have trained us to believe that retinal scanners can be fooled if you rip your victim's eyes out, but in MGS2* you are required to choke-hold your victim, then drag them to the scanner. The problem is, of course, that nothing in-game ever explains this choke hold thing, and nothing explains why retinal scans fail for dead or sleeping opponents. Apparently, you can't teach a new designer old tricks, as these problems with classic solutions are cropping up more and more these days. Perhaps it's not that we haven't advanced beyond adventure, but that we keep forgetting the lessons we've learned from it. -Chris Canfield *The Metal Gear series is a model for non-linear level design and gameplay, and should usually be looked upon for inspiration in design and execution. It is only mentioned here because of a particularly frustrating and memorable flaw. - Chris 4:30 PM [+] |
