Principal principles: a primer for adventure game developers?
First posted on 25 April 2009. Last updated on 25 April 2009.
There is a feeling that adventure games are not as well constructed as they are used to be. It is a common topic discussed online in message forums and newsgroups, where those of us who cut our teeth on classic LucasArts (or Sierra or Legend or Infocom or whichever) era games bemoan falling standards. Recently, Steve Metzler of metzomagic.com has written an article series titled "Where Have All the Puzzles Gone?" (1,2) which compares puzzles new and old. The article concludes that game designers need to make an effort to revisit the classics and learn from them. Certainly, no adventure game of recent years has, for me, felt as well constructed as Maniac Mansion: Day of the Tentacle—a LucasArts game and classic released more than a decade ago!
Those who make the games are under a lot of pressure. Sales and margins are not what they have once been. Trying to persuade games to work across a range of operating system versions and drivers and graphics cards takes resources away from the game itself. What players want ranges widely from the gloss of leading games in other genres to more of the tried and trusted. Is it worth wasting resources on aesthetic niceties? Games can be popular despite having the failings cited by Metzler. Adventure gamers complain more about getting dizzy from panoramic panning (in a first-person perspective) than about games not being well crafted. There is plenty of help available such as walkthroughs or cheats to smooth over these holes in gameplay. So why do we need to worry? I believe a few well crafted games will raise the respectability of adventure games outside the core adventure gaming community, and that is important for the adventure game genre in the long run.
What can adventure game designers learn from other games (both good and bad)? There have been some good writings from game authors and academics about design of adventure games that give some guidelines. A good article is "The Craft of Adventure" (3), written by Graham Nelson of Inform. It is written from a text adventure standpoint but has a lot of relevance to graphic adventures too. Another good article is "Why Adventure Games Suck" (4), written by Ron Gilbert of Monkey Island fame. Like Nelson, Gilbert has raised some good comments about the genre, albeit it is framed from a perspective that adventure games are officially dead. I want to stress the aesthetics of game design—perhaps not saying anything really new, just re-emphasising it.
I want games I can play—no pixel hunts, no confused sets of hotspots, no time critical puzzles, no arcade sequences, no sounds puzzles to force you to remember musical tones—at the least, unless there are alternative solutions. I am not colour blind so I do not mind personally colour based puzzles, but to someone who is colour blind these can be hard to impossible. Making commercial games must be quite a headache with so much to consider.
For me, the story must drive the game, the puzzles in it, and the way the action moves forward. In a bad game, the puzzles feel shoehorned in, as if the developer is squashing them in where there seems to be room. You get what Metzler refers to as "artificial obstacles", where the developer triggers an unrelated event by the player picking up an item, such as a screwdriver, (because it is needed to progress further in the game). This is bad because the player is reduced to randomly trying acts that are no longer part of the story. A similar situation is if the player has both a sword and a knife but only the sword (or the knife, but not both) can be used to cut a rope. This may be convenient for the developer, but not for the player. The best games for me are those where the game worlds and game characters feel like they do not just exist to provide the puzzles. An example of a game world with depth is that in Journeyman Project 3: Legacy of Time, where there is a distinctive sense of oneness and unity.
A game must be completable from any point during play, without needing recourse to saved games. This covers the need that games must not have any dead ends, such that a player must be able to trust the game to not leave the player in an unwinnable state. For many games I do not trust that far, perhaps because of other rough edges—and so rather than waste time down a dead end I often first check with a walkthrough or cheat. For me, the price to be paid is that this slightly diminishes my game playing pleasure. Completability also covers the need that clues to solve a puzzle must still be available when you are trying to do the puzzle. In some games, if you blink you may miss a vital clue for good. My preference is that puzzles are locally solvable—that is, the items and clues needed are not too far away, rather than a near endless trek across the game world.
A game must have consistency. An example from a classic game (it is too easy to pick holes in recent games) is Broken Sword: The Shadows of the Templars, where the left click is used to interact with characters and items and the right click provides amusing descriptions of them (the game's dialog is brilliant). However, there is a place in this game where the right click is necessary to advance the gameplay. This is bad because, up to that point, the player has been building up a mental model of how the game works and how the game world works. Breaking this mental model for the convenience of implementing the game harms how immersive the game is.
A game must always reflect the player's actions or changes in the game world. For example, a hotspot must not just appear on the ladder when you are asked to rescue the cat in the tree. If a new hotspot has appeared, then there must be a visible change to the scene. Often, dialog does not reflect what the player has done. An example is from Broken Sword: The Shadows of the Templars where you can insult a non-player character without knowing what you have just said really means. Even when you do eventually find out what you have said, clicking on the now insulted character still gives a "what did I say" response. The Immortals of Terra: A Perry Rhodan Adventure even manages a rare twist on this kind of slipup by recording actions in the game journal that the player has not yet performed!
A game must kind to the player. Suppose a character has asked you for a sword. You go to collect the sword by pulling it out of the rock or out of the umbrella stand where it has been hidden or whatever. A logical game dictates that you must then be able to give the sword to the character either by talking to the character or by clicking sword on the character. You do not want a dialog in which the character still asks you for the sword or in which you are told that nothing has happened when you click on the character with the sword. If a game is kind to the player, then a player is more comfortable playing the game and feels that the player's decision is being respected.
A lot of adventure games start off well—the developers have great intentions, and they are very familiar with the genre for which they are developing. When the game is released, there will be plenty of discussion about how it may be even better. (Discussing games is legitimate fun.) However, it is what comes in the middle that ultimately counts—how well the developers from writers to testers navigate the hazards of budgets and deadlines to reach quality.