You can now get an eBook version of this document that features expanded text and a new bonus essay. Check that out on the books page. You can get the whole reverse design series up to this point.
I propose a project to reverse-engineer all the design decisions and processes that went into classic videogames. There are a lot of great game design books and resources out there, but most of these focus on the fundamentals of game design. Those books are essential, of course, because it's impossible to pursue any vocation without good fundamentals. I find, however, that there is a strange over-emphasis on those fundamentals and a lack of attention paid to games that are obviously classics. To me, the study of game design theory and fundamentals is incomplete without the study of game design history focusing on the classics. It's like a musician studying scales but not Beethoven.
Accordingly, this post begins a series of large, textbook chapter-length essays on the masterworks of videogame design. I suppose I should lead off with an apology: I am not a an accomplished game designer. As such I will try not to theorize about anything which is not specifically evidenced in the design of the game under analysis. I don't claim to have any special insights, but with enough math, graphs and carefully reasoned (and textually faithful) examination, I think it's possible for anyone with enough time and resources to pick apart the best ideas present in classic game.
Each page comes with a kind of "executive summary" at the top which details the specific game design components disucssed there. This first page also has an overview, which you can find below. This overview details my general thesis for the design of the game (Final Fantasy 6) and a larger picture of how all the design elements fit together to form a whole. The last page contains an itemized list that acts as a quick reference for the design lessons taught by the game.
As for the game in question, I chose Final Fantasy 6 because, like it or not, it is certainly a classic. Seventeen years, two remakes and a lot of discussion prove that. The game exhibits quite a few design ideas that ought to be documented as fully as possible, so that we might understand why the game has endured for so long. I also chose FF6 for a practical reason: there is an enormous amount of information available on the game. I would not have been able to do this project without the incredible data sources provided by Kris DeHart, Jeff Chan, MasterZED, Meeple Lard, and many more including a few IGDA colleagues. The full list of citations are on the last page; thank you all very much.
So what makes FF6's design so special?
The foundation of Final Fantasy 6 is its large roster of playable characters. Everything else in the game was designed to accomodate this. A number of current and former Squaresoft employees have spoken to the fact that the game's fourteen characters were the first invention of the team. Of course, it's very likely that some work had already begun as this was happening, because many design features from older FF games carried over. What we're interested in, however, is what's different about FF6. What makes the game special? The features unique to FF6 are all design ideas created to accomodate a roster of 14 playable characters. There were three primary design ideas:
(1) The diminution of character classes
(2) The emphasis in design on entire dungeons as the meaningful level of content, rather than individual battles, forcing the party (or multiple parties, as happens) to play the "long game."
(3) The centering of the narrative upon (a) a villain who operates like a main character, and (b) a central theme rather than a central plot.
The Diminution of Character Classes
The most historically significant design concept is the first one, the diminution of character classes. Meaningful character/job classes were one of the most durable ideas from tabletop RPGs, appearing in virtually every RPG that featured more than one playable character. These character classes force the player to make tactical decisions. Do I use a fragile mage for more damage? Do I use a heavily armored knight to outlast the opponent? Do I bring two healers, sacrificing damage to ensure victory in the long term? Even in the case where a player is able to choose party composition, or is able to change the job classes of party members on a whim, there are always tradeoffs to be made for the use of one job class instead of another.
In FF6, these tradeoffs largely disappear. As the game wears on further and further, the significant differences between job classes disappear. Talented mages, which in other games had low defense, can often wear heavy armor in FF6. Rugged fighters can learn magic, and use it well. In fact, because of point #2--the "long game"--it's almost essential that every character class learn magic. Every single character in the game has a nearly identical pool of health points (for survival) and magic points (for casting spells), from the beefy martial artist to the ten-year-old artist. By the end of the game, and without over-leveling, it's possible to beat the final boss with any combination of characters, with ease.
There are two principal reasons for the diminution of character classes. The first reason is that there is little value in a game with fourteen characters, only a few of which are actually usable. There are only so many roles a character class can fill, and as such, the difference between those roles would come down to effectiveness. Why use Strago when Celes is so vastly better? As far as stats and armor go, Celes and Strago aren't markedly different, if they're at the same character level. By the end of the game, there's not a huge amout of difference between the best "light" armor of Strago and the best "heavy" armor of Celes. Their magic stats, HP, and MP are all closely comparable. Celes has an advantage in weapons; Strago has an advantage in his special ability. A player who likes to use Strago out of pereference will find no gameplay reason why he should not do so. Any group of characters can form the main party, and do well at it; and this is because of the diminution of character classes.
The second reason for the diminution of character classes is that, at several points, the player must use more than one party, and for the final dungeon must use twelve out of fourteen possible characters. (Actually, the player can use fewer, but why would they?) Accordingly, if every character class were subject to unique, complicated rules, restrictions and tactical drawbacks, it would be inordinately difficult to organize those multi-party adventures. In order to be useful to a player who has limited time and attention, the characters must be different but not too different; and thus we arrive again at the diminution of character classes.
Information about the diminution of character classes is frequent throughout this work, but is particularly concentrated on pages 4, 6, and 7. The dimunition of character classes is actually a design trend composed of many other, smaller design decisions, so naturally there are numerous things to discuss, and no single page that summarizes it--other than this one! Those smaller ideas consist of standardization of stats, non-linear distribution of gear, and an emphasis on a universal skill: magic.
The Long Game
The discussion of mult-party dungeons actually brings us to the second element central to the design of FF6, the use of dungeons as the level of content delivery. This terminology is a little confusing, for its brevity. Basically, this means that individual battles in FF6 are not the really significant unit of player experience. Rather, dungeons are constructed as a whole, of which individual battles are rarely an exceptional part. In other RPGs, even in other FF titles, there are usually a number of special encounters that require extra attention and strategy. The Mecha Head, the Great Malboro, the War Mech, Mega Tonberry, etc. Final Fantasy 6 doesn't really have kind of encounter; certainly not in any dungeon. Most encounters within a dungeon can be met with the same tactics.
Interestingly, if you look at the average encounter across every dungeon in the game, the data reveal that dungeon encounters are the same in several facets--in every dungeon. This is not to say that every dunegon in the game is the same, because they're not. Rather, using a statistic that measures the survivability of players and enemies, it emerges that the average survivability of the standard groups of enemies in various dungeons are very close to one another. Moreover, there is no clear upward or downward trend for dungeon difficulty based on encounter-level survivability. With characters of level equal to their opponents, it should usually take about the same number of attacks to finish an encounter. It's also true that players rarely have to change their strategy when facing these encounters in all the various dungeons.
What players have to deal with is not the inflating durability of enemies, but rather their own survivability, which becomes increasingly precarious, but not because of pure stat inflation. The change in dungeon difficulty arises out of the increased need for players to heal their party, more often and more fully. In the second half of the game there are no purchasable items that will fill this need, and so the player has to rely on the same MP pool for healing that powers their attack magic. Items that restore MP are either scarce or expensive. There are battle strategies that can refresh this MP pool, but they're not universally effective and leave the characters open to additional (and often exponentially more deadly) attacks. Thus, a player's strategy for a dungeon isn't based on individual encounters, but instead focuses on the party's ability to make it from beginning to end, or from save point to save point.
This is what I call designing (or playing) the "long game" in a dungeon. Especially when it comes to the final dungeons of FF6, where multiple parties are in play at once, this kind of meta-battle strategy becomes especially important. The player needs to organize multiple parties not just by power, but by ability to survive the attrition that is thrown at the characters by otherwise regular encounters.
The primary examination of the "long game" is on page 7, although there is a lot of relevant information on page 6 as well, and a few relevant facts on page 3. The constitutent design ideas are the standardization of stats, the sorceress problem, the caster's advantage, and a regularization of a meta-stat I call real durability.
Organizing a Charming Chaos
The final design principle that makes FF6 special is the way its story is organized. It is no easy task to write a plot for a dozen or so characters who existed in substantial form even before the script was begun. For another thing, it's not easy to write for so many characters when the total length of the dialogue runs less than 20,000 words, as it does in FF6. For a third and even more onerous hurdle: the second half of the game is non-linear. And all of this has to take place in such a way as to make the game not just readable, but playable!
To address this, the design team made two interconnected decisions. The first decision was to have the villain act more or less like a main character. Kefka is on screen an awful lot; he speaks more words than almost all the party members. Most of the sub-plots and quests feature Kefka, or his actions. Most of the party members are affected directly by Kefka's aggression. The other decision was to center the second half of the game around a theme (hope and resolve in the face of catastrophe) rather than a plot. The quests in the second half of the game are largely personal, rather than epic macguffin retreivals or decisive military action.
In order to make this happen, the second half of the game needed its own organizing factors. One of those factors was the expert use of NPC dialogue. (In truth, both halves of the game are well done, in this regard.) FF6 uses its NPCs to masterful effect, reinforcing the themes of the game not just with individual bits of dialogue, but with a shifting trend among an entire population. This population also serves to organize gameplay by use of what I call ironic communication, or NPC irony. NPC irony involves veiled directions from the designers couched in narrative format, so as not to break the spell of the imaginary world. A full survey of the categories for NPC communication is on page 5.
Another design decision that affects the organization of the game is the way that the non-linear half of it is constructed. Being that there is a definite goal (to defeat Kefka), how can all the disjointed quests form an artistically complete whole? Final Fantasy 6 solves this problem through the use of metonymy, to create what I call a metonymous sense time and completeness. (I admit that this is hard to sum up quickly, but you can see the full definition on page 3.) Distributed throughout the second half of FF6, in no particular order, are dozens of pieces of gear and magic necessary for the final dungeon. Also scattered are the characters that the player needs to reclaim. The strange thing is that because these things are not in a linear order, they have metonymous power. That is, because every dungeon, chest, and encounter can be equally important in terms of the gear and magic they give, or the characters they return to the party, those dungeons are really memorable. This design feature might be entirely incidental to a series of dungeons that had a different objective to fulfill, but nevertheless it's an important feature of the game's design. This is certainly the most esoteric aspect of the design. For greatly elaborated detail, see page 3.