The Preliminaries

Given some broad path to follow in designing a new game - one suggested by such things as which audience you intend to pander to, or what you are known for or interested in or think yourself capable of producing, or, perhaps least noble, what has been selling the best in recent memory - there remains the task of determining what the player is to be presented with in terms of actual gameplay.

Gameplay is ultimately and always the product of a system of rules. Some systems of rules are simpler than others. Chess, ironically enough, is an example of very simple rules. The rules exist on only a handful of low levels. They serve simply to define the game board and the pieces, and how the pieces move, and the condition for victory. All the rest of the intricacy of chess is a consequence of these rules and is not explicitly defined in these rules. It is strategy, and while it has rules of its own it is not part of the basic rules of the game. It is an epiphenomenon of the rules.

This reasoning applies in general. The rules are what make a game, and in the interest of accessibility they should be simple and well-defined. Then, built upon the rules, is game experience itself, which varies immensely depending on the type of game.

At this point you may be tempted to dismiss all this, saying, "I'm not trying to reinvent chess, I want to {make money, have fun, waste my own time} and this is too much to worry about." Don't. This is not necessarily a description of how to make a complex game of strategy, it is a way to make an enjoyable game period. Just because I am using chess as my example in no way implies that a well-conceived game has to acquire a reputation as imposing as chess. A game that maintains both simple rules on one level and sound gameplay on the other is accessible to casual dabblers and fanatics alike.

This is good. If you want to make money, it insures that your game appeals to the widest possible audience, with all the collateral benefits thereof. If you want to have fun, lo and behold you end up with a fun game. If you want to waste time, it should be the time spent playing that is considered 'wasted' (enjoyably) and not the time spent creating the game by and downloading and trying it by the player that is considered 'wasted' (because the game is a turd).


The Nitty-Gritty: Rules

Given rules as the foundation and the resultant gameplay, how can one approach the task of design? Once again, the broad ideas of genre and character should suggest what the player should be spending his or her time doing, that is the strategies the player will be executing. Some basic rules should also immediately come to mind. For example, a versus fighting game suggests two objects that maneuver about roughly as people would, in real time, launching attacks at one another until one is declared the victor. At this point, everything is still vague, and it is your creative vision that determines how you will proceed. You may have certain ideas for attacks or for players around which everything else will take shape. If you are emulating and building on the success of another game, you know more and can go right to ahead to adding and tweaking things.

In any case, once you have the basic parameters of your system down, the rules that give it life need to be formalized and the strategies need to be balanced among themselves. To illustrate the antipodes of these processes, I will continue with the previous examples - chess, and a very mature (rules-wise) line of versus fighters by the name of Street Fighter.

I will actually deal with a simplified subset of Street Fighter - the archetypical Ryu. For those unfamiliar with him, he walks toward and away from his opponent, can jump both straight up and in an arcing trajectory at or away from his opponent, and crouch. To attack he has an arsenal of punches and kicks of varying strengths and execution times. These are bound directly to buttons. He also has three basic 'special' moves, which are performed by a simple combination of joystick movements with the punches and kicks. They are the fireball, which is launched and moves horizontally toward the opponent, the dragon punch, which propels Ryu himself up and slightly forward to hit the opponent, and the hurricane kick, which sends Ryu spinning, leg extended, toward the opponent. There are also 'super' versions of these moves which are available only when a certain amount of 'super energy' has been accumulated by performing other moves, hitting the opponent, or being hit, and do more damage than normal. To defend himself he can block both high and low.

Everyone in Street Fighter has regular punches and kicks, special and super moves, and can jump, duck, and block, and they are all more or less on par with one another in terms of damage. These are the constants of the game. What the moves do, and hence when they are useful, varies greatly with ones' choice of character. Thus there are universal rules and universal strategies as well as rules and strategies specific to each player, and to particular matchups, that are due to the nature of each character's arsenal. (Don't let anyone tell you these games are stupid!)

Let's ignore the moves unique to Ryu and focus on these game universals. They are analogous to the rules of chess, but in this case the simplicity comes in the form of consistency and intuitiveness more than obvious simplicity. Unlike chess, where objects bear no resemblance to their real-world counterparts, Street Fighters behave more or less as one might imagine, without too many surprises. Chess' rules must be simple to compensate for their abstractness, while Street Fighter can get away with seeming complexity (or rather MUST be more complex) because it more closely emulates the way real objects behave. In either case, the key is that the rules are simple to understand, be it through straight simplicity or resemblance to something universally familiar.

It is also important to note that while Street Fighter may seem complex in terms of the number of fundamental game characteristics that exist, these characteristics are in themselves as predictable as the rules in chess. The difference was necessary to more convincingly approximate the activity of fighting. Blocking high will block all high attacks reliably, but not low ones, just as a bishop can move diagonally, unless blocked. Exceptions make sense and there are few of them; there is an inviolable lowest level of recourse that anyone seeking to understand the game can look to.

This level of design should be held sacred for just these reasons. Everything here needs to make sense so that the game overall has a consistent feel. A large number of special-case rules, valid only in certain situations, can often be more confusing than no rules at all. It is usually possible and very advisable to reformulate such systems in terms of a set of universally applicable and consistent rules. You may think it a question of semantics but it influences how the game is perceived and thus how easy it is to work with, both in developing it and playing it.

Also, consider a game like rock, paper, scissors where there are no absolutes. The game disintegrates into a crapshoot because strategy (rock crushes scissors but is engulfed by paper) is one and the same with the rules (rock beats scissors but loses to paper). Normally there should be a model built of rules that underlies the fact that rock beats scissors, but here everything collapses in on itself; high-level strategies are encoded directly into the rules rather than arising as a consequence of them.

Generally, when this happens the game suffers in terms of depth and playability. (This is especially evident in our extreme example.) This is fine when you want a tool to decide something randomly and don't have dice around, but not when you want to play a bona fide game. Games must be predictable on the fundamental level of rules and interesting at the high level of strategy. Don't take this as a condemnation of randomness, because the laws of statistics still provide predictable macroscopic behavior (though I as a purist always think of randomness as a degenerate shortcut to an interesting mechanic), but keep in mind the need to segregate how things behave in general from how particular things behave. A strategy may be flawed in and of itself, and should be fixed through rebalancing of the strategies, or in rare cases the rules themselves may be flawed and require revision, but it is important to make the appropriate type of correction or a vicious circle of corrections that screw other things up or ruin the feel of the game could easily develop.

Revising the rules themselves is the easier thing to do in any case, which is why it is often abused. I would advise you to resist the temptation at all costs, but if you are sure the fundamental rules are flawed the problem should be easy to spot. If you are trying to simulate something, look for something that doesn't really jibe with its real-life counterpart (most often an oversimplification of something that needs to be hashed out in greater depth). On the other hand, rather than fleshing something out in a more complete way, something might have a prohibitive amount of complexity and need to be simplified. All in all though, these two general directions are your only good options to move in. Again, resist the temptation to throw a finger in the cracks of the dam as it will come back to haunt you later.

Let me take an example from real-time strategy. Say you have trouble getting units to find a line of sight to attack the enemy. You can approach this problem in any number of ways depending on what kind of game you have in mind. If realism is a concern, you need to beef up the rules that determine how units find a line of sight to better reflect how this is done in the real world, and you will focus a lot of work on the relevant AI subroutine. If not, you can either eliminate line of sight and just have weapons go from attacker to target regardless of intervening objects, a major rules simplification that undermines realism and simplifies strategy. Either way, this is a good opportunity to look at the rules as the source of the problem because the problem is fairly universal.

An example of when not to monkey with the rules is when a particular strategy is unbalanced. The perfect example is Red Alert's tank rushing. Here, the strategy of building tanks in quantity was effective because tanks in Red Alert are hard to destroy, fast, and hit hard, especially against enemy base structures. No strategy really existed to balance the building en masse of tanks, because what was intended to was too hard to implement (walls were cumbersome to build and maintain and defensive structures were inadequate). You could only respond with a tank battalion of your own. Westwood responded by toning down armor in their next title, but also made changes to game logic governing unit production (multiple factories give no bonus). Neither of these changes really had the desired effect, where making other strategies potent enough to stand on their own would have.


The Nitty-Gritty: Strategies

This is a good time, then, to shift the focus to balancing strategies.

Westwood was apparently thinking the same thing, and decided to return the Engineer unit to it's status in the original Command & Conquer game, where it could capture an enemy building outright rather than just a severely damaged one. Unfortunately, in a heavily underbalanced game that also had easy means of sneaking into the enemy base, strategies revolving around the Engineer became dominating. The key advantages of the tactic were its decisively powerful results, (the enemy base could effectively be castrated and the enemy defeated just by capturing the Construction Yard), and the ease with which it was employed (in a game with a rather clunky interface for coordinating large attacks, this tactic involved only a handful of easily buildable units).

In evaluating and balancing the strategies available to a player, these two abstractions as well as a third are of use. The third is scope, and is a bit less straightforward, hence many overlook its existence. Each must be balanced to create an effective strategy that contributes to an enjoyable and diverse game experience rather than sitting around unused and irrelevant. If you could assign each attribute of each strategy arbitrary ratings, they should vary in the case of particular attributes but their average should be pretty constant across the board.

Let me define these a bit more explicitly. Power is simply how far a strategy goes toward winning its user the game. 0 is absolutely no benefit; 100% power means you win entirely as a result of the strategy. Ease is as simple, it is how much trouble you (don't) have to go to to use a strategy. 0 ease means the strategy is practically impossible; 100% ease means it is basically built into the interface for free. Scope is the least tangible of these abstractions, representing the diversity of circumstances under which a strategy is effective. If a strategy is inapplicable by definition or is applicable only in a situation not ever likely to arise in a game, it has 0 scope; if it is applicable regardless of the situation at hand, it has 100% scope.

Scope deserves more elaboration. In the real-time strategy context, scope represents how effective a strategy is against an opponent regardless of what strategies they are using. The tank rush had scope because it did what it was trying to do regardless of how the opponent tried to defend against it. This may seem like power or ease at first glance but it isn't, and treating it as such will derail attempts to balance the game. The problem with the tank rush needed to be solved by reducing its scope (creating effective defenses) and/or making the other strategies (most of which revolved around units and their combinations) more powerful and easy by comparison.

You can keep this mental model in mind and try to shoot for the second and third quartiles, the middle half of the continuum, for each attribute. For a rollicking, rip-roaring nuts game, try to balance high, and for a more sedate experience balance low, but beware that it is much harder to balance low, as this has the effect of highlighting any broken game elements like blood on a white tux. Don't necessarily keep a laundry list of strategies and cling to it like gospel truth, as this will encourage you to craft limited and limiting gameplay, but do try to explore your options within the game and think like you're playing to win.

Note that the definition of a strategy is quite broad. It includes everything not a part of the universal rules, and encompasses every choice a player of a game can make, including to chain together other strategies. Strategies exist at every point in the game, on low and high levels - low level strategies just above the rules, high level strategies that involve the combination of lower level ones.

An interesting sidelight of this way of describing strategies is that it applies equally well to actions in systems other than game's framework of rules. Consider applying it to the actions of characters in a story and how they influence the overall plot. Balanced characters tend to make for an interesting story, but on different levels - often supporting characters revolve around a main character. Et cetera. Food for thought.

In closing, I can only proffer a few caveats. First, don't constrain your thinking about rules and strategy to the genres where they are more obvious. The principles that were easiest to illustrate here in one genre are equally applicable and powerful across the board. Second, in crafting a game, make sure to treat people as what they are - ingenious problem solvers, not drones with inexhaustible patience for the mundane. Give your audience something of substance and they will be appreciative. This applies equally well to simple, pick up and play games as it does to Axis & Allies. And of course, make it fun.