Thursday, July 17, 2008

Key Features of a True Strategy Game

Something that's been bothering me for several years came back to me recently.

Gamasutra featured an interview with Soren Johnson, previously a designer for the last two Civilization games and now working on enhancing the gameplay of Spore. I enjoyed most of the interview, but one thing kept jumping out at me: I kept reading references to the "strategy" of real-time strategy (RTS) games.

There's no such thing. If it's real-time, it's not strategic.

RTSs are almost entirely about tactical action. They don't include gameplay elements that enable player actions at a strategic level. There's no coordination of multiple theaters of operation; there's no personnel management; there's no contingency planning for the gain or loss of key transportation nodes; and so on. The usual "send units out to collect resources until the source is depleted" feature is marginally strategic in that it's about locating, defending and exploiting resource nodes. But it's not fully strategic because resources don't persist over time -- they're used up so quickly that they're more for short-term tactical advantage than long-term strategic dominance.

Even the rare RTS such as Star Wars: Empire at War that does offer a strategic feature like semi-simultaneous theaters of operation (in this case, different planets) undercuts that gameplay by making it real-time as well. This demonstrates a fundamental misunderstanding of what "strategic" means. When even the high-level gameplay is real-time, when you're given insufficient time to gather data, assess its value, and weigh alternatives, then that gameplay is no longer about planning. What's left may be a kind of gameplay, but it's not strategic gameplay.

I've discussed the definitional differences between tactics and strategy in multiple essays already, so I won't belabor that point again here. Please visit my Strategy vs. Tactics and Combat Modes and Player Ranks in a Star Trek MMORPG essays for an in-depth discussion of what these terms of art actually mean.

In brief, though, here's what the Wikipedia entry on military tactics says: "Tactics should be distinguished from military strategy, which is concerned with the overall means and plan for achieving a long-term outcome, and operational art, an intermediate level in which the aim is to convert the strategy into tactics."

There's nothing hard to understand about the distinctions between these levels of the application of force. So it's very strange to hear an experienced game designer like Soren Johnson -- who presumably understands the difference between strategy and tactics -- repeatedly talking about RTSs as though they give those gamers who long for truly strategic-level gameplay any chance for that kind of activity. Calling these things "real-time strategy" games is misleading.

Under other circumstances I might push this as just a technical gripe, similar to fussing about the way people misuse the word "aggravate" (which means "to worsen") when they mean "annoy." But there's no point in ranting about some mere error in word usage. Besides, Johnson clearly does have pro experience with the strategic level gameplay in the Civ games.

The reason I'm highlighting the meaning of the word "strategy" is not to complain about words; it's to point out that there seems to be a game design opportunity here. We talk about RTS games as though they're strategic, when they're actually not... so why not make a game that really is about strategic play?

To some extent, of course, we already have those: as the Wikipedia entry on strategy video games points out, they're the hex-map wargames and 4X-type turn-based strategy (TBS) games from whence RTSs sprang. But lately even the TBS games (Civilization, Master of Orion, Galactic Civilizations, and to some extent Sins of a Solar Empire) aren't much about actual strategy; they're more about moving units around in a tactical fashion than about making strategic-level decisions. These days about the closest we come to strategic decision-making in games is allowing the player to set a single overall tax rate and then allocate that tax income among military, scientific, or cultural production. While from a strategic perspective it's better to have that than nothing, it's not much better than nothing since for most players it's a "fire and forget" function -- it's not something that can really be counted as active gameplay, which is what sells games.

I think there's an opportunity here for some developer to make a game that highlights actual strategic functions as active gameplay, that's a different kind of fun because (unlike most games) it rewards competent high-level thinking. To highlight this as a key product differentiator, and to avoid making the game overly complex, some of the tactical features people take for granted in TBS and RTS games today would need to be minimized or even eliminated.

Some readers may be shaking their heads at this point and thinking, "typical noob designer mistake, putting the setting before the gameplay." Not so. I'm not looking to define a set of game rules solely according to whether they fit some finicky technical definition of strategy -- that wouldn't be a game, it would be a military strategy simulator, and that's not what I'm after here. What I'm trying to do is take a closer look at the real-world definition of strategy to see if there are any aspects not currently being exploited as rules in a game. My goal is not to try to fix existing strategy games, but to explore a more rigorous definition of strategy to see if it can be mined for new gameplay ideas... or, if not new, at least ideas that have been neglected for a while.

I haven't finished my thinking on this subject -- this essay (like my earlier essay on a "Living World game") is just a first draft of some ideas in the direction of something different. So don't expect The Answer here.

That warning given, here are some design concepts I'm currently considering for a strategy game that would actually support strategic play.

1. Consider eliminating individual units altogether.

Having units always seems to incline game designers to tactical-style design thinking driven by the conceptual simplicity of unit-against-unit gameplay rules. Once you decide to have units that players can move, it's just too easy to start building your gameplay around assigning strength and defense values to those units and making up rules for how they interact... at which point you've got another game of tactics.

Instead, allow players to place directional pointers (arrows) or goal markers (stars) as guidance to strategic forces. This would let players suggest the high-level direction of movement of resources without shifting the focus of the game to personally managing each movement of specific units.

If you just can't bring yourself to try something so radical, then consider allowing units, but treat them as large-scale objects with very limited movement and behaviors. Rather than being specialized units with this combat strength or that movement rate, units would mark locations of force projection.

The idea here is to take the focus off of the management of individual units, which elicits tactical thinking, and base gameplay instead on mechanisms by which players can define and perceive patterns of force spread over relatively large swaths of space and time. The goal should be to allow players to make choices similar to that faced by Gen. Eisenhower in the European Theater of Operations of World War II: support the "knife-like thrust" into the heart of Axis territory favored by Field Marshal Montgomery, or order massed movement along a "broad front" (which is what was actually adopted)? That's the level of strategic thought that would be fun to enable in game form.

2. Focus diplomacy features on strategic-level concerns.

Diplomacy as usually implemented in TBS games is actually more of a grand strategy feature in that players are able to negotiate directly with opponents who control all aspects of all possible assets. In effect, players (including opponents, whether human or AI) are empowered to act as the supreme rulers of the faction they represent. While that's a kind of very high-level strategy, it's maybe a little more high-level than the textbook definition, which is what we're considering here.

A truly strategic-level diplomacy game would impose some constraints on the power of players to determine the selection, collection, and allocation of resources. This might take two forms. First, players would need to establish and maintain positive relations with multiple representatives of the dominant cultures in a theater, rather than with a single fully-empowered entity. Part of the challenge would be to acquire strategic assets (access to key resources or intel) through various means, such as playing representatives against each other, offering inducements, making shows of strength, or even veiled threats. The challenge here would be to try to achieve strategic ends through non-military channels.

Second, the role of strategic leaders is to develop plans for achieving the overall vision of those who do control the entire institution. So strategic-level diplomatic gameplay might also involve taking very high-level direction from the leader or leaders of one's faction. For example, the player might receive direction from civilian commanders to take and hold -- by whatever means necessary -- some strategic resource (e.g., a large refinery or a FTL jump point) and the transportation lines leading to it from secure bases. It would then be up to the player to determine and set in motion the best means of accomplishing this task, whether that would be negotiation, force, or some combination of those forms of persuasion. A strategic leader might also be faced with the interesting challenge of working with political interests in his own faction in order to obtain desirable internal resources such as negotiating authority, advanced weapon systems, or troop increases.

3. Movement of resources should always take time.

This significantly increases the value of planning-type thinking, which is key to good strategy. Since resources need time to move from where they're produced to where they're needed, part of the strategic game is to anticipate both production and application requirements, as well as transportation and distribution capabilities. In other words, a strategic game is also in part a logistical game.

During a game, new sources of raw materials or constructed goods will be located or captured, and existing sources will be depleted (which should be rare), replaced with better sources, captured by the enemy, or destroyed. Similarly, different resources will be needed in different locations due to advances or retreats, and to the types of challenge present at different locations. As all of these events occur, players of a strategy game will need to modify their transportation models continuously so as to insure that every need is fulfilled as accurately and as soon as possible.

Transportation systems also factor into this part of strategic gameplay. The number and location of routes in the player's transportation network, the maximum speed and loading of the mobile elements on that network, and the placement, number and capacity of storage depots, all determine the player's distribution capability.

In a way, maintaining and enhancing this distribution network (and trying to foul up the distribution network of enemies) could be seen as a defining element of a strategic-level game. This doesn't mean embedding a version of Railroad Tycoon; letting the player lay down specific transportation lines is too low a level for a strategic game. (See item #1 above.) But it does mean having gameplay that enables players to demonstrate an understanding of logistics. Functions allowing the player to invest in different kinds of transportation modes (where different modes have both advantages and disadvantages), to assign priority weightings to different resources, and to establish the general number and capacity of storage depots are examples of gameplay elements that could be found in a true strategy game.

4. Make personnel assignment part of the gameplay.

Something that's a crucial factor in strategic decision-making that might also be fun to replicate in a game is the fact that no one can do everything. A strategic leader acts through subordinates who are entrusted with performing the operational design tasks necessary to successfully achieve your strategic goals.

So I'd like to see a game in which you have a vast number of subordinates, some of whom you'll need to select to fill the roles required to meet your strategic needs. I imagine this might work as one window where you can define strategic goals and the functional roles necessary to meet those goals, and another window containing a tree of names showing the current hierarchy of assignments. In the tree of names you could click on any name to bring up each person's "service record," and then assign individuals to desired roles. Each subordinate would have enough freedom (defined by the game rules) to translate your strategic goals into tactical actions for their own subordinates to carry out. They could even provide evaluations of their fellow subordinates, or suggest (competing) ideas to the player for which strategic actions to take.

Notice that this game feature has several ramifications. Imposing a time delay from assigning a subordinate to getting the results of the tactical actions they order, as well as simply getting a subordinate into the right location in the field for their mission, creates an additional kind of logistical problem for the player to solve. And if subordinates are implemented as non-player characters who can react to your decisions affecting them (and perhaps even the different factions they might represent), choosing which NPC to assign to a high-prestige role could be its own fascinating political subgame.

To return to the example of Eisenhower during WWII, the complex struggles among Montgomery, Bradley, Patton, and other top generals for Eisenhower's favor are legendary. Even in a relatively simple form, a feature allowing for such machinations could add significant depth to a strategy game.

5. Minimize the tech tree.

The tech tree as usually implemented in turn-based strategy games is probably not right for a true strategy game. What a tech tree feature really amounts to is the imposition of policy controlling scientific research. Some games (like Civilization) implement this in a kind of unit-like fashion; the player is able to choose specific technologies to research. Other games (like Master of Orion II when the "Creative" perk is not selected) only permit the selection of broad fields of research; the specific technology learned from a particular field is randomly selected.

In both cases, however, the player over time is able to specify something like a research policy. (I say "something like" because in practice most games implement so few technologies that by the end of the game the player has usually learned them all, which means that no real policy preference can be expressed.) While I've personally had a lot of fun with that kind of thing, it strikes me as somewhat suspect as a feature that's appropriate for the strategic level of action. Eisenhower, for example, didn't get to tell Rolls-Royce to do the R&D that led to the Merlin engine used to power the P-51 Mustang fighter. Instead, he defined high-level mission goals; some subordinate translated that into equipment procurement requirements; and the Allies got whatever was available to meet those requirements.

Something similar might be appropriate for a strategy game. Rather than exposing the elements of a tech tree to the player as a gameplay feature, if a tech tree is appropriate at all for your game, keep it internal and let the game rules determine when some new technology becomes available. The player would then need to react appropriately to this development -- perhaps to exploit it, perhaps to prevent an opponent from obtaining it, perhaps (as with the atomic bomb) both. (Note that a game in which new technologies can occur should take care to define only technologies that have strategic value. There'd be little point in implementing a "Coffee Grinder" tech, for example, but even the mere threat of building "Ballistic Missile Shield" technology could have major strategic effects.)

The one caveat to the advice to minimize the tech tree might be to implement some version of the "military-industrial complex." In such a game, the player -- in addition to other forms of negotiation -- would need to manage relations with production sources. Working with equipment suppliers to negotiate supplies of vital goods or services might be an interesting gameplay feature in a strategic-level game. (Or it might be lethally boring, or just not a good addition to more obviously strategic play. If there are enough other features, this one might not have enough value to be worth implementing.)

In any case, as with management of individual units it's probably best to forego the tech tree feature in a game that aims to let the player focus on strategic decision-making.

There are some other features that a truly strategic-level game would likely need in order to be fun. After all, if it's going to be a game, the player has to be allowed to do things!

I'll probably return to this "Real Strategy Game" concept in a subsequent essay as I give some thought to what those other player activities might be. For now, these suggestions seem like a reasonable starting point for thinking about what such a game ought to look like.

No comments:

Post a Comment