1. Introduction: Interacting Mechanics and Holistic Design
A good strategy game is about relationality: game mechanics and systems that interact in deep and variegated ways, enabling player choices that affect many interrelated aspects of the game at once and lead to a fun gaming experience. Relationality can only be achieved through holistic high-level design, not by tinkering with mechanics and systems one at a time. Stardock seems to have largely ignored holistic design as a means to relationality, which is, in my mind, the most important reason why Elemental is not as much fun as it should be. I fear the recent avalanche of well-meant feedback may accentuate rather than alleviate this problem, which is why I suggest a more holistic (as opposed to patchwork Frankenstein) approach to future design decisions.
This post is divided into two parts. First, I will provide motivation for the discussion by showing why relationality augments user choice and how holistic design contributes to relationality. Second, I will give examples of disconnects in the interaction between Elemental's mechanics and systems and suggest possible amendments. I will not attempt to outline a macro-level design for Elemental. It is my belief that macro-level design is best left to either one person or a handful of people in close communication with each other (i.e. not random forum posters), because relational design is only possible given a thorough holistic vision of all mechanics and their common purpose.
I will use the interrelated concepts of element, mechanic, and system to describe a design. Elements include such basic building blocs as base gildar income, gildar income multiplier, and current gildar balance. Mechanics, by contrast, refer to the wholes constituted by elements and their relationships, such as the gildar mechanic or the arcane research point mechanic. Systems denote bundles of mechanics that form relatively independent aspects of the design, such as the economic system or the research system. There is nothing absolute about this threefold framework, of course, but it does provide a basis to work from. The best strategy games dilute the difference between systems to a point where it is difficult to tell where one system ends and another begins. Solium Infernum is by and far the clearest example of this on the computer, but I will use the less brilliant Civilization 4 as a reference point due to its familiarity.
2. Motivation: the Frankenstein Design
In his review of Elemental, Tom Chick described Stardock's design approach as follows:
– – Stardock doesn't work their way up from a basic design. They craft parts and piece them together and create something not unlike a Frankenstein monster. It evolves and changes and reacts and morphs and mutates and lurches and sometimes spits up on itself or farts noisily. It can be scary and ugly. It can be awesome.
In other words, Stardock's designs are "patched together" from separate parts, taking interesting-looking mechanics and systems and mashing them together with only fleeting concern for their mutual interaction. As Chick alludes, the Frankenstein design is not only Stardock's weakness, but also their strength. Nothing is fixed, rendering the creative process open to user feedback.
The problem with the Frankenstein approach is, however, that it does not facilitate breadth of user choice in the end product. When a system – say, the economy – is designed largely apart from a game's other systems, it will not be as integral to the other systems as it would in a holistic design, leaving the user with fewer ways of exploiting their interactions. In Civilization 4's relatively holistic design, the economic system and the research system are very tightly bound to one another at the macro level – money and research points come from the same basic resources (commerce or direct bonuses, both produced by population) that can be directed one way or another by the player. In Elemental's Frankenstein design, some buildings, tiles, and city traits produce money (or money multipliers), others produce research, period. The difference in scope of player choice is obvious; a Civilization player chooses, turn by turn, whether the commerce mechanic should be part of the economic or the research system, whereas an Elemental player chooses once whether they want money or research out of a city and are stuck with the decision for the rest of the game.
In addition to the breadth of player choice (how many alternatives exist at a given juncture, such as deciding between research and money every turn), holistic design of relational systems also contributes to the depth of player choice, or the player's influence over the meanings of mechanics and systems. In Civilization 4, the player may expend a Great Prophet (great people system) to produce a holy place in the holy city she acquired upon being the first to research the relevant tech (research system). The holy place provides money (economic system) based on the number of cities that have been converted to the religion in question (religion system). This mechanic has the potential to significantly alter the meaning of the religion system and the religion spread mechanic, as it allows the player to partially substitute religion for economy. The substitution is not perfect, however: the holy place produces money, which is not analogous to commerce because it cannot be directly used to obtain research. This design contributes to deeper user choices than simply giving +1 food for casting a spell on a city. It allows the religion system to take on a different guise, changing the meaning of the religion spread mechanic without overriding the commerce mechanic. Such depth would not be possible without the relationality of the economic and research systems.
To sum this section up, holistic design renders relational systems and mechanics possible. Relationality, in turn, enlarges the scope and depth of player choice. Since player choice is probably the main component of "fun" in strategy gaming, any serious strategy game should be designed holistically. The most important difference between Civilization and Elemental is that between holistic–relational and Frankenstein-individual design.
3. Disconnects: Here Be Research
Elemental's research system is a prime example of design disconnect. It affects the other systems exclusively by enabling them (research treaties, you get treaties; research spells, you get spells) or by giving gradually increasing modifiers or bonuses (get Swords of Plus Fifty Percent, a gold mine, or the Spell of Plus One Food). The same is true the other way around: money and materials buy the champions and buildings that enable and contribute to research, but that is all. Map tiles and buildings that provide research are only for that purpose, and no connection to the other systems can ever be made – you would be a fool not to build on lost libraries and arcane temples as soon as possible because the benefits are always the same and always important, effectively eliminating player choice.
The research system is also internally disconnected and non-systemic. Many of the research mechanics, such as the arcane and tech research point ones, do not interface much. The branches of the tech tree are separate. Individual elements, such as individual techs, interface with maybe two or three other elements, and their meanings are very stable with respect to each other. The Tech of the Sword of Plus Fifty Percent is exclusively a Tech of Swords of Plus Five Attack – notwithstanding champions –, since no tech (as far as I know) affects peasants' attributes or changes the one that is used to determine attack strength.
There is an unlimited number of ways in which the research design disconnect could be fixed. The interesting and depth-inducing way, however, is not to add a building that both produces one research and gives a fifty percent gildar modifier, or a tech tree branch that gives peasants plus one strength per breakthrough. The player should not be a spreadsheet optimizer, but a creative decision-machine who reacts to circumstances by playing the systems and mechanics against each other. Some examples of interesting, relational alternatives at different levels of the design might include:
a ) Allow quests to unveil techs for research at some branch of the tech tree. Give these special techs special effects. For example, allow one of them to warp your genome in a way that renders your children physically frail (may give birth to random monsters, guaranteed low str, dex, and con) but magically enhanced (a random trait from a list of things like +1 mana regen in battle, a one-shot global spell that forces peace between players, and two extra hand slots for equipment).
b ) Allow the players to purchase tech resources from each other by relocating arcane temples or ancient libraries to the plot of the recipient's choice. This would require sufficient diplomatic capital, relationship, and money (or other tradable). Note the difference to tech trading: both players lose something.
c ) Make some techs mutually exclusive, even between major branches such as Warfare and Civilization. Make the player choose between Nationalism (armies gain morale and lose two points of intelligence [make these matter]) and Open Canon (caravans to other players contribute to research instead of food). *cough* My political views show through here, sorry about that.
The above examples are, of course, not constitutive of a relational framework by themselves. They cannot replace a proper holistic design. It may also be that they are unfeasible to implement due to AI or other similar considerations. I nonetheless hope my examples confer a flavor of what I mean by relationality and the general direction in which I think Elemental's design should be taken. Relationality as such cannot be impossible to implement, since it has been done many times to different degrees.
4. Conclusion: Where Hope Lies
I think there is hope for Elemental. At the moment, however, relationality and the resultant breadth and depth of user choice are not present. In light of some of the discussions on the forums, it also seems that relational design is presently impossible to mod in or implement, since several crucial logical operators are missing from the XML parser. Changing this should be a high priority after bug fixing is complete.
I believe the most important thing Stardock should change, however, is its design philosophy. Rather than resorting to Frankenstein design and creating more spreadsheet-y and linear-feeling mechanics, future Elemental design should be holistic and aim at a relational, network-like end result. Ideas should not be implemented on the coolness factor, but only after their capacity to introduce interesting interactions between different parts of the design has been weighed. Vanilla Elemental should work as systemic a whole as possible, and the cool but non-relational ideas should be left for modders to implement.
If a new essence mechanic is implemented, it should not be singularly because the old one "feels bad". It should be because the old one could not be reconfigured in a way that integrated it sufficiently with the game's other mechanics, whereas a new one is judged better in this regard. New mechanics and systems should be designed to interact meaningfully with the rest of the game.
