If you consider gameplay in terms of making progress through the game, you find that it typically exists on multiple levels. A very crisp example is Archon or similar ‘battle chess’ variation. On the high level, you play chess. ?On the low level, you have action battles to determine which of two pieces is victorious.
Chess gameplay is ‘higher’ because actions here map on to large steps of progress toward winning the game. Battle gameplay is ‘lower’ because individual actions culminate in the resolution of the current battle, which can only result in as much progress as a single step on the high level. As we can see, in this notion of ‘levels’ of gameplay, higher levels are associated with broader concerns and lengthier steps of progress than lower levels. Another common differentiator is as follows: the high level is associated with ‘strategy,’ meaning how to win the war by choosing battles to fight, the mid level is associated with ‘tactics,’ meaning how to win the current battle by choosing which moves ?to attempt, and the low level is associated with ‘execution,’ meaning accomplishing the current moves successfully.
Archon’s two levels are very visibly distinct, essentially separate games connected with a thin thread. It is more common that multiple layers are integrated into the same mode of interaction. In Black Ops, regardless of whether the player is concerned with protecting the headquarters (high level), guarding a particular doorway (mid level), or tossing a grenade at an approaching enemy (low level), they are in the same engine with the same view and controls. In these examples, player input always maps ?on to low level actions: moving an avatar, using a weapon or tool, etc. There is no input that accomplishes mid or high level actions, e.g. no ‘Increase research capabilities’ or ‘Defend HQ’ button. Instead the player accomplishes higher level play through a series of lower level actions.
In this sense, the higher levels are the goals of the game and lower levels are their subgoals. As this demonstrates, gameplay does not always decompose neatly into three levels. The reality is typically messier, more stratified, and evolves as the player’s goal chain gets simpler or more complex during the course of play. For example, if a player is attempting to complete a mission objective in Thief but decides they will need more water arrows to pull it off successfully, this will spawn a temporary goal tree under ‘Complete Objective’ whose head node is ‘Collect Ammo’ and which itself could be fairly complex before it completes and goes away.
However, the high/mid/low grouping often makes sense due to characteristics commonly found in each level. High levels are often characterised by designer-imposed rules about how the game is won (victory points in Civilization) or what happens as you proceed through the game (levels and story in Thief). The high level is frequently one-way progress; the player marches forward until the game is won. In comparison, the low level often appears chaotic and entirely under the player’s control ?if seen from the perspective of any higher-level goal or structure. From moment to moment, the player might take any available action: jumping, shooting, switching inventory. The mid level is where we see the most ambiguity of authorship. The objectives of Thief and the scoring system of Black Ops are examples of designer-imposed mid-level structure, but in these games the player is given ample freedom to author their own sub-goals, as the Thief water arrow example illustrates. The mid level is also where we see the most structured repetition in games, as play here tends to arrange itself in loops. Mid-level loops are one of the more versatile and nuanced tools in game design, and I’ll ?cover them further next month.
What is the purpose of this type of ?academic approach? Formal analysis can help ?us understand some common reasons game designs succeed or fail. If you are working on ?a risky, unconventional concept, it might be possible to apply these patterns to give your game more structure and your players a familiar sense of control and progress. They can also be referenced while generating new game ideas, although following the patterns too closely can result in ‘paint by numbers’ design, assuming there is no innovative kernel to wrap the structure around.
Formal analysis can also be used as a tool to generate experimental concepts when applied in a way that defies expectations. For example, it tends to be the case that players are given increasing flexibility as you proceed from high to low levels. What if this were shuffled? What would a game be like if the player had lots of control over the high level – how the story played out or something similarly unexpected – but very little control over the mid level, such ?as how individual encounters resolve? Does such a game already exist? Perhaps the player is giving orders on a mobile phone to a bumbling henchman? Turning these tools on their heads is a creative exercise that can produce the ?initial threads of interesting game concepts.
Randy Smith is a game designer and principal at Tiger Style, which has its first game, Spider, in the App Store now. See and follow all his columns on his topic page.