The Stances Battle Engine concept is a RPG battle engine concept based most strongly on the idea of preparation. The name itself comes from martial arts stances, but the system is expanded in my head to include a new way of envisioning magic abilities, again based on preparation, but most specifically the way in which your preparation affect the range of your mage's abilities.
Ideally, the engine core would be based on a hex grid. The hex grid is important in particular for martial stances, but it is also useful for magic, as hexes can approximate circles reasonably well (many mythologies and superstitions around magic view the circle as a fundamentally pure thing, but more than that, it suffices to describe the range of a magician's abilities in terms of circles).
We'll start the discussion with martial stances, as the concept really started there. In creating the Demonsword Project - Draco setting, I created a magical weapon called the N-sword, which was several swords bound by magic to act as a single sword; when one was held, the others would hover in formation around the wielder. "In formation" is a critical point here: these swords had multiple formations, and each formation was fundamentally different. Certain formations would automatically defend your rear, while others would attack twice, for instance. In my mind, the N-sword is a tactician's weapon; the more adept a wielder is at determining the most advantageous formation, the more effective he is in combat.
That concept, once generalized and formalized, becomes the basis for the Stances system. In martial arts, there are fundamental ways of positioning and moving your body that alter what areas of your body you can guard, and in which directions you can attack. These range from very specifically forward stances, which are meant to put the most power into your blow (usually, you attack while moving into this stance, but we will get back to that), to defensive stances where you are expecting attacks from the side of your body, to responsive stances where you can most easily move to another stance depending on the circumstance. Taking that basic into a more extreme level, it is not hard to create a system of martial arts, based on a hex grid, where stances provide natural and meaningful impact on the battle.
Fundamentally, the Martial stances system alters your character's relationship with the surrounding hexes. The six adjacent hexes can be characterized by the following criteria:
1) The character can/cannot attack this hex from his current stance
2) The character can/cannot defend against this hex from his current stance
3) The character can/cannot move instantly into this hex from his current stance (as a reaction ability, not a normal action, which is unrestricted)
4) The character can see the square clearly, can only spot large movements (periphery), or cannot see it at all.
This system is fundamentally most interesting when a small number of player characters are fighting a large number of enemies, or vice-versa. In a two-man duel, both characters will tend to pick a stance that focuses on the front; this is only interesting if the system goes into particular depth and nuance as to the consequences of each stance, and it requires a very fine-grained turn structure, where making an attack may leave a character without the ability to slip into a defensive stance.
It is worth a dalliance to say that that system, if combined with a deep study and interest in the martial arts, would be worth having as its own game. I would gladly encourage anyone so interested to get involved in such a project, but it is not the project currently under discussion.
The interest here as a game design conceit is that stances are a fundamentally restricting factor; compare this to the following three tactical RPG conceits:
1) Character facing and position is not considered in calculating hit chances or damage
2) Any character is vulnerable to attacks from behind
3) Any character is vulnerable if attacked from two sides
When character facing is not considered, the battle engine has no restrictions on player tactics. If fighting multiple enemies, the correct solution is based on reducing the enemy's damage dealing ability and nothing else.
When characters are specifically vulnerable from behind, or made vulnerable to being flanked, this restriction narrows the number of correct tactics. In particular, enemy units become living features of the battle map; a unit in one position is twice as dangerous as the same unit in another position, and a character in one position is in twice as much danger as the same character in an adjacent square, when fighting the same units.
Martial stances take this increase to the variability of the field and double down on it. Individual units have vulnerable zones, but also enhanced-danger zones. They may also be able to take reaction movements, invalidating an opponent's entire turn by stepping out of range, or making a counter-attack action that may change the tides of battle. However, critically, all this information is available and predictable. In particular, any game that makes use of these martial stances systems would by necessity have to be consistent, allowing the player to recognize at a glance the style of the stance and intuit the other abilities of a character.
At this point we have to take a step back away from focusing solely on martial arts and expand the same system to generic combat. For example, a wild animal has natural stances, but has not studied martial arts; they will have vulnerable and danger zones, escape tactics, and so on. Similarly, archers or other ranged units have vulnerable and danger zones, but those danger zones are largely not adjacent to them. For ranged units in particular, which squares they can see clearly and which they can only vaguely see can define their safety and danger zones; consider the difference between a sniper and a pistolman. A sniper can clearly see a certain section of the battlefield, and dimly understands others, but is so blind to other squares that they can only be deployed to specially secured areas; a pistolman in contrast functions as effectively as a melee unit at short range.
This does not quite suffice to move directly into a discussion of magic in the stances system, but instead we will talk about magic-themed special abilities, which are not the same. Magic-themed special abilities and their interaction with stances are fairly simple to understand; they have an area that they effect and special consequences to the affected area, for example, you may leap four squares forward, doing damage and freezing enemies, and come out facing back towards your original square. This is a fun ability, but its tactical significance is mostly in the fact that it moves you, and forces you into a certain stance with a certain facing. You would not, for example, use this ability if it meant exposing your character's vulnerable squares to enemy melee units, or if it put you in range of other enemy abilities.
For the most part, these special abilities come with predefined constraints. They must be launched from a certain stance, put you in a certain stance, and can only operate at certain distances. What is more interesting is magic without predefined constraints, or alternately, with highly malleable restraints.
One of the fundamental conceits of magic in RPGs tends to be that magic works like a simple combat ability, but it either does more damage or less damage inherently, or has status effects like blind, paralyze, etc. Tactical RPGs expand this by allowing you to affect multiple squares within range, but I have never seen or heard of any interesting method for expanding or defining the range of a magician's abilities. It is here that the Stances system becomes most interesting.
In this conception of magic in a tactical RPG, I must admit that it is necessary to fall into a discussion of magic in fiction. A good starting place here would be to recommend to the reader the Recluce Series by L.E. Modesitt, Jr. Of particular interest here is the idea that a character can only affect what they sense, and their senses are naturally extended only under certain conditions, for example, a mage attuned to earth may be able to reach deep, a mage attuned to the sky may be able to call great storms, etc. Further, Modesitt's characters tend to spend a great deal of time setting up for major engagements; it is not merely a matter of working hard enough, it tends to be necessary to find the best possible location, and then making do with what you have.
Intuitively, the same mechanics are at play in my concept of magic. I stratify magic by element, and generally assign a school of magic per element, although that conversation is a bit advanced here. Rather, consider the following: a fire mage's range is dependent on his energy stores, and if he is depleted, he must recharge it by absorbing fire. For him, a battle in a forest is a wonderful thing; he may set a small fire, absorb it, set a much larger fire, absorb it, and then either set a fire that forces the enemy's position to change, or directly attacks them with fire.
By constraining the mage's abilities, their position and abilities in the ensuing combat are more interesting. Consider another mage, whose range extends freely through water hexes, or whose range otherwise is dependent on them being adjacent to water. This mage will choose an engagement near a lake or river if at all possible. Consider another, whose range extends through hexes of solid stone, or who must be up- or downwind of an enemy force. Consider another who is peerless in storms. Consider another who has greater range but lesser abilities. Consider a mage who can take their energy supply with them.
Then take all these individual mage types and allow one unit to use several different types of magic. Consider rules of mixed magical range where a wizard can reach you from well out of sight by combining wind and water and earth and fire magic. Fundamentally, the range of the unit is not difficult to compute, and so it is easily possible to have this unit in a tactical RPG. Fundamentally, one skilled mage facing an army on familiar ground can whittle them down piece by piece without ever presenting a target. This is an ability with malleable restraints; the unit is still only able to affect a small piece of the battlefield, but clever positioning lets them decide exactly which piece of the battlefield to affect, to their own best advantage.
How this fits into the idea of Atances is not entirely unfamiliar; the mage must make a choice ahead of the encounter to decide what tactics he is going to pursue. He must set up his own location to minimize his vulnerability and maximize his danger zones. He must be aware of the many ways in which the battlefield may turn against him based on enemy movements. But more than anything, when he makes a move, he lays his cards on the table. An enemy who knows they will fight a water wizard will not march up a river road; those facing an earth mage will fight on loose dirt when they can; those fighting a fire mage will take certain other precautions.
Stances battle engine is fundamentally tuned to at squad-sized engagements, but it can go down to individuals or up to armies. Magic in particular can, at the highest level, affect the world map, allowing the complete reversal of a enemy's advantage and potentially winning a war without a proper engagement. Most interestingly, Stances is a system based on preparation and establishing canon; that means that effects that may otherwise be relegated to narrative, such as magically scanning people as they pass through a gate, can be modeled with the engine instead, allowing things to be built and destroyed rather than hard-coded to the game map. Ultimately, I would love Stances to be part of a game with such a grand scope that you could simulate nations fighting in what amounts to a sandbox, while still fundamentally being built upon an RPG engine.
Alas, I never fail to set my expectations too high. It is only a concept, and something that large would be a monumental amount of work.