At the same time, the other resource in the game is the unit's skulls. As you kill your enemies, you can eat their skulls to gain more health. After you eat three skulls, your unit turns into a demon, which means that unit can attack twice per turn if you assign it one of your five orders, as well as having twice the hit points of a standard unit.
If a unit dies, any skulls they've eaten drop to the ground -- free for anyone to eat. As units die during a game, that naturally means there are more powerups. So the last fighting units will typically be heavily powered up, given a sense of increased stakes for the player.
A cavalry unit that's been powered up to a demon by eating three skulls.
Typically units take two to three attacks to kill, depending on the unit and the attacker. As demons have twice the health of a normal unit, it can take as little as three or four attacks to kill a demon (again, based on the units involved). So while a demon is a huge advantage because of its extra action and the value placed on the limited actions per turn, another player can still kill a demon in one turn, probably with an action or two left to eat some of the dropped skulls.
On top of this, melee units can knock back their targets while attacking. Getting knocked back over a ledge of any sort (cliffs or water) will cause the unit to die immediately (and not counterattack).
You can protect your units from most knockback by grouping them together to form spirit walls. These also act as collision barriers to block enemies and defend your own. One mistake in positioning and an observant opponent means any unit is always one short step from possible death.
Also, given that your general's life is the only thing keeping you from the losing the game, one small mistake in using him can cost you a match that you were otherwise winning.
An infantry unit knocking an archer off a ledge.
In this way, we provide mechanics and advantages that are high reward, but high risk (or which can easily lead to a loss). That can give players a sense of power and accomplishment during the match, but they must stay on their toes in order to maintain any advantage. So the player that is ahead in the match can fluctuate wildly, but players feel like that change is due their own skill or mistakes.
Most turn-based and real-time strategy games boil down to some sort of rock-scissors-paper mechanic. We thought it would be interesting to purposely step away from this model and try something different. The high concept of the main three battle units was to give each one distinct strengths and weaknesses, where the strategy was to make the most of their given strengths while avoiding allowing the enemy to exploit their weaknesses. The real strengths are allowed to shine when intelligently used together, as well as minimizing their weaker attributes.
For example, the infantry units have great defense, and can take a lot of hits. They are the equivalent of a linebacker, holding your defensive line. They don't move far per turn or have particularly great attacks, but they are great at advancing a strong line of defense, and they are the cheapest unit to buy. They're usually too slow to go on any major offensive push but an excellent support and defensive unit.
The archer, on the other hand, has very weak defense but an amazingly powerful ranged attack. By keeping him protected by infantry, you can soften up enemies safely from afar, while having an infantry bodyguard rush in to finish off the damaged enemy before he can counterattack. One isn't particularly better than the other, but used in conjunction well, they become much stronger.
Most games in this genre can boil down to puzzle games, where a certain unit is always the right antidote for a certain enemy. Anti-air vs. aircraft, for example, is not a lot of fun when you have a great fleet of aircraft and the enemy just quickly builds up an anti-aircraft front line. This gives players a range of expression for the strategies they want to follow. Even more importantly, though, it helps prevent multiplayer matches from turning into slugfests where one player has rock and that player is just beating the player with scissors for over 10 minutes.
Forcing ourselves to convey all needed information to players as quickly as possible forced us to streamline our mechanics and only keep what we could explain. Making sure the mechanics and the resulting dynamics stayed open and didn't force players into one best approach -- with opportunities for dramatic power shifts -- gave us a whole range of strategy so we didn't water down anything while we were trying to present that information as clearly as possible. Having a clear mandate to apply to any new possible mechanics that came up let us stay true to the spirit of the genre we loved, while giving it a needed kick in the rear armor.