High Granularity Vs Low Granularity Space


Preface (3/23/2015): I originally wrote this article three years ago, when I was just beginning my design career. It was half proposal, half exploration of design concept. I’ve cut a few parts out that aren’t really relevant to my readers; if it seems to start and end abruptly, that’s why. I had not yet worked on a MOBA, where I learned quite a bit about tanking in PVP, and I overlooked many examples of high-granularity 3D games that had readable positional and facing gameplay. Nevertheless, I think the basic gist of this article is still quite accurate: there is an inverse relationship between spatial granularity and tactical readability.

I’ve been doing a lot of thinking about the game industry’s embrace of fully 3D spaces, huge numbers of possible viewing angles, and the incredible amount of fine granularity introduced when you can move or look anywhere. I’ve come to some interesting conclusions about the effect this all has on gameplay. I want to throw a wrench in the tanking paradigm; bring a focus onto positioning, level design, and the environment; and improve the tactical readability of combat.

Top-down Grids vs. Fully 3D Spaces

Readability and teaching: Tile-based environments have enormous advantages when it comes to judging distances and range, clearly defining the boundaries of different terrain or effects, and teaching the player consistent properties of different tiles. Basically: it’s easier to learn, and it’s easier to read.

The Legend of Zelda: Link's Awakening DX

Players know that their sword has a reach of one square, they can jump two squares if they’re good, grass can be walked on, bushes cannot be walked though, pots can be lifted, and that green squares turn yellow and then red then into holes. Armed with all this knowledge, they can take in an entire gameplay situation, right down to precise spatial relationships, at a quick glance. Even a space they have never been to before. Discovering these properties came from experimentation, but it was experimentation with clear and consistent results that can be applied to similar tiles in the future.

Aiming and targeting: Aiming abilities or projectiles in a 3D space can be a hell of a lot harder and less accessible than aiming in a top-down grid-based 2D space, where projectiles or abilities usually fire in one of four clearly defined directions, relative to the avatar. It is also a lot slower. By contrast, in most top-down 2D games you press a single button and the projectile or ability is immediately fired in the direction you are facing. (And, because aiming of actions is strongly relative to facing and position, they become extremely important!). This is also true of targeting non-projectile abilities: when AOE attacks are defined in relation to a spot on the grid rather than a spot in finely granular space, you can move away from things like targeting lock-ons and trying to click on a precise spot, because that sort of control and feedback is built into your space.

Megaman Battle Network

Aiming a simple projectile is quick and easy: it will fire in the direction you’re facing, and it will move straight through the tiles on the grid, hitting things in the same row as you. More complex targeting, like cones and other shapes, also becomes easy to judge and give strong feedback on. This goes for enemy attacks, too – when the AI takes actions, the player can learn patterns very easily.

Positioning and Facing: It’s a lot easier to design with positioning and facing as major factors in gameplay when things are positioned on a tile (a manageable, defined unit in space) and facing in one of 4 directions. Largely because it is a lot easier for a player to read the situation and understand those factors. Compare to a 3D game, where players could be facing in 129600+ different directions, there is limitless granularity when it comes to positioning, and collision boxes overlap and collide in thousands of different ways.


#1 is tile-based movement and 4 facing directions. #2 is more granular movement, and 4 facing directions. #3 is full 3D, and offers huge facing and movement granularity. In Pokemon, stepping into the line of sight of an enemy triggers a battle; note how that gets progressively more difficult to judge as the granularity increases.

The Environment

By breaking the environment up into a grid, you can treat each square of the grid as an object of its own, with its own properties and status. Properties like ethereality to the player or projectiles, difficult terrain, height, elemental amplification, cover. These properties can be clearly shown and defined to the player, as demonstrated above, and they can be used to drive gameplay directly: actions can be performed on the tiles, the properties of the tiles can modify the events that take place on top of or around them, and gameplay events can change these properties.


Properties of a tile can be explicitly defined and have a big impact on gameplay. The top screens show mountainous tiles selected, and units who stand on them gain bonuses to defense or avoidance. Units who stand on the selected tile in the bottom gain penalties to electric resistance, bonus to acid resistance, +25 to attack, and +40 to defense.


You can allow the player to take actions that target tiles on the environment, such as placing traps, and expect the results to be predictable enough for solid tactics to emerge.

The ultimate extreme of this concept is real-time terrain editing – or destruction. If the tiles of the environment are treated as individual objects, then they can be destroyed, created, raised, lowered, or moved with impunity, while retaining the gameplay strengths of tile-based environments – especially if the movement on all three axes is kept to a rough granularity.


Kodu Game Lab, a game editor/toy/game. The terrain is comprised of blocks that you move in real-time. This is taken to the extreme, and there’s a lot of fine granularity in the different brush sizes and the scale of the tiles, so it doesn’t quite capture the other concepts I’m talking about. But it illustrates this one well!

The Tanking Paradigm: Let’s break it and leave it in the dust

The ultimate goal of “tanking” is to protect weaker units who cannot take being the target, and to concentrate limited healing resources onto fewer targets. Tanking in MMOs typically revolves around two concepts: making a character that can hold the attention of the enemy, and surviving that attention. If you’ve got a mage in your party, it is the knight’s job to ensure that he, and not the mage, has the “agro”. He does this through “taunting” or “intimidating”, which arbitrarily attracts the enemy’s attention, and by doing damage. A key element is dumb enemies: these agro systems require enemies who are too stupid to act in their own self-interest, and instead have simplistic behavior that is controlled by the actions of the players. If the enemy was smart, for instance, he’d say “screw you knight, I’m going to kill that mage!”… and then walk over to the mage and kill’em dead. There’s a reason that agro-based tanking mechanics don’t survive a transition into PVP.

A more interesting approach would be one where “tanking” revolves around physically blocking access to the weaker units, or by putting an enemy into a position where switching their attention to a different target is a poor tactical decision (dropping your guard, turning your back, splitting your attention). Both of these concepts work really well in a grid-based based environment where facing matters.


Blocking Enemy Movement

In a fully 3D game with fine granularity for facing and position, you run into all sorts of readability problems:

  • it’s difficult to tell whether a gap is big enough for an enemy to move through.
  • You can’t judge that sort of thing for an area far enough in advance to make tactical plans.
  • It can be hard to tell whether you’re directly facing an enemy or not.
  • It’s really hard for multiple people to line up into a solid wall.
  • It’s hard to tell whether an enemy got through because it’s a special sneaky enemy, or because you’ve got a hole in your line.
  • Environments need to be narrower, but players feel cramped easily in 3rd person 3D space.

It is also hard on the development team: when designing a chokepoint, you need to worry about how wide the chokepoint can get before it’s too big for players to block it, you’ve got physics and nooks and crannies and intersecting collision boxes that create weird unintended routes for monsters to take and spots for them to get stuck in, or backed into corners, or perched, etc.

In a top-down grid-based environment, all of these problems go away. You can easily read a large area at a glance and make tactical decisions. When a tank takes up a defensive position, they block their clearly defined square, no more, no less. You can clearly see the difference between an enemy that’s flanking you and one that’s coming head-on. Realistic scale of environments and objects is less important, so it’s easier to make a space that doesn’t feel claustrophobic.

One of the most interesting things that can arise here is the need for multiple tanks, or for non-tank characters to take up defensive postures to help hold a line. Maybe you have a compelling need to block a corridor that is three squares wide – the solution might be to get two tanks and have your ranger plug the remaining hole, or maybe your geomancer can pull a temporary pillar of earth out of your terrain to block that spot, or your two tanks can try to be quick on their feet and block all three squares with a combination of shielding and knock-back attacks.


Four Swords adventures allowed you to lock four Links together into a solid line. With shields up, you have a solid shield wall, exactly four tiles wide, with no chinks.

Turning your back on your enemy: A bad idea

When facing is readable and important to the gameplay, you can create mechanics where turning your back to an enemy you’ve already engaged is a bad idea. Thus, “tanking” can mean something as simple as “being engaged face-to-face with the enemy, and if I turn my back he’ll cut me down – and vice versa.” Maybe you’ve got a big kite shield, but you can really only face it in one direction. If you’re engaged in melee combat, turning to face a different direction is asking for some attacks of opportunity, and is thus an action that has an inherent cost.

Note that this also means that flanking and positioning becomes very important.

Misc. Sundry

There’s a lot of room to play with mechanics like these, as far as progression, enemy variation, and player actions go. Maybe you have a “stand your ground” stat, and the higher it is the longer you can keep your guard up without having it battered down. Maybe some enemies (and players) have a tumble ability that lets them bypass opponents on guard, and other enemies are physically large and can trample right over the defenders. If physically blocking enemies is important, then there are all sorts of ways to play around with a spellcaster who specializes in modifying the environment, or a rogue who actually creates traps, or an engineer who makes weapon emplacements straight out of a tower defense game.

Level Design

All of this drastically increases the importance of level design. In combat, level design in most 3D MMOs and action games boils down to high-level navigational stuff, or supporting scripted sequences and events. When it comes time to actually engage in combat, it typically happens in clear open spaces. Level design’s job is to get out of the way and let the combat mechanics take over. And when it doesn’t it’s often annoying or cramped, or you can’t see what you’re doing, or you get stuck on the collision of some object, or you lost line of sight so your heal spells fizzle out, etc.

With a top-down camera, a grid, facing as an important mechanic, clear feedback about gameplay impacting properties on tiles, abilities that can target the environment,  and “tanking” as described above, level design reclaims its rightful place as a tactically significant factor in gameplay.


With a space like this there is a lot of room for players to devise different tactics that take the space into account. Blocking different chokepoints while ranged support flanks enemies, knocking enemies off the edge, avoiding them, jumping over holes to flank – and the level design plays an integral role in defining how that all plays out.

Applying these concepts

Nearly any genre or franchise could benefit from being a top-down grid, as long as you’re very aware of what you’re gaining and why you need it in your game. One with player-generated content? Well, the granularity and clarity of working with tiles would be extremely important. An MMO with a view and a grid like this would be very different from most of its competitors. You want an interesting fusion of gameplay styles and real-time modifiable terrain? We’ve already been over that one!

You could be extremely literal and make it tightly bound to a grid and even turn-based, or you could loosen things up and find middle-ground where you gain some of the benefits without sacrificing fluid movement and real-time gameplay. You could set it on a grid but leave it 3rd person. You could have a top-down view but throw out the grid.