Dungeons Part Seven: Flow and Area Encounter Design

July 23, 2014

A much more indepth look at the ideas I spoke on in Running the Dungeon.


So remember when I said that dungeon design is like videogame level design IS game design?  In part one I linked a bunch of things from analyzing good dungeon set ups in a macro sense, or Metroidvania style game design.  Here we come full circle, except this time I’m looking at the flow of individual rooms, or areas, as they pertain to making a good combat encounter.  Here’s where mostly it comes down to the logic in stuff like MOBA or FPS game design.

Unlike those games, you’re not going to have playtesters run through it thousands of times, nor will you get to reuse most of your areas many times, which makes it a lot harder to come up with good, novel areas on a consistent basis.  Oh well! You don’t need perfection, you just need good fun.

Positioning as a choice

Boardgames and war games rely on positioning as a key point of strategy in play.  Tabletop roleplaying games… well, sometimes.  The problem is that many include maps or grids but ultimately don’t do a lot with positioning.   If the ideal form of play is to simply run up and hit each other without much movement during the actual battle, it’s functionally no different than a JRPG console game where the teams line up and hit each other until one side drops.

Tabletop games have attempted to add some of it with flanking rules or spell area effects, but it often only results in minor back and forth shifts during play and not a lot of heavy movement.  Because so few games give good support in the core mechanics, it falls upon the GM to set up areas where movement is encouraged.

The Crashpoint

What’s the goal in a combat game?  To beat the enemy.  So what’s the best place to be, tactically?  The place where you can dish out damage the quickest and receive the least damage.

In an empty room or area, with no features and no tactical factors whatsoever, the “best place to be” is the place where you can do damage to the enemy the quickest – which often ends up being the midway point where the two groups meet – the Crashpoint.

The Crashpoint is like the center square in Tic-Tac-Toe – it’s where all the interesting stuff of play happens, and, if you know where it is, you can build areas and rooms that are tactically interesting by putting stuff there to make it hard to get there, or to make other places more tactically valuable to be – and thereby, move or split the Crashpoint.

Remember, fun tactics sits in fun choices.  If there’s only one place that’s the tactically best place to be, what choice is there?  You want to make a few tactically interesting options, so players have to start thinking about what’s going to be best.

Movement Tax

So consider this – anytime characters or monsters might want to move somewhere else than a direct line to do damage, they’re paying a “movement tax”  (actually, it’s a time tax, as they’re losing actions doing something other than direct damage… but since we’re talking about moving a map and how to set up areas to affect this, we’ll stick with Movement Tax).

If you want folks to go somewhere other than the Crash Point, you have to either put enough hazards/dangers in the way, and/or advantageous positions in other locations to make the Movement Tax worth paying.   In other words, this sets up choice – players can try to go straight for it, or try to avoid danger, or try to get something that should be worth more than doing straight damage.

Area Encounter Design

Blockers and Funnels

Blockers and funnels are obstacles, walls, debris, rocks, trees, crates, whatever you want that basically stop movement and funnel movement into particular areas.  This changes where combats happen and change where the Crashpoints are.  These also tend to set up some fun tactics about ranged attacks and chokepoints for melee fighting.

It’s also worth noting what’s a Blocker or Funnel for one type of creature or character may not be for another – a half ruined wall might be a big obstacle for a normal character, but the giant spider just climbs up the side of it like nothing.    This is worth considering if you want to load the map to favor one group or another.

Also consider that blockers and funnels may be destructible depending on what’s going on.  A monster might knock them out of the way, an earthquake spell knocks down the trees, the bridges everyone is fighting on start collapsing.. etc.  When you put Blockers and Funnels in the way of the Crashpoints, you move where they go, just like how water flows around things.

Alternating Wide and Tight Areas

Generally, you’re not going to go wrong by playing around with alternating more open areas with only one or two hazards or obstacles, and tight areas.  Usually my rule of defining “wide” vs. “tight” is whether the party can all stand side-by-side and fight without having to do tricky maneuvering around each other (for grid based games, don’t forget this includes diagonal positioning).  Tight areas should often include alternate paths to flank each other.  You can take a big area and set up enough Blockers, Funnels, and Hazards until it’s effectively a tight area.

The big effect this has on play is a matter of things like how well groups can focus fire on a single target, how well players can quickly distribute healing, retreat at short notice, surround an enemy or use area effect spells.  It also strongly impacts the advantages of having good movement abilities, whether raw speed or skills to jump over small gaps, climb over things, fly, etc.


Hazards are a negative incentive- everyone generally avoids them, if they have a choice.  That said, putting Hazards next to Crashpoints or near places with advantages will cause players to have to start weighing their odds.  Be careful if the Hazard can move around, because it might completely change the set up of the encounter area.

Single Use Stunts / “Power ups”

Things that can be used to an advantage on a short term are positive incentives.  This could be a pile of logs to knock over on enemies below, a catapult to be shot, a lever to close a gate, etc.  If you can only reasonably use it once in a combat, it’s a “single use” advantage, and so, it needs to have a good amount of effect to be worth the Movement Tax.

Alternatively, you can make it a thing which can be reused at the cost of time – a catapult might be reloaded to be used again and again, but then the cost turns into time spent.  Again, consider what the benefit is compared to the amount of time the players could have been just using direct damage to the enemies.

Sandbox Encounter Considerations

Unlike an FPS map or a videogame encounter zone, combats can flow across areas, or move into spaces you didn’t expect.   This means your initial guesses on crashpoints or hazards might be completely off – the party isn’t going to care about the fiery pit in the middle of the room if they’re stuck in the hall when they’re getting attacked.   Again, this is the reason you don’t want a lot of empty rooms and halls without anything tactically interesting in them – the more the dungeon as a whole has interesting stuff in it, the more likely you are to have that intersect with your encounters for entertaining results.

Work vs. Payoff?

“Wow, this all seems like a LOT of work, to consider all these game design factors in making a dungeon!”

*Tired voice* “….yes, yes it is.”

I generally don’t do a lot of dungeon crawls for this reason – I set up short dungeons, maybe 12 areas/rooms or less, with more emphasis on Logistics and Tactical play, and not so much on Exploration.

Unless you plan on publishing a dungeon or having several groups run through it, you’re probably not going to get enough payoff for thinking this hard about it.  Of course, if you’re the kind of person who actually cares enough to want better dungeon design for your games, and you read this massive amount of words and thoughts I’ve thrown at it, you probably WILL use some of this to whatever level fits your needs.

If you find my blog entertaining and valuable, consider supporting me on Patreon.

%d bloggers like this: