Desktop Dungeons: Design Analysis (Part 2) 1

In Part 1 of Desktop Dungeons: Design Analysis, we looked at the core game play, interconnected choices, and tension of Desktop Dungeons. We saw that, despite its simplicity, this “ten-minute, one-screen” game offers the player endless variety through cleverly stacked core mechanics. In this part, we continue looking at what lies beneath this award-winning game; how choices are given meaning not only through their complexity, but also through the role they play in structuring an experience — an experience that progresses with the player, an experience that speaks a story.


Frank doesn’t care much for your attitude, chum. — From the description of Frank the zombie.

Conflict with monsters is the main action in the game. The ultimate goal is to defeat the dungeon boss, only possible by fighting past the lesser monsters. Monsters are killed to gain experience, to gain piety (from certain gods), and to open up unreachable parts of the dungeons. (Of course, the boss is killed to complete the dungeon.) Killing monsters drives the entire game. A monster can be harmed in several ways:

  • attacking it directly
  • casting a fire spell
  • turning the monster into stone (not possible for boss characters)
  • death gaze (gorgon only)
  • knockback damage (half-dragon only)
  • life steal (vampire only)

The first two are often used in combination – the spell is used to weaken the monsters, while the direct attack is used to finish them off. On the game’s Wiki, attack patterns are referred to ‘two-one’, ‘two-two’, and so on, to indicate the number of each type to use. For example, a ‘two-one attack’ means casting the Burndayraz spell twice, followed by one direct attack). The pattern to use depends, of course, on various factors – the most important being the player’s available resources.

The special classes offer interesting attack strategies for the player – many provides the player some of the benefits that some monsters already had, making them particularly gratifying. One of the more interesting abilities is the Martyr ability of the crusader: just before you are killed, you get a final blow in, dealing three times you normal amount of damage. Generally not useful – you still die – except that if you happen to kill the dungeon boss in that blow, you win the dungeon – even if you are dead!

Killing each monster is a little puzzle: can it be done? should it be done? how can it be done? And there are no easy answers for the player. The player has to consider the current situation (looking at his resources), the abilities of the monster (and they are just as varied as that of the player), and, most frustratingly, what his deity will make of your slaughter. Sometimes mercy must be shown (if just until the experience can be conveniently harvested during the boss fight)…and sometimes, the player is just really screwed. The point is: fighting is the most entertaining aspect of the game, because winning (and loosing) can happen in so many interesting ways.

Wall_libraryInformation and Progression

Woe unto the fool who does not heed the lessons provided here. — Desktop Dungeons tutorial.

Information is carefully managed in Desktop Dungeons, and ties in closely with progression.

In-Game Information

The key information of the dungeon – the monsters, items, shops, gold, and the gods – are all gradually revealed to the player, and the player has to adjust his playing style accordingly. The earlier the key elements are discovered, the better, and some characters get early access to certain information: humans can detect similarly levelled monsters; wizards can detect spells. For the most part, the information becomes available as the player explores the dungeon. But exploring a dungeon has two more effects:

  • It rewards the player with health and mana. When the player is maxed-out, this reward becomes a punishment of waste.
  • It lets (non-poisoned) monsters regenerate.

The player controls dungeon exploration (at least partially), but because of the consequences of revealing information, it has to be done carefully to maximise the benefits. And it is integral to the strategy of playing the game.

Game Content

New game content is also revealed gradually as the players unlock them by defeating the boss characters in the dungeons.

The unlocking feature serves several purposes. Only the first is directly related to information; the other relates to other aspects of the game.

First, it makes the amount of information the player is asked to deal with more manageable at the beginning of the game. The player has a chance to get familiar with a subset of the content and to develop some basic strategies, without suffering from acute information overload.

Second, unlocking items also marks progression. This locates the player in “skills space”, and gives the player a sequence of games that, on average, increase in difficulty.

Third, new content is the reward for playing well. The reward is so compelling because of the unique properties of each game element. New items are not merely fads, but they open up more gameplay possibilities: new strategies, new ways to combine the game’s elements.

Fourth, the nonlinear way in which content is unlocked makes different orders possible, leading to different paths from no-unlock to full-unlock. This increases variety and replayability, and gives the illusion of an overarching story where there is none (see the ‘Narrative’ section below for more observations about this).

Character for the Gods

Gods have complicated behaviour that gets revealed to the player very carefully (assuming the player does not consult external sources).

The player must encounter a god to know that it exists. Even when an altar is visible, the player must first pray at the altar to see to which god it belongs. And while players can see the blessings and conversion possibilities of the gods, they cannot see the prices (in piety) they will have to pay for these before they start to worship the god. Although the description of the god might give some hints about the god’s particular reward and punishment system, it is fully revealed only when it is triggered. It is possible to worship a god across many games, before finding out that the god frowns upon converting your glyphs and punishes it severely.

Discovering the gods’ characters is an implicit campaign: it helps to tie together a series of otherwise unrelated games.

Tactics and Strategy

The most important information that players can gain is about tactics and strategy. But they can get this information only gradually.

They have to wait for appropriate situations to arise during which they can test out strategies and tactics, making it much harder to discover and master them.

For example, the Wanafyt spell, which summons a monster of the same level as the player to an adjacent square, is very effective very early on in the game when playing as the (Human) Rogue, especially when already worshipping an appropriate god. To learn the value of stringing Wanafyt spells together can take many games, because the ideal circumstances for using this tactic arises very rarely.

Class Progression

classes Although there is a thematic link between a vertical of classes, and each class is more powerful than the one on the previous tier, there is no linear progression from one class to the one in the next tier (in the same vertical). A beserker is not simply more powerful than a fighter; it is a different character. This makes playing different classes more interesting, and makes it much more fun to play the special challenges with all classes. What is true, though, is that a higher tier class is more complex than a lower tier class – it is harder to work out the best strategy.

Monster Progression

The difficulty of a specific type of monster depends on its level. Players choose which monsters to fight, and generally will fight more difficult monsters when levelling up. But new monster types are unlocked when finishing dungeons, and although these monsters are more difficult, the difficulty comes not from increased statistics, but from their special abilities: poison, mana burn, damage resist, etc. Players have to learn the best way to deal with each monster, and cannot simply draw conclusions from what they already know.


As mentioned already, the game becomes more difficult as the player progresses and unlocks more content. It is worth looking at the core skills that the player is developing:

  • Numeracy: This involves looking at the numbers, and understanding their implications. The interface helps with this in short term goals, but for longer term goals, the player still has to do a lot of calculations.
  • Memory: Although a lot of information is presented to the player on demand, the player needs to remember the particulars of the gods. The most common reason for dying is forgetting some fact about a monster (despite the information being easy to get through the interface).
  • Analysis: As the complexity of the game increases, the consequences of the relationships between the game elements become harder to analyse, and strategies become harder to develop. This skill is the key to winning later levels.

HealthPotionResource Management

Currently out of supply. — Part of Desktop Dungeons tooltip for empty potions.

The player has to manage several resources effectively to solve a dungeon: health, mana, gold, piety, unexplored dungeon, glyphs, items, experience and other character abilities (although not strictly resources, they still enter into the equations in a similar way), and even living monsters.

Every action in the game has an associated cost. What is clever about the game (and aids greatly to its replayability) is that these costs typically depend on so many things: the action itself, on the character the player is playing, the god he is worshipping, and other previous choices.

We already mentioned how information is revealed as players explore the dungeons. But exploring also gives the player resources, either as reward, or in the form of power-ups, pickups, potions, etc. As if this does not complicate the balancing act enough for the player, different strategies are required when mana burn or poison is involved. The Lemesi spell that reveals hidden dungeon blocks poses as an information tool; however, its most important use is as a resource management tool.

To see the intricacies in resource management, consider the following strategy for dealing with the dungeon boss described in the game’s Wiki:

The ideal position to be in when starting a boss battle is at level 7 with 34 / 35 experience, with proper runes [i.e. glyphs], all health and mana potions remaining, and an experience farm containing 96 experience worth of monsters (enough for level 10). The first phase of the boss battle involves the player, using all available health and mana points (but not potions) against the boss before killing weakened monsters to gain a level. This phase ends when the player no longer has enough weakened non-boss creatures to gain another level, or has reached level 10. Phase two of the boss battle involves the player using other available resources (health points and mana potions, etc.) to defeat the boss.

Levelling up – used strategically – gives the player more health and mana than is available from potions and unexplored squares alone. To enable levelling up at will, the player needs to build an experience farm: a collection of monsters that have been weakened so that they can be killed with a single blow, so that the player can level up quickly when running low on health or mana. Building a large enough experience farm is a challenge in its own right.

Balance and Difficulty

Steady worship is rewarded. — The Pactmaker.

‘Balance’ is often used as a catch-all concept. Generally, it means all roads to success are equally difficult or equally likely to succeed.

Appropriately assessing the degree to which the game is balanced cannot be done without analysing a lot of data. However, there are some clues:

  • The game is fun (or, to keep the lawyers quiet: many players enjoy playing Desktop Dungeons).
  • The way an intermediate player can almost win most games suggests that just a few better choices in most games would make victory possible.
  • The player is always close to empty on resources, at least in the end-game (for most characters).


Dungeon Balance

I’ve mentioned that not all dungeons are created equal. Indeed, occasionally players may encounter dungeons that are impossible to solve. The tutorial warns us not be afraid to fail, suggesting that unsolvable dungeons arise by design. The occasional unsolvable dungeon increases the uncertainty – and consequently the tension as well. It also motivates the player to look harder for lateral solutions, such as using spells in clever ways to escape impossible situations.

Character Balance

Except for the three special races that are both races and classes, all characters are compiled from a race and a class on one of three layers. Only the first layer is unlocked initially. The properties of the races and classes cause certain combinations to fare better than others. For example, Elves are good for wizards because wizards use a lot of mana, and Elves have more max mana than other races. So clearly, ignoring other aspects of the game, not all characters are equally likely to succeed, and, in this strict sense, are unbalanced. Of course, this is still useful: “Poor” combinations are handicaps that lead to more challenge for veterans, and certain combinations have been identified as good for secondary gaols such as gold farming.

Unlock Branches

Items, characters, spells, and so on, can be unlocked in different sequences, depending on the order in which the player wins dungeons with specific characters. Certain sequences might make other aspects of the game more difficult than other other sequences. For example, by winning the first dungeon with a fighter, the wraiths are unlocked, which may make it more difficult to play the wizard class. This is a balance problem, with almost permanent consequences. Although the player essentially chooses which branch to take, a new player will not have enough information to make these decisions sensibly.

Bad Combinations

One strategy while playing the game is to keep trying one’s luck as opposed to improving one’s skill. The general idea is to explore the dungeon quickly for specific glyphs or gods; if they’re not present or they’re hard to find the first time round, the player restarts the level, much like the card games Patience and Solitaire. This is especially common for the special challenges, where the bosses are known in advance. The short game sessions encourage this strategy, and (judging from advice on various forums) it seems to be unavoidable. What is bad about this is that the player is not really playing while restarting for better conditions. Players’ willingness to invest time in this type of strategy shows just how rewarding the gameplay is when the combinations are not unfair.

Emergence and Complexity

You just woke up the wrong dog. — From the description of Nine Toes.

The game elements are quite simple, yet they combine in complicated and unexpected ways to strengthen or weaken a player’s position, and to allow the player to adapt using a wide variety of strategies.

It is important to note that randomness alone does not create this variety and interest: the special properties of game objects also play a part. Monsters are not merely numerical variations on a base template. There are monsters that take away mana, prevent health regeneration, or always strike first. When combined with spells, items, and god boons, the player can’t rely on the same fighting tactics every time.

Similarly, gods have properties way beyond mere statistics: they have different requirements for worship (money, health), different requirements for dealing out piety (using spells, killing monsters, or continued loyalty), different ways of getting offended (using magic, converting glyphs), and different punishments (reduced damage protection, blocked health regeneration). In the right circumstances, items can have unexpected uses. Every player has a few such stories. Here is one of mine:

I was at a reasonable level, and had just uncovered a large part of the dungeon. Low on health, I needed replenishment before launching my final attack. I was out of potions, and a creature blocked the unexplored part of the dungeon. Although he was at the same level as me, I was just a bit too weak to take him on. I was about to give up when I noticed I hadn’t converted the Wanafyt glyph (which summons a monster – if it exists – of the same level as you to a neighbouring space). And just as I was about to drag the spell over, I realised that, with some luck, I could use it to move a monster out of the way. I used it, and it worked. It saved my game.

An important aspect that emerges from the game elements and structure is the battle tales of single games, and the overarching war story of game sequences. We’ll look at this under ‘Narrative’.


Strike a dramatic pose, my love. I absolutely ADORE the grandeur of it all. — Medusa.

Desktop Dungeons does not emphasise story. The plot emerges spontaneously between entering a dungeon and dying or retiring. These stories are all unique, because of

  • a player’s choices and actions,
  • random dungeon elements,
  • the arbitrary unlock sequences.

The most unique (if not most epic) sequence of events occurs during the early parts of the game, when elements are still being unlocked.

Of course, Desktop Dungeons shares this with all role-playing games: players make the story. What is special about Desktop Dungeons, is that so much is done with so little. There is no explicit link between one play session and the next, except for the (limited) stash of gold. However, the unlocking scheme adds a clear sense of progression, and together with information that gets revealed gradually, gives the player the feeling of a journey rather than just quick session without any context (as is the case with Minesweeper, for example).

These emerging stories and journey would not be possible without the explicit narrative context that the game provides: the roles that the player takes on, the characters of the enemies and gods, the game world (the dungeon, with its mysterious shops and other items scattered about), and the achievements summarised at the end of the game. In addition to giving meaning to player actions, the narrative elements also allows a lot of humour to be sprinkled throughout the game, not only in the witty descriptions and dialog, but also in the curious mix of characters and items. Without the narrative elements, the game would be a dry exercise in mental arithmetic.

The mythical world of Desktop Dungeons is eclectic. As in other role-playing games, many elements are taken from prehistoric, Norse, Greek, and Eastern European mythology (Earth Mother, Jörmungandr, Medusa, Vampire). But there are also references to other games as kudos: from the genre starter Rogue, there is the Amulet of Yendor; from Dungeon Crawl, The Orb of Zot; from Diablo there is Gharbat; from contemporary indie games we get Meat Man and Goo Glob. There also seems to be a healthy amount of invented mythology (Wicked guitar – or is this a reference to Guitar Hero…or something I just don’t know about?). The spell names, at first blush, look like they come from some archaic language; but they are really quirky names in modern lolspeak (Weytwat, Burndayraz).

It’s also interesting to note how the gameplay offered by the religious system supports the religious aspect of the narrative. Like the gods of real-world religions, the gods in Desktop Dungeons are not easily known, as mentioned before. Contrast this with the earthly (if undead) monsters, which are fully known when a player encounters them.

Also, there is a sense of arbitrariness to the gods’ behaviour: they all operate very differently in terms of what they require as offerings and how they reward and punish their subjects. There is greed and insanity. Most gods are finicky, and unforgiving. These characteristics emphasises their divinity: each god makes its own rules.


  • Short, easy-to-get-into games does not equate to simplistic gameplay and limited choice.
  • Randomness and content that combine in intricate ways can help to create emergent play.
  • One can create tension by combining aspects to create unique situations for players to overcome, and by structuring the game in a specific way.
  • Non-linear unlocking sequences make it harder to ensure that all campaigns are equal.
  • Revealing complex information gradually and unlocking content cleverly can help bind otherwise unrelated mini games into campaigns — making experiences that tell stories.
  • One can increase difficulty by adding complexity to the gameplay, rather than just fiddling with hit points, mana, and other numbers.

Thanks again: for Diorgo Jonkers and Jacques Jurgens for valuable feedback. 🙂

More on the Design of Desktop Dungeons

Desktop Dungeons Wiki Detailed info on game elements and strategy.

Tropes in Desktop Dungeons

Picking the lock-box A discussion of the unlocking system of Desktop Dungeons. But see what Rodain has to say in Meaningful Unlocks:

Philosophy and Inspiration

What I Learned from Dungeon Crawl

Podcast #15: Rodain Joubert and Desktop Dungeons

@Play: Interview – Enjoy A Coffee Break Of Victory With Desktop Dungeons

Full-version work-in-progress design discussions

About Herman Tulleken

Herman Tulleken is a game developer co-founder of Plinq. He thinks computers are a necessary evil to make games, and sees frying a CPU or two in the process as a worthy sacrifice.

One thought on “Desktop Dungeons: Design Analysis (Part 2)

Comments are closed.