What I Learned from Dungeon Crawl 12

There are a few ways to learn about good game development. One is to frequently read up on game development articles and opinions. Another method involves just practicing and refining your work through, well, actual game development. Both of these concepts, however, are enhanced gloriously when married with the concept of simply playing other people’s games. When you come across something that you love, taking the time to study it from a game development perspective can yield untold rewards and glory.

It’s for this reason that I really, really like Crawl. Not only is it quite possibly my favourite Roguelike in existence, but it also includes an in-game “design philosophy” text outlining a few of the game’s key development goals, offering justification for each. This is a staggeringly nice touch for people interested in development themselves, and makes it quite easy to appreciate the game on a deeper, more fancypants level.

No, not THOSE fancy pants.

If you know about Crawl but haven’t yet looked at its design philosophy, read on for a study of its goals and methods. For those who haven’t yet played Crawl at all, do yourself a favour and download it now so that you can while away the rest of your existence within the confines of its fiendish and treacherous dungeon walls. Then come back here and figure out how the devs got you so hooked.

Lesson the first: Tricky and sometimes unfair

“Haha, you’re f**ked!” – Crawl guide to dealing with Sigmund.

The most delicate aspect to sort out in any game is its difficulty. Challenging games tend to be very rewarding for dedicated and hardcore players, while easy games promote accessibility and tend to be met with a broader audience. Crawl sits firmly and unapologetically in the former category: like all Roguelikes, it requires dedication, patience and an army’s worth of player deaths to finally get right, and even then it continues to humble even the most veteran gamers.

There’s nothing wrong with making a very difficult game as long as you accept two things:

  1. your audience may well be narrowed down to the truly dedicated and
  2. you’re going to be challenged on your decision.

A common error for new and inexperienced devs is to make their games way too freaking difficult, and sometimes casual observers will look down on a tricky game for just this reason. From the start, you need to get the message across that your game is challenging, that the difficulty is intentional and that you’re tackling it responsibly.

The Crawl design justifies the difficulty decision wonderfully. In fact, the team brazenly admits that there are times when your character is quite frankly screwed. To lift a quote from the docs of Stone Soup (a visual, tile-based version of Crawl):

The possibility of unavoidable deaths is a larger topic in computer games: Ideally, a game like this would be really challenging and have both random layout and random course of action; yet still be winnable with perfect play. This goal seems out of reach. Thus, computer games can be soft in the sense that optimal play ensures a win. Apart from puzzles, though, this means that the game is solved from the outset … alternatively, they can be hard in the sense that unavoidable deaths can occur. We feel that the latter choice provides much more fun in the long run.

A related and seemingly cruel decision is Crawl’s inclusion of out-of-depth (OOD) situations: for example, placing a high-level monster on an early level that would almost certainly destroy any regular player. But lo, it also makes sense in its own weird way:

[OOD events] serve as additional motivation: in many situations, the OOD monster can be survived somehow, and the mental bond with the character will then surely grow. OOD monsters also help to keep players on their toes by making shallow, or cleared, levels still not trivial.

The text also mentions a dungeon location known as the Abyss: a nightmarish realm filled with overwhelming hordes of demons and shifting portals, where life is at best temporary and exit isn’t guaranteed. Having personally screamed “SHIT!” on several occasions when a low-level character of mine has been banished to the Abyss (those bloody elven demonologists being the usual offenders), I can attest to the intense fear that the very idea of this realm evokes, and the staggering relief and sense of victory that comes from successfully escaping it.

Now that I think about it, this series kinda kicked ass too.

It’s quite likely that this idea of difficulty will always be heavily contested: Jeff Vogel from Spiderweb Software recently wrote about how it can be a pretty bad idea to bushwhack your players and raises a few pretty good points, though the comments offered below his post would suggest that some players do indeed crave the higher stakes.

But Crawl seems to do things well enough, at least for its intended audience. Combined with the inability to save and load games in a conventional way (once a character dies, that’s pretty much it), it manages to create emotional experiences in the course of play that one would find hard to encounter anywhere outside the realm of the hardcore roguelike. Difficulty hasn’t simply upped the challenge of this game: it’s been used as a tool to create feelings and inspire players because they’re genuinely operating in an unsafe environment. Which is great, even if at times the dungeon master seems to be acting like a dick.

Lesson the second: stay away from no-brainers

“Liberal use of resources WILL kill you.” – common Crawl wisdom

Crawl is full of meaningful decisions. And by that, I mean decisions which are tricky. Decisions which require short-term tactics and long-term strategy to be taken into account. Decisions which always require the sacrifice of one resource in exchange for another. Decisions which, in proud Roguelike tradition, will usually make the difference between claiming the Orb of Zot and dying a horrible electric death because you didn’t keep a resistance potion handy for that horde of blue dragons.

I actually referred to this design trait in another article about Crawl not too long ago, and it’s probably the game’s strongest trait. In the same way that OOD monsters can keep even experienced players on their toes, the scarcity of resources and the ever-present dangers of the world mean that there’s no clear-cut cases: sure, we’d all like to stockpile ridiculous quantities of health potions and food so that we’lldefinitely be able to tackle the next level of the dungeon, but what about our character’s weight concerns? And will those health potions necessarily save you in all instances?

Where’s the fun in “definite”, anyway? A lot of games rely on rhythm and rehearsal (and are rewarding in their own right), but Crawl demonstrates how a gaming experience can actually retain its depth and uncertainty even after players have gone through it hundreds of times. It’s a gaming paradigm which many other titles strive towards and fail to replicate: true replayability. It’s the kind of stuff that forces veterans to keep taking the game seriously. Heck, it’s the sort of thing that allows a level 15 character to starve to death because he swigged a levitation potion and couldn’t reach some food that was lying on the floor (hey, true story).

Avoid no-brainers!

I’m tempted to compare my love of this design to my absolute and unnecessary hatred for World of Warcraft: Crawl’s philosophy being about variety and adaptability, while the latter places more stress on repetition, reinforcement and refinement (and don’t you dare try telling me otherwise after I’ve spent an hour hunting down and killing a bunch of freaking gnolls for their freaking stupid teeth). If I already know how to do something in advance, and the only real challenge facing me is a time investment, then is the challenge actually worth it? It depends, of course, on what you’re actually playing games for: if you subscribe to the psychological hooks of progress and reward, or the raw joy of unlocking things (outlined most clearly by the excellent and startlingly more-ish Upgrade Complete), then this won’t really apply. And of course, sometimes we’re willing to grind now for awesome things later (20-man raids ftw!). But gamers looking for a different paradigm should find this advice welcome:

Wherever there’s a no-brainer, that means the development team put a lot of effort into providing a “choice” that’s really not an interesting choice at all,” says the Crawl dev doc. “And that’s a horrible lost opportunity for fun.

Hear hear!

Lesson the third: keep the fun, cut the rest

“Yet my observation of gamers who boast about “beating the game” is that they often appear not to have enjoyed the journey — that is, even for them, sometimes the game is more like work than fun.” – Lew Pulsipher


The above quote comes from a rather fascinating article on Gamasutra about game challenges and the idea that games can accidentally slip into the realm of “work” if developers aren’t careful.

It doesn’t strictly deal with what’s being discussed here (quite the contrary, actually: it talks about removing the focus on game challenges completely — that, and the author seems to like WoW a lot more than I do). However, the core danger of games turning into work is a very real and present threat whenever a developer sets out to make … well, anything. Often enough, games aren’t 100% fun 100% of the time: there’s the tedium of clicking on menu buttons, or sitting through loading screens, or running halfway across an map in Borderlands because your ride got blown up and it’s taking you years to move to any meaningful locations. We have a lot of names for gaming moments like these: “the grind”, “in-betweeners”, “pure unadulterated tedium” and “unskippable cutscenes that we’ve seen over and bloody well over again”.

The problem is that there are a lot of very necessary, yet very lightweight chores that accompany a game like Crawl: marking unidentified scrolls, map backtracking, item management and killing off low-level monsters when they show up. Fortunately, Crawl knows about this — in fact, the project philosophy specifically mentions grinding, defining it as “activities that have low risk, take a lot of time, and bring some reward.” Such design flaws “encourage players to bore themselves.”

When the optimal route to gaming success spawns directly from tedious and repetitive activity, the game arguably suffers as a result. Funnily enough, a lot of gamers are quite used to putting up with tedium and we’ll all accept a certain amount of this if it furthers the overall glorious agenda. However, a game that wants to be “truly great” instead of “pretty neat” should probably avoid this sort of thing as much as possible.

Shops, for example, are a pretty common flaw in games, even if people don’t quite realise it: they often encourage “item hoovering” and constant back-and-forth behaviour for buying and selling. When you’re ploughing through a desert tomb in Diablo 2 and find out that your inventory is full, is your first reaction one of delight? Are you genuinely eager to take a break from the action, summon a town portal and sell off all of your crap to some gormless merchant before continuing the slaughter? If your answer is “yes”, you’re a damn liar. As wonderful as the game is, the only time I’m happy to go back to town is when I find a particularly tasty drop and want to get my stuff identified straight away.

Crawl sidesteps this elegantly enough: shops in the dungeon don’t buy items, so the only time you’ll want to visit them is when you’re particularly keen on buying something. If a monster drops a sword that’s useless to you, you can drop it then and there and never bother about it again. Crawl also makes a point of not messing with arbitrary and annoying issues like dungeon lighting: candles and torches are a pain in the ass and rarely offer real gameplay value.

The idea of this “valuable” gameplay ties in with that of easy access and navigation: when it was first released, Crawl had a broad variety of commands that could let you get what you want in just a keystroke or two. This interface was improved even further with Stone Soup: chores such as dungeon navigation were streamlined with auto-explore options, and information was conveniently presented with mouse-over tooltips and shortcuts. In a nutshell, everything that could potentially turn the game into an uninspiring admin riot was either cut off or reduced as greatly as possible. Funnily enough, studying this game reveals a staggering amount of design minimalism in action.

Interface design: it's important. Also, what the hell?

Interface designers in particular could stand to benefit by playing Crawl for a while: it’s an amazing demonstration of how a bajillion functions and ideas can be made accessible without ploughing into dozens of nested menus and subsystems. At most, the game will only ever require you to take one (maybe two) steps away from the main game screen to achieve or learn something. And for a veteran player, this is golden.

Wrap-up: Crawl is awesome

Crawl isn’t for everyone. Let’s make this clear. It fits into a niche genre, the win rate is nearly non-existent and the graphics can be generously described as “functional” at best. It does, however, sport one of the finest game designs around and a lot of the lessons that it teaches can be carried over to more accessible and colourful projects.

If you’re interested in learning more about the Crawl design philosophy, go ahead and grab the Stone Soup build from Sourceforge and look up the doc from the in-game help. It’s a refreshing read, and if more designers take care to outline their decisions in this way, we’d probably see far more great games being produced by developers of all levels. It’s a good read and you won’t regret it — unless, of course, Crawl ends up swallowing and destroying your life. But, well, you have to take the good with the bad. And on that scale, Crawl is damn well legendary.

About Nandrew

Rodain Joubert is a South African game developer based in Cape Town, currently working for QCF Design. He likes his job. He likes being opinionated on the Internet. He likes fighting evil with his heat ray vision. And he also likes cats. [Articles]

12 thoughts on “What I Learned from Dungeon Crawl

  • DukeOFprunes Post author

    Loving this little gem of a roguelike since your write-up. Got me a troll monk at level 10, it’s a pretty potent combo.

  • Richard Terrell (KirbyKid)

    It’s hard for me to get a feel for Crawl from this article. I like rouglikes too (as you probably already know… http://critical-gaming.squarespace.com/blog/2010/10/29/desktop-dungeons-differences.html )

    Comparing the game you love to a game you hate (WoW) doesn’t do much to explain why the game you love is good. You post quotes of various design philosophies, but you don’t seem to to put them into proper context or challenge them.

    And I think it’s an big exaggeration to think that menu clicks and load screens are considerable moments of not having “fun.” This is why I tend to not use fun in my arguments. First of all, if a sound effect can be rewarding/fun, then anything can be fun including the anticipation for interactivity. Second, people need waves of difficulty including moments of low challenge/interactivity and high moments. Even roller coaster rides have this.

    It’s neat that you cover so much design philosophy ground including difficulty design, “interesting” decisions, and tedium in the medium.

    • Evan Greenwood

      @KirbyKid. Did I read that correctly? You seem to argue that spending time in load screens is desirable? Do you really mean this?

      I think Nandrew uses “fun” in the sense of: something “fun” is something you would do even if there were no reward. If that is true then any loadscreen after the first is easily a “considerable moment of not having fun”.

      I would agree that riding through the wilderness finding a colossus in Shadow of the Colossus is a vital part of that game’s gameplay. But it works because it is characterising the looming battle as an epic feat, whereas menu screens draw the player out of the gameplay and break that very sense of anticipation.

      Some games rely on compulsion and not fun (like Farmville and World of Warcraft). I will quite happily join Nandrew in saying these are not “fun” or “good” games. They do have incredible compulsion and I believe there is a lot there to be learnt from them (but not “fun”).

      Also @Nandrew, just a point about Crawl’s interface, although Stone Soup’s interface may be a lot more powerful than previous incarnations, I feel it is still pretty unacceptably unintuitive. I think the interface drives a lot of players away who otherwise may have really enjoyed the epic adventure Crawl provides.

      Of course such an interface hurdle isn’t a problem for hardcore Rogue-like players, but I believe a large part of your own game’s success (Desktop Dungeons) arrises from providing a similar gameplay experience without the steep interface challenge.

      Nice article Nandrew!

      • Evan Greenwood

        (To be fair I don’t think Nandrew ever said World of Warcraft is not good, just that he hates it and for the games he is trying to make it offers a lot of examples of what not to do, because frankly, if World of Warcraft was not multiplayer it would fail dismally)

  • vivify

    Hey Nandrew, I very much enjoyed what you had to say about game play and share the views you expressed in this article.

    I was quite in spired to make several comments but ultimately i think i’d just be agreeing with you, so the only comment i have left is to ask your opinion of Powder:
    I’ve been playing powder for years, and i truly love it. It doesn’t have the depth of Crawl by any means. But it has just the right amount of accessibility of interface and depth of game play to make it enjoyable as both a casual and a hardcore game. It does help that i am able to play it on my DS.

    BTW i very much enjoyed DD, thanks.

  • erik

    i liked powder, but i think it was too random/hard. desktop dungeons rocks because it’s both random/hard and steady progress in one game without being too much of either (avoided boring, stayed sweet for a long time). i really liked the backward strategy for vamp!

  • dpeg

    Hi, I wrote the philosophy section for Crawl (out of boredom, as it happens — I cannot code 🙂 It is very nice to see that the pas-de-faq is appreciated. What is more, is has helped to get an idea of where the game should be going to (Crawl has devs come and go, as is probably usual in projects of this type and size).

    It is a funny coincidence that Crawl’s orignal developer Linley was motivated by trying to avoid the mistakes of Nethack and Angband, and similarly for me (with the difference that I only learned from Nethack how not to do certain things) — Nandrew seems to get lessons from WoW.

    As a developer, however, I see the flaws. Two aspects of the roguelike genre make me quite unhappy, and I have no real idea how to address them (within the dinosaur that is Crawl): players are essentially forced to restore health and magic after every fight. The food clock is not enough to get around this (Rogue could get away with an extremely tight food clock, Crawl cannot). Secondly, monster spawning is too meaningless. For the latter, I found Keith’s 100Rogues interesting: he only spawns groups in rooms, so that every room is a little tactical encounter.

    Regarding baroque interface: Crawl now has a tutorial (in five stages) and a hints mode. Feedback on how they (don’t) work is very much welcome. Thanks!

  • Markus

    I must say that Crawl is sometimes too unfair. Suddenly a dozen tough critters pour over you can you don’t even have a chance to escape. That’s not fun and I doubt why anyone like it. After all, it’s a game, not reality. Look at Diablo, which is also roguelike but cleverly put all very difficult fights away from necessity. Sure that the Butcher is difficult, but you don’t really have to fight him, so is the Skeleton king as you can escape into deeper dungeons. There are difficult levels but you can always adjust your gaming style. Crawl does not even give player the chance to adjust. But of course we all have different gaming philosophies, so, well, so be it.

Comments are closed.