On the Commercialization of Roguelikes (And Some Other Stuff)

(David recently did a State of the Dungeon post, and I guess this is my State of the Dungeon.)

For the most part it is rewarding for us, as developers, to hear that people are excited about the work that we’re doing and how badly they want to pre-order our title. (See, we’re listening!) I think that this is true of any creative endeavor: the artist thrives on the energy of his or her public. At the same time, it is also worth noting that it is very easy for a developer to get unexcited about a project when people keep telling you that you’re doing a lousy job. Most developers will tell you that the secret to this is not to listen to people on the Internet – after all, what do they know? That said, most successful developers – Valve springs instantly to mind, under the capable direction of Gabe Newell – will tell you that the success or failure of a game, and of a game company, is dependant upon your fans and your customers. Listen to people, get them excited, and you will prosper. Alienate your fans and you will alienate your customers, and your customers are the people who pay you money (or who will pay you money once we get the pre-orders going.) So that’s why we listen: we secretly want to be Valve. (Who doesn’t? Ben McGraw, our executive producer, recently pointed out that Valve is one of the few game companies in the so-called “Industry” that he would work for. Like indies, Valve brings joy to people. Other companies, he says, just make games.)

So what are our users saying? In a recent poll on our Twitterfeed – which you should all be following, because it really is the best way to keep tabs on us – one user wanted to know more about our development process, and the day-to-day decisions behind game development. We hope that this will oblige you, but today’s blogpost is *really* inspired by something from The Internets.

In a recent online discussion about Dungeons of Dredmor, somebody said – and I paraphrase: “Commercial roguelikes will never be as good as free roguelikes because the multi-year, evolutionary development process that results in amazing games like Crawl will never be commercially viable.” Here we have somebody who likes roguelikes, and who should like Dredmor. Hopefully,  he will support us – here is a man, after all, who could be a customer, and any failing to attain him as a customer is a failing on our part as a business – but his concern is legitimate. Can a commercial roguelike be as good as Crawl, or Nethack? Well, I think we can… but let’s talk about this.

First off, it is worth noting that Dungeons of Dredmor has had a multi-year evolutionary development process. Dredmor, as a product, started development in 2004, about four years before Gaslamp Games existed in any way, shape or form. Dredmor originally started as a project code-named “Orion” – in fact, this is still the name of the Visual Studio solution – and until Ben McGraw made his major contribution to the project by naming it “Dungeons of Dredmor” it was going to be called “Heroes: Wanted”… blech.

Now with 95.2% more programmer art

The original game would see players taking on one of ten character classes, including some traditional ones (Archaeologist, Tourist) and some new ones (Pirate, Viking). Some elements of this have survived – you can still choose Archaeologist as a career path – but the fixed character classes disappeared after I realized that the amount of animation required wouldn’t be cost-effective. This is also why David has redrawn everything in the game except for our character art, which was done by two very talented outside contractors, Bryan Rathman (now at Blizzard) and Tim Wexford (who I believe is still doing the independent contractor thing. It’s been awhile.) I brought the game to Gaslamp in a semi-finished, and quite frankly a confused state, and the Gaslamp crew has been working within the framework established by my early efforts six years ago, trying to straighten it out. For the most part, we have suceeded, and the game that we are preparing to ship is very close to my original design parameters.

Enjoy more programmer art from 2005.

I wanted to create a roguelike which had the depth of Nethack or Crawl, but without the burdens that traditional roguelikes impose upon the player. This succeeded. I also wanted the game to be numberless and parameter-less. This was inspired by Weird Worlds in Infinite Space as well as an article on Gamasutra by Ernest Adams. This was a disaster, and I am very pleased that David and Daniel finally managed to talk me out of it. The game is just so, so very much better with numbers. You can see the original, numberless game in our early trailers. Compare that to Trailer #4, and tell me which game looks more inviting and easier to play.

Some of the design decisions that I have made on Dredmor have been very poor ones indeed. Some of the coding decisions have also been very, very bad. We fixed most of the bad design decisions, and a lot of the code decisions are still causing us pain and agony. Still, you live and you learn, and I have another game coming up in which I can apply the lessons from Dredmor. (I’ll just have to create new ways to screw up.)

We went from ten character classes to a set of mutable skills based on the game “Lost Labyrinth“, a roguelike written in PureBASIC which did a lot of fairly interesting things. The mutable skills evolved from being regimented skills in a Nethack vein – before Gaslamp got ahold of the game, you learned spells from spellbooks and they were deployed using a spell ring, similar to Secret of Mana – to the skill tree that you have today. This was based on an observation by Citizen Daniel that players needed to be able to do something if they didn’t want to play wizards. At one point, we had a templated system which let you combine abilities (knockback, fireballs, etc.) with attack patterns (T-shapes, L-shapes, boxes, odd circles, et cetera.) It was… interesting, but it didn’t really work out and so we ended up killing it in favour of the current set of warrior abilities, which are a mixture of passive boosts, randomly applied special abilities, and targeted abilities with cool-down periods. The only thing we ended up keeping was the concept of cool-down periods for non-wizard spells, as well as the actual code for drawing oddly shaped attack patterns.

with spell ring lifted from Secret of Mana

The actual mechanisms used to select skills and work with items evolved, and evolved, and evolved. At one point, we had a belt full of items, but the belt could also contain skills… it was, in hindsight, very odd, but finally we ended up doing what we do now, which is so clearly the right thing to do it’s not even funny. Before we added tooltips to the game in 0.8 or so, we had an All-Seeing Eye tool that you could click on objects – in a two-stage process – to figure out what they were. Again, a bad idea. Again, we figured it out.

We had polymorphing at one point in the game, but it didn’t make the cut. It adds a lot of depth and strategy to the traditional roguelike, but… well, it just wasn’t fun. I may see if we can sneak this in later.

Shopping! Shopping went through a number of evolutions. I always wanted shops to adhere to the roguelike tradition that Theft is Good – but it turns out that this is very hard to make work in a graphical system. We only cracked this earlier this year when we added Shopping Pedestals. Before that, life was… not good. (Again, you can see earlier shops in earlier trailers.)

So, yes, Dredmor has evolved, and the evolution has taken us years. Not all of this evolution has been in the traditional form, where layers upon layers of content have been layered on the basic frame of the roguelike like a deranged wedding cake. (That said, this has definitely happened in a few places.)

This brings me neatly to my second point: much of Dredmor, in its current form, did not exist until fairly recently… and by fairly recently, I really mean “March.” After a sequence of false starts, the Gaslamp development process has turned into a cycle of repeated iterations. At each stage of the process, we ended up with a playable game. Whenever we get through a milestone, we do some playtesting and then take a look at what’s working and what isn’t. We keep revisiting the sticky points until we figure out how to fix them. It has turned out, oddly enough, that there is a Right Way to do a lot of things. Stairs, for instance, were an unsolved problem that caused us no end of trouble until David finally solved the mystery a week and a half ago. (I’m going to let him write about this one.)

Our current magic system didn’t exist until fairly recently. We had Earth, Fire, Water, Air, Light and Dark Magic. This was a holdover from as far back as 2005, and it was not really a good holdover. Finally, somebody snapped and we decided, okay, we need to come up with new, Dredmor-appropriate magic schools. This is how we ended up with Fleshsmithing, Golemancy, Necronomiconomics, Sun Worshipping, Viking Magic, Promethean Magic, Mathemagic, and Psionics. Much better. We wrote a blog post about this that will tell you exactly when we came up with it. Heck, our combat mechanics are new as of the start of this year. We went through three systems – my original collection of hacks, a system designed by Citizen Daniel which included, amongst other things, logarithmic damage multipliers, and finally a system designed by David to be idiot-proof and simple, which was then lovingly prototyped with a sequence of Excel spreadsheets until we were sure we had a winner.

Traps did not exist in their current form until our giant push in March and April. We tried a bunch of stuff, but we only figured out what we wanted and how to make them work much later. Another recent addition is the existence of special treasure rooms – scriptable ones, driven by XML used as a domain-specific language – that break up the dungeon’s monotony.

In short, much of what makes the current iteration of Dredmor special and sparkly, fun to play, entertaining and cheerful… is pretty darned new. Some of it isn’t. Artifacts have been around since 2008, our basic movement schemes have been around since 2006, and there are little footprints from the various iterations and builds of Dredmor that have happened over the years sprinkled across the game like so much confetti.

Development time, as it turns out, is not the only factor in terms of making things great. You don’t make a set of great decisions, one after another, in game development. Instead, you put down a floor and rip it up time, and time again, until at some point you feel comfortable putting up the walls of the house. Implicit in the assumption that it takes ten years of development to make a brilliant, deep roguelike is the idea that everything you do in that ten years becomes part of the finished product. Game development doesn’t work like this. As we are increasingly realizing, there is no game developer who has The Brilliant Formula that turns everything that they touch into gold. This is why a game developer is only as good as their last project. (Ask John Romero.)

Here’s the third thing, however – what does it mean for a game to be commercially viable? The Bay 12 Games guys release their income figures every month, explaining how much cash they’ve hauled in as a result of Dwarf Fortress. Most months they seem to make about $2500 USD/mo., some months they make about $4000 USD/mo., and when Dwarf Fortress had a major release in 2010 they pulled in $16,000.00 in the month of April. This isn’t a success on the level of Minecraft, but it is enough that Dwarf Fortress continues to be developed actively and on a full-time basis. Remember, this is by *donation*. Commercially viable means a lot of interesting, wonderful things when you’re a small studio. If, by the grace of the game-buying public, Dredmor can pull in four thousand dollars a month and can do so consistently with occasional spikes, we will have leeway to keep building weird and wonderful things. My Business Acumen people, of course, want to pull in way more money than four thousand dollars a month. We think that the game will do better than that in terms of turning a profit, and we still have a few tricks up our sleeves to make this happen. (everybody still likes buying 200 hats worth of DLC, right? Right. Good, we’re done here!)

I think we’ll do fine… but it’s important to note that we don’t need a AAA return on Dredmor in the range of hundreds of millions of dollars – or even Notch-level, in terms of tens of millions of dollars – for Dredmor to be a financial success. I have always viewed Dredmor as a means to an end – it has enabled us to build a company, to build an infrastructure, and to discover how to make independent game development work. From these perspectives alone, it has succeeded beyond my wildest dreams. Finishing Dredmor has put us in a great position for whatever comes next. We have some technology, unrelated to Dredmor, that we are going to finish and release and which I am very excited about, and then … well, there will be another game. We haven’t finalized what we’re doing – we have a bunch of ideas, and we’re going to do some prototyping in-house to determine which ideas are good ideas and which ideas are just awful – but looking at the short list we have so far, I can tell you that pretty much all of them are 100% TIGSource Bait. (A surprising number of them also involve wizard towers.)

So, yes. We think Dredmor is a pretty darned good Roguelike. It certainly delivers an experience that no other game does – neither Crawl nor Nethack has amusing descriptions of common minerals, for instance – and there is definitely depth and skill involved in playing it. David recently survived a near-death scenario in a recent play-through by building a crossbow out of spare parts in his backpack, then used it to blast his way out of a nasty situation by deploying his carefully hoarded Squid Bolts (don’t ask).

Is Dredmor on the level of Crawl, Nethack, or any of the other ridiculously content-filled rogues out there? We’ll see. I can guarantee, however, that it has more depth than Pokemon: Mystery Dungeon. (Actually, it’s a lot like what Pokemon: Mystery Dungeon would be like if Squirtle got his name for a terrible, foul reason that would make the game entirely inappropriate for children. But I digress.)

My personal belief is that skeptics will be pleased. I hope, regardless of anything else, that you will give us the opportunity to make you laugh, make you smash your keyboard, and to keep you blissfully entertained by buying Dungeons of Dredmor. I know we still haven’t announced a price point yet, but we have one in mind, and I really don’t know what game out there – indie or non – represents more of a gaming bargain. More than that, I hope that today’s post has convinced some skeptics out there that yes, you can make a go of it by building a commercial roguelike. We’re not the only ones trying. Desktop Dungeons has made a go of it, and has been rewarded for its take on the genre by none other than the Independent Games Festival. Dwarf Fortress… continues to be a fortress, full of dwarves, and that’s about all you CAN say about it. And then… well, there’s us.

I will also note that there *will* be stuff happening to Dredmor after we ship. I don’t know what, exactly, but I think there will be a bugfix patch after 1.0 (stuff always, ALWAYS shows up post-1.0), a well-deserved holiday, and then we will see what happens. I have a shopping list of things I’d like to see in the game, and I know that David does as well. There is a bugtracker target labelled “2.0”, where all our bad and good ideas go to be locked away until the universe is ready for them. (A few random samples: “bees you can throw at people”, “yoghurt maker”, and “add marinating and grilling mechanic for food items”… okay, the last one we might do anyway because people keep clicking steaks on the BBQ and expecting it to work.) One day, soon, we may crack that vault open, and wonder what on earth we were thinking. I can say, with some certainty, that one of my priorities will be improving Dredmor’s mod support so that people can graft their own content onto it. An open source release, in the fullness of time, is not outside the realms of possibility either. Both these things are fairly obviously the Right Thing to do, and it’s simply a matter of time.

So there you have it. Can a roguelike be a commercial success? I don’t know, but I think it’s a mistake to underestimate what a commercial roguelike can, and cannot do, and what it means to be successful. Have we suffered as much as any other roguelike developer? Oh, yes; there has been a great deal of polishing, refactoring, pain and suffering in Dredmor that is not visible in the current product, but which is integral to making Dredmor the game that it is today. Is Dredmor a product of over six years of learning? Oh yes, yes indeed. What’s it like to be a professional game developer? Why are you asking us?

The future is fabulous, and I am incredibly excited. Happy Easter, everybody.

8 Comments

8 Responses to “On the Commercialization of Roguelikes (And Some Other Stuff)”

  1. Tyrell says:

    Squirtle got his name after a little incident that happened after he heard about the impending release of Dungeons of Dredmor!

    { reply }
  2. J Brian says:

    I love making games for a living, it’s really hard to imagine a better job. But I do wish I could work for a company that truly made it a goal to delight their customers, rather than just making something good enough that customers will pay for. Definitely an issue at the publicly-funded company level. I wonder how long Valve can hold out as a private concern.

    Also 4 grand a month doesn’t sound like enough for 2 full-time employees and some additional help. Is Gaslamp not your full time job?

    { reply }
    • None of us, at present, are full-time employees of Gaslamp (although we would all like to be.) So we have other sources of income, mainly derived from a combination of contract work, unemployment, and people being various sorts of students. Obviously, we would like this to change. $4k… is probably not enough to do anything real, but if it was a consistent $4k it would be enough to let us bootstrap our way to making a game with significantly higher production values than Dredmor *and* getting some of the technology stuff that we want to do done (which in turn increases revenue.)

      Valve is estimated to be worth $4 billion – you got that right – by none other than Forbes magazine. They’re not going anywhere and they’re certainly not short of money.

      { reply }
  3. Ben McGraw says:

    I’m pretty sure I contributed much more than just the final name, Nicholas. 😉

    { reply }
  4. Brian says:

    Good stuff. Though I’d say it is a tad more of ” western viability on PC” as Chunsoft has been rocking on along for…over a decade now. Though they haven’t done any PC ports, and certainly now localized, in a in a good many years. Dredmor have any internationalization aims towards Japan or elsewhere down the line? Voyage to Farland on Android has been trying to make a go of it on that front and, IIRC, has been relatively pleased with the reception.

    As with Valve, Minecraft, DF, etc—there are myriad examples of little good ideas that any game developer, indie or otherwise, can harvest and apply as appropriate as while The Magic Formula has yet to be alchemized, Enchanting Equations are at least well demonstrated.

    { reply }
  5. exit says:

    “I wanted to create a roguelike which had the depth of Nethack or Crawl, but without the burdens that traditional roguelikes impose upon the player. This succeeded.”
    you’re kidding, right?

    “[Chunsoft] haven’t done any PC ports”
    what? yes, they have.

    “certainly now localized, in a in a good many years”
    uh, how about the two that were localized LAST year?

    i guess the mangled development cycle explains a lot about the present state of the game, anyway.

    { reply }
  6. Pingback: Les Rogue-like et les Dungeon Crawler (ou le masochisme dans toute sa splendeur) – Part I » Gato de Bento

Leave a Reply

Your email address will not be published. Required fields are marked *