Das Geisterschiff is a dungeon crawler developed by Graverobber Foundation and released on September 19 on Itch and November 24 on Steam. It initially started as an idea for a very short game inspired by Carmine (developed by ArrowSoft and published by MicroCabin in 1986), set in a distant future and taking heavy inspiration from HP Lovecraft.
The goal was to create a project small enough to deliver within a reasonable amount of free time, on my own, on ~$0 budget and with a good amount of polish. It was driven by arrogance and frustration about the game industry as a whole and especially in my country -- a place that used to offer interesting projects, such as Chasm: The Rift, Cossacks, S.T.A.L.K.E.R, Metro and Cryostasis, but now mostly focused on mobile and social gaming with only a few worthy exceptions. Since the job I had at that time wasn't satisfying and game development-related employment wasn't offering anything of interest, it was decided to start clawing my own way into the industry.
Of course, it wasn't without its fair share of things done right, things done wrong and a handful of happy accidents. That's what I would like to talk about here, but first, let's take a look at the whole dev cycle.
Development
The game managed to go through four iterations and changed drastically with each one of them. What started as a very simple adventure game disguised as a dungeon crawler and inspired by HP Lovecraft eventually transformed into a Blade Runner-style adventure by the end of second iteration. Unity was chosen as an engine due to my familiarity with it. In early 2017 an artist (Victor) joined the project to handle 2D artwork.
However, even by the end of second iteration, the game was a mess -- random bugs, less than stellar performance on target hardware and frankly, it wasn't fun to play. Short-lived third iteration attempted to fix that by rewriting it from scratch, but by mid-2017 it was evident that it needed some time on backburner while I improve my programming skills and take some time off. Around that time Greedwalker joined the project as a guy who could bounce ideas with me, playtest the game and shitpost about it online from time to time.
On 2017/09/19 development started again from scratch. It was decided to cut down on unnecessary tasks and put it together as quickly as possible and reach playable alpha state before the year ends. An ambitious plan, and it didn't fail -- Tutorial Demo released in November after about two months of work proved that it's more than possible to deliver it.
Around that time Victor had to leave to focus on his studies, and the amount of finished artwork from previous iterations set a strict limit on what could be done in terms of story and character interactions. Story was cut down from six short chapters to two lengthy with a short intro and epilogue -- it became clear that the sick and disturbed layer of the setting had to be conveyed through other means.
By early 2018 I've managed to speed up the creation of game content and had the majority of systems in place. Another demo was released in February -- this time it offered two small segments from Gangland chapter and was also sent to GTP Indie Cup to get some more eyes to take a look at it. Reception was mixed, from 'I don't get why this game exists in 2018' and 'unplayable' to praise of its style and execution. Things went according to plan and by April the Gangland chapter was ~90% complete, while the Ghost Ship chapter was ~50% complete as well. I aimed to release DG on June 1st and spend the summer on support and bugfixing.
However, by mid-May it became apparent that the game would need more time, so it was decided to release a new demo along with completed half of the game as a beta to gain more time and (hopefully) some additional funds to finish the game. It also became apparent that the game will need some sort of tutorial to help players familiarize themselves the game mechanics. In late June I had to switch jobs, which somewhat slowed down the remaining work on the game.
Eventually the tutorial was released and the last chunk of the game was completed in August and early September. The remaining week before launch was spent on tweaking the game and checking how long it would take to make a Linux and Mac version. As it turns out, both versions didn't take too long to make and the game launched on September 19 on Itch with Linux and Mac support on day one.
After some patching and tinkering the game launched on Steam, but not without a fair share of problems. At v1.0.2 I've changed the way RNG is handled in the game, which made the encounters more deadly, but also made the encounters less frequent, which required changes to the game balance and made the game somewhat shorter than expected. As of v1.0.5a I've managed to tweak it to the point it was back at intended difficulty. Currently I'm working on the update to expand the game a little bit and do some final tweaks to make an already polished game shine even brighter.
Considering that final iteration was only 12 months in development, I consider that a success and a good first step.
What Went Right
1.
Well-documented engine
Unity gained notoriety among players and some developers for a variety of reasons, but good documentation combined with my (somewhat limited) experience with it proved vital to get this project out the door. Available solutions, such as Aron Granberg's A* Pathfinding plugin and ProBuilder also helped a lot. It was also an attempt to prove that Unity can dance in the right hands.
Another reason lies in the fact that despite being a modern engine, the base editor is pretty lightweight and C# ended up being a pretty easy language to start with. Since the game was designed on low-end hardware and targeted at it, choosing a different, less-documented engine could cause more issues down the road than it would be possible to handle.
2.
Simple, interconnected and nuanced mechanics
Exploration and interaction were tightly connected with each other from day one. Each object type had to have as many use cases as possible to keep the simple mechanics interesting enough to carry the whole game and they had to be tied to stats to make sure that RPG elements aren't just a bunch of meaningless numbers on screen. Even combat was always conceived as part of interaction with the game world and almost every mechanic had to be usable outside of it as well.
At the same time dice rolls had to be used for something more than determining a simple 'success/failure' result. I let the imagination run wild for a while to figure out what could happen in this situation and implement as many quirks as possible into the game. As a quick example, during a battle attacker could:
- Miss the target, but hit a landmine behind it and thus get the desired result, albeit in a different manner.
- Miss the target, but the target miscalculated the trajectory of the shot, evaded in a desired direction and both the projectile and target ended up in the same spot.
- Miss the target and hit a door behind it, which could render it inoperable and cause trouble after the battle.
All three can be encountered in the game and in my opinion it works much better than a simple 'shot didn't connect/target hit' result. In my next game I would like to find ways to expand interaction with game world and add more quirks to keep the game more complex, unpredictable and fun to play.
3.
Tight schedule
Last iteration took just 12 months from start to finish with a few months of additional patching on top to fix bugs and improve the game. Rebuilding a game from scratch in 12 months is something even large teams would struggle with, and with one-man team this problem is multiplied.
On the bright side, it left no time to waste on fluff. Things that didn't add much to the game but would require a lot of time to implement were cut at the design stage, especially when it came to our weak side -- the graphics. This saved time and allowed me to focus on more important things.
4.
Sticking to the vision
Das Geisterschiff is different to the games in the same genre, and that was intentional. I made no compromises with those who weren't familiar with the genre or weren't 'getting' what the game is about. Certain features were intentionally not implemented (e.g. full map, automap, etc) to make even a simple task of finding your way through the dungeon more challenging and more rewarding.
Instead of making the game more 'accessible' by dumbing down an already simple game or making it closer to the 'generic' example of games within the genre, a lot of effort went into giving the player enough hints, both obvious (tutorial) and subtle (details) to make the learning curve less steep without lowering the barrier of entry. The remaining challenge was up to the player to tackle, and that trust in player's ability to experiment and catch hints within the limited amount of information may have narrowed the audience, but I don't have any regrets about it.
That doesn't mean that feedback was ignored -- listening to what people had to say helped make the game better, but it was carefully filtered to avoid compromising the vision and 'fixing' stuff that wasn't broken in the first place.
5.
Ideas kept pouring in
Some stuff was added at the last minute to add more spice to the game. Crippling? Added the night before submitting the demo to the Indie Cup. Terrain not only affecting the damage, but also giving evasion bonuses and penalties? Added around August 2018. Experimental 'carnage' feature that connected missed shots with the rest of the environment? Added during the beta stage.
There was no shortage of ideas and every minor addition helped shape the game into a better, deeper and more nuanced product.
What Went Wrong
1.
Poor level design process
The way graphics style was implemented eventually turned level design into a tedious process that felt like designing two levels at once. Also, due to technical limitations each chapter had to be cut into smaller levels, which in the end worked out to my advantage, as I could design massive, multi-layered dungeons from these tiny segments. But even then, process took longer than expected and wasn't half as enjoyable as designing a Doom or Quake level. This is one of the first things I would like to address in my next game.
2.
Marketing
You'll be surprised to learn that a lot of time was invested into attempts to promote the darn thing, but it didn't pan out in many ways. First, the game is not that interesting to watch, so attempts to get streamers/let's players to try it out didn't work out well -- those who actually went to try it ignored even the basic prompt like 'Press F1 to review controls' and spent a large chunk of playtime ramming doors and struggling with the basics.
The game did get a few previews and managed to get some interest in smaller communities, but that didn't attract much hype before release. The last hope was getting someone with a more or less good taste to review the game once it's out and hear what they have to say, which worked out rather well.
3.
Constant shifts in design and content
The game used to shift a lot during the development due to various roadblocks along the way as it became apparent that some content and features required more time or resources than I was willing to allocate. The project had to be refocused too many times to fit into the design and resource limits, and because of that the story had to be simplified until it became just a background and glue for the gameplay and levels, which, on the bright side, forced me to make them work better on their own.
4.
Part-time work on the project
Even though I was working on Das Geisterschiff full time during the last few months of development, the game would benefit a lot if these 12 months were spent working on it full time rather than part time. It would give more time to develop better solutions instead of relying on 'fastest' in terms of development time, which would eventually come back and bite me in the ass. Sometimes this would also drag the progress, but eventually I've managed to find the right balance to keep the game on schedule without compromising its quality.
Happy Accidents
1.
Lack of NPCs
Due to low amount of artwork the amount of NPCs and chatter had to be cut down to the barest minimum. I firmly believe that if there's no artwork to support the ingame text, then the text has to be very good to make up for that. While this decision may have turned away story-oriented players, in the end, it added to the barren, lifeless atmosphere that I wanted to create.
2.
Graphics style
What was initially planned as a shortcut using a quick edge detect effect to mask my lack of art budget (and skills) ended up being reworked into its own solution with handcrafted lines in 3D space. This made the game lightweight while also created an eye-catching style that was easy to use and stylish enough to look good and accidentally more 'retro' than the majority of 'retro' games out there. It's a polarizing style, but I firmly believe it gave DG its own identity and made it instantly recognizable among modern games.
So, What Now?
Das Geisterschiff is out, it received positive reception and it is probably the first dungeon crawler where terrain matters in combat. It sells pretty slowly, but it already found a small audience, and the sales are goind well enough to keep me afloat and start working on another game. Overall, it ended up being a good game that I'm proud of, but I can't deny that there's room for improvement that separates good games from great ones.
There are plans for a small offshoot to DG as a 'recreational' project for myself and an attempt to deliver a better and more complex game in 2019. Afterwards, I would like to try other 'crazy' ideas, but eventually I would like to come back and make a true DG2 -- a familiar, yet different beast that takes the original ideas to the new extreme.
