Gratuitous Post Mortem

Background

Gratuitous Space Battles is an indie space strategy game that aims to bring to life the role of a spacefleet designer / admiral in classic sci-fi movies such as star wars or star trek. The game was developed over the course of one year by lone indie developer Cliff Harris (Positech Games).

The original design for the game was totally different. it was not even set in space, or a war game, but was a ‘virtual dictator’ style game where the player had the role of a ‘virtual saddam’ trying to desperately cling on to power. When I started working on the code for that games map, I ended up experimenting with using space and nebula backdrops instead of maps of the middle east (I don’t know why), and pretty soon the dictator sim was morphing into a space battle sim, almost as if it had a mind of it’s own.

I’m a big fan of strategy games, but find that many of them suffer badly from the feeling of lack of control during a battle. Although naturally we enjoy playing big impressive looking battles with huge armies, it’s actually close to impossible to control large numbers of units in real time.

When thinking about this, I decided to try and design a game where that frustration was ‘designed-out’ by deliberately making the lack of control a core game mechanic. Could I make a game where the whole point was to give up control of the actual battle, but put in place everything you needed pre-battle, to ensure victory? The online challenge possibilities were a natural progression from designing non-interactive battles. The game’s name sort of jumped out at me. I hate games with generic sounding games, especially sci-fi ones, and cliches such as using ‘Nemesis’ or ‘Redemption’ or other over-used words. Gratuitous had not been overused, so I thought it would be ideal. I don’t have a billion dollars to use to tell you what my game is about, unlike movies such as ‘avatar’, so I had to use the name to say where I was coming from.

Development

Like many indie gamers, I’m basically a one-man company, so at the start, before I knew what I needed I was using any old spaceship sprites to try out the ideas. I also used coder-art for all the UI and backdrops, so the game itself looked terrible, but the basics of hands-off gameplay could be tested and seen to work well. I was originally using a fixed size map, but pestering by a fellow indie who happened to be staying at my house persuaded me I should support zooming in and out as well. (good call!).

Although eventually I decided to go with a semi-realistic look using 3D models rendered out flat, I did experiment with a Tron/Darwinia style neon retro look, the idea being that the whole game was played out on a strategic viewscreen of a space admiral. Eventually I discarded that, although it did look cool.

GSB was self-funded, and was a full time job, so with one eye on the budget I was very tempted to use stock spaceship models. I bought the licence to use a number of 3D models and did some test renders and playtesting, but they just didn’t look the part. There was the dual problem of them being ships seen in other indie games, and them being very low poly models, because they were designed to be used in 3D, not rendered as 2D sprites. With custom-made models, I could basically go poly-insane, as they would be rendered to sprites anyway.

The Search for Art

A long, long search for the right artists ensued. Eventually I found a guy (Joshua combs) working on a freeware Masters of Orion clone who produced excellent quality background nebulas for the game. A long while later I finally found my starship modeler, Charles Oines. I got quotes from a number of artists, and ended up going with the most expensive one. A pretty scary decision when you are self-funding an original game and your income from previous games (which is what I’m living on) is going down and down every month…

Anyway, it turns out to be well worth it, the spaceship models were perfect, and looked fantastic in-game. Originally I had planned on re-rendering out many variations of them, but as luck would have it, the models were done in Lightwave, I only have 3D studio Max, and learning + buying lightwave was not an option. Converting between the two proved extremely fiddly. I also had the task of doing the damage textures myself, which involved about 10 different layers in photoshop and using a clone brush and a picture of an oil rig at night to get the right ‘burning spaceship insides’ look. I also spent hours taking hundreds of still frames from a Star Wars DVD to get the right textures and effects and style for laser effects and explosions. I ended up writing a whole particle system configuration and testing program to get better explosions. (see img)

This takes us about 9 months into development, and I had a semi-finished working game. I then showed it to some people to gauge their reaction, and got tons of great advice on how to improve the ‘flow’ of the UI, which was frankly a disaster. You had to build ships, then build a fleet and save it, then load that fleet into a deployment, then fight a battle. The fleet composition UI was fiddly and silly, it all belonged on the new deployment screen.
I’d also coded a semi-multiplayer asynchronous online challenge system for the game, which I thought was a great idea but was wary of assuming anybody else would. Eventually, I found out people loved it.

Once the UI was improved and a decent logo done (thanks to a separately hired UI artist) and the excellent music was delivered (from Jesse Hopkins, who I’ve used for several games now), it was time to put GSB into beta testing, something I’d never done before. I decided to sell pre-orders for the game at a discount, and open up the beta to everyone who pre-orders. This was great because it meant getting real unbiased feedback from players before release, and meant I didn’t have to panic and rush-release the game to pay the bills, which was becoming a real threat.

Release Day

The beta went on from August 31st 2009 to 4th November, and was invaluable. About 12 days later the game became my first title to be launched on Steam, and had already been available through pre-order on a number of other sites such as impulse, gamersgate and direct2drive.

The minute all the bug fixes were under control (there are always bugs!) I threw myself into getting an expansion pack ‘the Tribe’ put together for the game. this was another first for me, I’d never released an add-on for a game before, and that seemed to work out well too, acting as a useful test of how moddable the game could be at the same time. In between doing the pack, the game continued to expand and get improvements, with the current total being twenty eight updates since release. In the middle of all this, I moved house, which added to the general chaos and panic and lack of sleep around that time

In total, GSB took about a year of full time work for me, plus help from various contractors. I work pretty long hours, so I’m wary of working out eh hourly income in case it’s tragically low . The source code for the game comprises 581 source files of C++ for a total of 76,048 lines of code for the game and a further 20,852 lines of code for the base engine. The game was developed using Visual Studio and Visual Assist, and uses directX9 in a custom-written engine designed around the game. The size of the code is roughly double that of my 2nd best-selling game which is Democracy 2. There is also a bunch of php and SQL code that runs on the server, outside of the main game, and piles of config data for all the ships and weapons.

Budgets

In terms of budget, the rough cost of everything outside my own time is $14,000 for the basic game. I add to that the cost of advertising so far which is about $5,600. The game is actually making money, which is a relief, because my games take a year to make, so the plan is to have another game ready by the time the income from this one runs out

Comment on this post in our forums