The Pile of Fail
or, more accurately. "The pile of iterations from which I learned to do better"
We're seeing a huge trend of new companies and projects starting up around here. It's really great and amazing watching us grow like this! Yay SA gamemaking!
I'm worried that the more established among us are greeting this passion with what seems to be heavy handed scepticism. We know why we're doing it, because we understand survivor bias and have some experience that leads us to react "badly" to exciting statements like 'making my own engine,' and 'open world MMO'.
This thread is to help people see a better picture of where we're coming from. Help take down that survivor bias and tell us about your most embarrassing moments. Tell us about the stupid legal tie-ups and broken promises and all those assumptions you made along the way that you were too stubbornly ambitious to let go quickly enough and hurt your progress.
It can potentially also serve a resource, people could see parallels between their current struggles and your history. We need more post mortems for the incidents where we fell over and got back up.
(I'll post my own too, once I evaluate a couple of things, like how much embarrassment I'm ok with being here)
We're seeing a huge trend of new companies and projects starting up around here. It's really great and amazing watching us grow like this! Yay SA gamemaking!
I'm worried that the more established among us are greeting this passion with what seems to be heavy handed scepticism. We know why we're doing it, because we understand survivor bias and have some experience that leads us to react "badly" to exciting statements like 'making my own engine,' and 'open world MMO'.
This thread is to help people see a better picture of where we're coming from. Help take down that survivor bias and tell us about your most embarrassing moments. Tell us about the stupid legal tie-ups and broken promises and all those assumptions you made along the way that you were too stubbornly ambitious to let go quickly enough and hurt your progress.
It can potentially also serve a resource, people could see parallels between their current struggles and your history. We need more post mortems for the incidents where we fell over and got back up.
(I'll post my own too, once I evaluate a couple of things, like how much embarrassment I'm ok with being here)
Comments
To start off with what might be a somewhat more charming example: back when I first got access to a computer, I played a number of games and decided I'd try my hand at making one myself, as one does. I was still pretty inexperienced with computers but I had just enough knowledge to be dangerous (there are many other stories where I almost destroyed the PC). Now I knew programs all had .EXE extensions, and I knew Paint could make bitmaps with .BMP. So I opened Paint, drew a player avatar, a maze that the avatar was meant to navigate, wrote out some instructions in words and then pasted those in the corner of the canvas. Then I changed the extension from .BMP to .EXE (I was pretty proud of knowing how to do this) and tried to run the file. I obviously had ridiculously high hopes for what a computer could do back then. Fortunately, it just gave an error and failed over. I shudder to think what might have happened if that image was interpreted as machine code on those old OSes, but I still recall being kinda sad that my creation wasn't successful. So I guess that's when I started making games failing... :)
I think my next attempt was an open world RPG with Visual Basic. Needless to say, that' didn't get very far. Unless you consider wondering around an open empty void a great game, in which case I nailed it ;)
I've learnt an enormous amount, mostly thanks to the incredible people I've been fortunate to work with, but I'm really happy that I didn't have to take the risks to make these games on my own pocket. Currently, I'm building up skills and savings so that if one day I want to make my own games, I can do so with the greatest chance of success. And I also know that I'll show it here really early on, because if it's a game worth spending more years of my life on, then I'm pretty sure I'll know it by the number of people who want to help me work on it... :D
I worked on many failures. Oi!
One of the most embarrassing projects was an Angry Birds clone. (But with cats). Although some of it was somewhat interesting, the brief from my boss was to mimic their features as closely as I could. I actually completed playing Angry Birds (close to getting all three stars for everything), which really made me hate a game that I would otherwise have just dismissed. At one stage, Unity's physics was not good enough and I had to investigate using Box 2D physics instead. It was a totally demoralizing project; almost every decision we faced was "solved" by looking at what they did... not by reasoning or testing. And it all happened as the Desktop Dungeon clone drama played out... Being paid to clone was quite a low point. (I ended up quitting the job after the project was finished, and as far as I know the game never ended up being released, nor did I hear anything from the company again.)
Another one was a type of role playing adventure game... The Adventures of Gi. My job was to make sure we ship the thing. There was a fixed budget, and little room for being late. There were several problems with the project, but the sheer scope was just unbelievable. There were several fights to keep the game not going MMO... and although we cut and cut and cut, it was still too much. And all the cutting made the game strangely empty (not elegant). So we were late, and the game was not fun, and the single most unfun thing was something I had to build (the controls), which I just couldn't get right even after I-don't-know, 9 rewrites. (Another programmer also did 3 versions, if I recall before I took it over.) The game was published, but with no money left no marketing could be done, and the company went semi-bust.
During a my contracting phase I experienced many horrors. There were two projects that I could not deliver that I took on because I was so desperate for money at the time. (Contracting work pays well, but the ebs and flow of cash is not consistent, and I fell for the "client did not pay us we can't pay you" trick a few times and landed in temporary cash crises more than once.) In both these cases I thought I would be able to do the work, when it turned out I couldn't I did not say so immediately. Ugh it was really horrible. Having to admit it, having the client's project fail (which happened to one) was really low-points.
Another semi-horror was a fairly simple game (a prototype originally scoped for 3 days) that went fairly smoothly. The client requested some changes, and then more, and then yet more... And then he could not download the game (I made web builds up to the point of delivery), so I had to post disks to him. And then he couldn't run it because of Norton. And then ... it went on like this for some time. And it started to overlap with two other projects that were scheduled to run one-after-the other, but because of delays happened simultaneously. And both those involved specialized hardware that required me to work on the clients' premise. And both had very strict deadlines (the one was for an event, another was to showcase to someone from China that was only in the country for a few days). At some stage I was working from 6 am to 2 am (which you know you cannot really do). By some miracle, I managed to deliver the other two projects ... but neglected the smaller one that was still lingering. That went into a bit of a legal dispute (which was silly, I think it was like R8 000 for the entire project).
Then there was the Dota-clone (but clone in the ethical sense). I actually enjoyed working on this project, and I wouldn't group this under the embarrassing heading. I joined it fairly late, taking over from another programmer. A lot of art was already done (some 50 robots - that was the theme), and a large chunk of the programming was done. I had a list of very specific things to do, and I could roughly do them in about 6 months I think. And I basically succeeded, but it was very rough around the edges. The problem only really came in when the client realized that someone had to actually design the whole thing - set the attributes of the 50 robots (each had like 30 numerical attributes), several dozen items (with several attributes each), and that it will be a lot of work. He initially thought that it would be quick to do, and I think he planned to do it. At the time, he spent quite a bit of his own money on it, and did not want to spend more money to pay a designer. Like I said, this was a fun project... but a good example of how to not build a game.
I could probably list about 10 more. Of these, I enjoyed some, and some I were proud of, but none I'd like to repeat in the same way. There was a few death marches, a few projects with really crappy graphics, projects taken over from other programmers with really really bad code (in one case, with intent on deception), one project with really horrible technology (Blender!).
Because I worked on so many marketing games, a lot of the projects under my belt is not particularly... glamorous. I sometimes wish I worked on more projects that I could really love (not just technically, which I normally do, but game-aesthetically too).
Then there were a few that ended up being cancelled despite being fairly decent. The one was a space-hover craft racer which was really fun (but for "business reasons" it was decided to make Angry Birds clone instead); another one the client changed their marketing strategy and cut off all funding. Another very cool project was actually an educational board game for a power company, that also for reasons on their side did not go ahead even though people that played it enjoyed it. (What is particular sad is how the artist managed to make something cool and pretty within the strict bounds of their corporate identity that I thought was hideous and not suited for games at all...)
I did certainly learn a lot. One nasty lesson is that despite intentions it's possible to make certain mistakes twice. (If eminent failure was obvious, I guess, there would be less of it around). And I cannot say that there are no things I regret, but there are not many. They do make you stronger. And of course, despite all these things, you work with amazing people, and there is a lot of joy; and that I am very happy about.
@FanieG: We have them, but I have no idea if they would be usable or not. We made them in an ancient Quake 2 modeling program. Vertex animated and everything (no skeletal system).. That, and they look like eldritch abominations that hit every branch on the ugly tree on their fall from the sky of unskilled hands..
So for my part, I'll go through this briefly.
In about 2008 I spent 3-4 months building an animated intro for a game that I barely had a prototype for. I eventually realized that I needed a game to make the world building I was doing worthwhile, and when I started considering how to approach that I saw that the task I was setting myself was insurmountable given my skillset.
At the end of 2008 I made a flash game in a two months that I put out there and got some reasonably positive feedback about (it was kind of an Asteroids game, but with upgrades and had levels to explore). I then decided to build the game "for real", because I felt that the project was already falling over itself in Flash (this was in Actionscript 2 back then, which wasn't object orientated). I hit a massive programming wall as soon as I exited the familiar confines of Flash, and spent months trying to scale it, and eventually decided I'd have to do something simpler first.
So in 2009 or so I started working on a casual game, a much more sensible venture. I had been doing so full time for about 6 months (as well as learning C++ at the same time, so progress was slow) when a Norwegian company saw it and, in exchange for employment and a copy of Unity and an iMac, bought the game.
That was a pretty rad deal, but after purchasing the IP I got a designer from Norway telling me what to do, from changing the name, the visual style, tacking on a narrative about an Inuit girl, and, in my mind, simplifying the game by removing many of the better features.
It was a pretty negative experience, and I nearly got fired, but it did eventually launch, and I heard it made a little money, but I don't think it was profitable. https://itunes.apple.com/en/app/northern-lights-hd/id364912193?mt=8 In my mind it was a failure, it was an awful experience, but it did leave me with a lot of useful information and a better means to make more games (at that point Unity only worked on Macs).
I then spent 3 months working on a game called "Pirate Climber" with the same Norwegian company. They had a vision for the game that was a lot like Fez, before Fez, but with a pirate chef. I basically got fired from the project two months in, after I hadn't been making good progress, and the game, after another couple months, released and was a commercial failure. https://itunes.apple.com/us/app/pirate-climber-hd/id364603189?mt=8
I then worked for 6 months by myself on a project that never saw the light of day. It was about ragdoll Kaiju (like Godzilla and Mothra) battling over cities. In truth I regret not releasing it, if I'd tightened up the scope, I could have at least got it out there, and learnt a bit more from the experience (even though the game wouldn't have made me proud). It actually worked on a iPod 2, but I killed it instead.
Next up I worked on a game called "Dragoo" alongside Tasty Poison (which was a game studio that at one time had 13 people on staff, but has subsequently closed down). I came in to save a project after the initial programmer left, who had replaced me on Pirate Climber, and three months later the thing released. I didn't save it, and the thing was a little embarrassing, but it launched, to little fanfare and little sales. https://itunes.apple.com/en/app/dragooo/id371490135?mt=8 I suppose it's only fair to say that some players thought it was pretty okay, but it was a muddled game by a fairly inexperienced team, and it didn't make it's investment back.
Then (now in 2010) I worked for two months on a game called "KickyFighter", which was sort of making progress, I even paid for two months of an artist's work, but I saw a better opportunity in working with Tasty Poison on a game called "Pocket RPG" (which was an actual success, my first commercial success), and so canned KickyFighter.
After Pocket RPG I started working on a game called "Slice Fight", which was an on-rails slice-em-up for iPhone. I even hired a programmer and an artist and worked for 7 months on it myself. The project felt like it was floundering, struggling to draw excitement, and there wasn't an end in sight, the programmer who was helping me was completely demotivated and basically got fired (which was the first time I fired anyone and it felt shitty as hell) and we eventually canned the project in order to pursue Broforce.
This obviously doesn't mention all the prototypes that lasted from a day to two weeks to a month that didn't go anywhere, and all the levels and mods I made that barely anyone played. Like for instance I made six distinct platformer prototypes before prototyping Broforce. I think we all agree it's great to have a failure that lasts a weekend. It's not really a failure at all really (except that one time that zero people enjoyed our game about debating business strategy with pigeons). It's the times when that failure has been drawn out for months, or have cost other people's salaries that I couldn't afford where I've found it to properly suck.
I think, like anyone, my failures have come from inexperience and naivety. But I think they've been in many cases compounded on by a kind of stubborn ambition. I'm quite proud of the fact that I've canned projects aggressively, I see it as a sign of savviness that I haven't succumbed to sunken cost fallacies and haven't clung to many doomed projects long after they were clearly doomed. I've also been pretty lucky that I've been the master of my own destiny much of the time.
But I also think I could have gotten more experience if I'd been prepared to release some half-assed games, instead of insisting on levels of quality I simply didn't have the skills to achieve.
I think over time my skills have caught up to my expectations, or at least come closer, but I think I've also learnt to accept my shortcomings better. And that's been an important lesson for me. I think I'd still balk at being told to change a game by an outside designer when I thought that change was for the worse, but I'd hope I'd have some more grace about it, and be more prepared to co-operate under constraints. I think I'm a more pleasant person to work with now, and because I'm on every team I am a part of, that really helps me lead a happier life.
(If anyone who reads this thinks to themselves that, actually, this guy is pretty shit at accepting imperfections, I can only respond that I'm a lot better than I was).
VR
In 1999 I got into an argument in a bar about VR being silly and games already doing everything better. That turned into an internship at the CSIR doing VR, I was mostly there to play with the Silicon Graphics machines they had, but it was fun doing different sims for people. We'd have to really understand the stuff we were making sims about, so we'd spend weeks driving mining equipment or in the secret parts of the Sterkfontein caves digging up early hominids. It was pretty cool for a curious hobo like me, but the division died of politics and us interns we offered jobs doing HTML stuff for tiny salaries. I said no thanks and got retrenched instead.
In 2000 my old boss wanted to start a new company to do more 3D VR stuff with all his contacts, so I joined in on that AND started a CS degree at the same time because why not? I thought that us founding members would have some form of equity, which I tried to negotiate for, but found out years later that I had zilch. We did a lot of strange projects, many of them full of fighting with strange software tools, I feel a lot of these were failures but they're not things I own, so talking about them feels weird. By 2002 the tech support team for our main tool, EON VR, would email me when people had problems. It sucks getting your own support tickets mailed back to you to solve...
I left early in 2003 to focus on my studies - instantly going from scraping 50%s to 90%+ for everything. Working and studying full time is not a good idea. Since the VR days I've always tried to evaluate tools and not be married to one above all else, I know exactly how frustrating being stuck in the wrong environment can be (yeah, rich coming from me, right?). I also try to find out really really early if something I'm working on actually solves a problem for someone, we got stuck on too many dead-end projects that didn't earn us any real money and just made everyone hate each other. I also learned a TON about team dynamics and what you don't do if you have a team that's self-motivated.
Ovid
I spent most of 2003 writing a 3D engine, game and simulation environment at university with a team I'd pitched the idea to: We wanted to build something that could replace the inflexible engines that I'd been using to make VR things in. Ovid used the then new C# and .Net reflection to recompile itself on the fly: You could write objects to be used in the engine and when you saved your source it would serialise any object of that type, recompile and re-instantiate them. It also used delegates extensively to allow your objects to get method call backs from central update, timing, input and rendering managers and the editor for it all was built inside the engine itself (you could also add your own managers and extensions to the editor as well, using the same systems). Maybe this sounds like a familiar design these days? ;)
Ugh, look at how hideous this is... Why did I think this would convince anyone of anything?
Ovid worked really well and by 2004 we gave it to a bunch of people to build things in that I'd worked with over the years in VR, they all said they wanted it, so we worked on what it would do next. We ended up looking for funding to build a stable editor/visual coding environment (I wanted to build a visual coding system that could output logical representations of games/sims in different base languages with extensions to allow people to write their own exporters for say Unreal or Torque, etc) and to flesh out the default behaviors and libraries for the engine. We pitched to a lot of different people, but got furthest with a place called UpStarts, the scouting division of Shuttleworth's Here Be Dragons VC fund. Here's what we wrote about games in the "What trends have you spotted that support your business?" part of their opening questionnaire: (The X was a placeholder, I'm not sure what it was supposed to be in the end). We had loads of meetings with UpStarts and had to do all sorts of competition studies (Virtools, Quest3D, EON and Torque, Unreal wasn't a thing yet) and market breakdown research. They said no after about 6 months of this because they said we weren't aiming big enough, but they didn't think that the market could scale very well. They were totally right: We were business noobs and I cringe when I see the cost breakdowns and terrible time estimates we did (everyone that's worked with me is nodding right now), but they got the scale thing very wrong. I learned that I had literally no clue about business, and I'd only start feeling vaguely comfortable with it YEARS later after many more failed startup attempts.
Right after the whole Ovid thing fell over and I turned out not to be an entrepreneur, I posted on the NAG forums about maybe making some games together just for fun. I wanted to get out of a bit of a rut and enjoy myself without feeling like I had to do super serious stuff all the time that would affect my careers for ever and ever. That post ended up spawning Game.Dev...
Molecules
I did end up using Ovid myself to make things though. My then-girlfriend had a hard time visualising carbon ring flips and arrangements for her molecular bio courses, so I made some animations in 3DSmax to help. I figured that I could maybe do that in real time with some tricks from an article about Hitman's ragdoll system and so I built a thing I called Molecules. I kept giving it to people to see what they did with it. I also kept thinking that each time I did that they'd love it and I'd be finished and could start selling it. I was wrong every time. It took like 9 months to build part-time and ended up looking like this before people started offering me money for it:
It did allow you to do neat things with massive molecules tho, yay buckyballs!
Selling a thing is also really, really difficult. Once molecules was "done" I tried to shop it around at universities and to different departments. Chemistry depts all forced their students to buy R200 physical kits with balls and rubber joints in them, to build anything of any size at all you had to pool with 3 other students and each buy different kits. I thought I could destroy that market at R150 a pop for a CD that let you not only make carbon rings, but had them self-organise into low-energy stable configs in real time. I think I sold like 20 CDs to people who just liked the software. No departments picked it up, but UP did offer me a spot doing a masters in chemistry if I could make the electron cloud visualisations better...
I still kinda feel like Molecules would work as software, especially on tablets these days, but I know a lot more about distribution and getting sales now. It would have to have some sort of awesome story behind it and, frankly, the simulation it was doing was a filthy hack, so I didn't have the courage to stand up and shout about it. I don't think I'm any more confident now, I just listen to people who like things I make a bit more. Molecules turned out not to be the springboard to a stable income that I was hoping for, nor did it serve as a launching point for the engine it was built in.
(Also, the interface sucks. Why past me, why?)
As I was leaving high school in 2010, I decided I needed to move away from "baby languages" like GML (I know right?). So, I started learning C#. Over the course of the summer between high school and varsity I learnt a bunch of stuff and eventually started building my own engine.
"Hot damn @Stray_Train! You're in the big league now! Just wait until you start pushing out all those master piece games and they'll never have bugs in them because you can track them down yourself now :O" (0 games were made by me with engine that shall not be named)
Fast forward several months and I happened across @edg3 in the computer labs at varsity. He told me about a game developers group that met up monthly. Up until then I was operating as a lone wolf, this news made me so happy! FINALLY! PEOPLE WHO COULD UNDERSTAND MY TRUE GENIUS. Fast forward a month and we're at the meetup. I remember Sven was running it that evening (he's one of the experienced programmers that has since left to work overseas).
Towards the end of the program he asks "does anyone else have something to show off?"
My time to shine had arrived.
"I DO"
So I swerve on over to the front and set my stuff up and I start talking about all the stuff I did. I don't remember what I said but it must have been a load of ignorant BS because I can clearly remember the face of Sven being spectacularly unimpressed as I went on. After I was done I asked for feedback (which at the time I thought meant stroking my ego).
Yes friends, I got me some feedback. Over the course of what felt like forever he pointed several reasons why my engine was... not so great. To top it off he made a remark about the art I used in my demos:
"All that art, you stole it from the torque game engine didn't you?"
"Yeah, how did you know?"
"I worked on torque"
"....."
":<"
It was pretty demotivating at the time but in retrospect it taught me some really valuable life lessons:
1. Failure isn't always bad, in fact it's a necessary part of getting better at stuff.
2. The best feedback is usually the stuff you like to hear the least.
10/10 would fail again.
@damousey had very kindly offered to do the artwork for Worst Warriors and I'd given some incredibly vague direction for the poster art. The first draft I saw wasn't quite right, so I thought sending a bunch of googled images back might be helpful. It wasn't :( Don't get me wrong, @damousey did a fantastic job. I just did a terrible job communicating. I'm getting better by inches now, but that incident is seared into my memory about how not to handle working with artists.
I was told in these very forums that is how you learn the tricks of our trade and have accepted it to be the way of things and not to view it as failure, but a way to grow instead.
Here's an incident that happened as recently as this past weekend. I'll title it "Backups!".
So, I spent a large chunk of my private time prototyping the new Unity 5 networking stuff after a bout of sudden inspiration and free time.
Things were going along swimmingly until I got load-shed'ed Saturday night. 2 hours later when the darkness departed, the large c# file, that I have managed to spaghetti together, failed to open in Visual Studio or Notepad ++. It had one long line of <null> characters and I seemingly couldn't get the code back. Suffice to say the sea couldn't wash clean the Eskom executives after my tirade.
I cried real tears and had a seventh glass of red wine. (can't remember which came first)
Very long story short. I managed to reverse engineer the dll using ILSPY which was not ideal, but I managed to re-create 2-3 days of effort in a few hours again. A personal embarrassing and costly moment, luckily saved and I have hopefully learnt from this mistake.
Backup your code in a geographic separate location, either a repository or a cloud space to avoid heartbreak and a tale of fail to tell others. ;)
Pt. I - The kickstarter
http://gamasutra.com/blogs/PeterCardwellGardner/20150629/247050/When_Your_Best_Isnt_Good_Enough_A_Tale_of_Failure_Part_I.php
Pt II - The feels
http://gamasutra.com/blogs/PeterCardwellGardner/20150629/247124/When_Your_Best_Isnt_Good_Enough_A_Tale_of_Failure_Part_II.php
However, what I regret is all the time I spent trying to make 3D games in Game Maker. I learned that just because a tool has a feature, doesn't mean that the feature is particularly GOOD. I spent more time trying to get GM to push polygons in the name of wow-factor than I did making compelling games. Not a total waste in that I learned a lot of what I know about 3D from my efforts, but I didn't get an awful lot of games done. Also, they looked awful. Shoehorning is never a good look.
Extra hilarity in that initially GM was all there was for accessible 3D, but I then held off on migrating to Unity when it came out because of my stubborn insistence to bend GM to my will. Many valuable learnings were delayed because of this.
Moral of story: Choose your tools wisely, and you'll be more productive.
My most embarrassing moment in terms of games was my first ever GameMaker Game JAM, in Oct2013.
I just left my very stable high paying chemical engineering career in end June2013 and was playing around with GameMaker and telling everyone this is what I want to do... so of course I enter a JAM competition. I cannot do art, and I have no idea how hard it is to actually sit for 3days and program after being a manager for 2years just talking in meetings.
Of almost 200JAMMERs only 80 of us submitted our games, and I should have been disqualified because I accidentally submitted the stupid GMS installer instead of the zipped exe file, and then after all the excitement I came in dead last number 80! I made a stupid text game, MonsterCraft where you need click a button until you get the target DNA string from the starting ones you are given. LOL FAIL.
All my failures are up on gamejolt for the world to see, I leave it there as a reminder.
Now 'our' strategy is to just make a better game with each iteration (not to beat Blizzard). It seems to be a decent strategy. I'm of the opinion that my biggest success (in the game development world) was probably to find team members, whom can fill in my blanks, and pick me up when I fall down. So I can get a kick to fall forward again. And after I shake off the dust, to have a laugh and start on a new project with them.
Here are some screens in all their 2001 glory:
We did learn a lot during this time, I think the main lesson was how to work as a team. We turned this into a 3D graphics company that is still going 12 years later. On the programming side, I went from Basic to Pascal to C to Actionscript and then Unityscript, although I can use C# I have used Actionscript for a decade so US was easy to get into.
Chris worked his way from POVRAY to 3D studio to Max and onwards.
It would not have been possible for us to just start with C# and 3DMAX - its the wins, failures and passes that have aided us along the way - so a pile of fail maybe, but the good kind. Heck, Stasis is also as a result of this partnership.
-Nic
Like @Gazza_N, I would say an overarching fail was also spending too much time on prototypes. I don't regret any of the prototyping, but I do feel like I kept working on some prototypes too long and could have been better. Sometimes it was because I made a sunk cost fallacy and figured a little more time would somehow turn a bad prototype into a great one, and other times I'd say I continued working too long because I got to a point where I didn't actually enjoy it any more and would have enjoyed it more if I started working on a new prototype.
I do remember one old project where I was working with a friend. We started by planning out the story and project (which was obviously way too large) and then decided to start making a short teaser trailer to get people excited. Naturally, the only thing that came out of that project was the trailer :/
A potential failure we (Clockwork Acorn) just narrowly avoided, mostly due to luck, was doing fixed cost for a project as our first bit of contract work. Sometimes this is unavoidable, and doing fixed cost for a project is the only way to get paid work, but if you can avoid it you should definitely try. When we started looking, there was a potential project we could work on, but it never came to pass because of delays on the client's side and some other fortuitous events. Looking back, I am almost certain that we would have made a significant financial loss on that project if it had gone forward and not enjoyed it at all. In just a few months, our ability to accurately scope and plan a project has drastically improved while we have been pursuing paid work by the hour. At this stage I feel comfortable doing fixed cost work, but it's taken quite a bit of learning to get here and there are still significant risks involved. If you need work, and fixed cost is the only thing available, please try and get outside experienced people to look over your (hopefully) detailed plan, at least when you are starting out.
When the game released, I wrote strategies. I wrote some of the first 6-pool defenses, I played games against the dude that invented Reaver-popping (and prompted patch 1.03's drop-cooldown change) and wrote about that. My page drew the attention of someone from Prima Publishing and they approached me, a spotty 17-year-old, to write for the official Prima Starcraft Strategy Guide.
I dithered on the reply because I didn't think it was legit and I didn't think they'd be able to pay me. Also, I was afraid that if it wasn't fake, I'd be shown up for not knowing enough. Eventually someone else wrote the guide. I was in school and completely overwhelmed, but passing that up was pretty daft of me. My girlfriend owns an original Prima Starcraft Strategy Guide, how cool would that have been?
...
My school friends and I used to write detailed design docs for games that we wished people would make and then send those off to game companies. We spent ages on a design for a new Dune RTS game and sent that to Westwood. I have no doubt in my mind that it was a terrible game design... I sincerely hope that Westwood instantly incinerated it as soon as they saw it was an unsolicited game design - I only found out that this was the standard response to those docs years and years later. I have felt silly for doing that ever since.
My attempt at a summary: