We've all seen it a million times. A new startup, an Internet organized group, a few friends saying they want to do it. It's not surprising that people want to make their own games, it's a natural want to express creativity in a new medium. If it's not making your own amateur movies with cheap effects, or running a garage rock band, it's writing your own game. Yet how many amateur game projects can you count, that have succeeded? I'm pretty sure the list is quite small compared to the projects attempted.

Before we can answer what happens, lets identify whats involved in
most amateur game projects.

1. A healthy amount of code. This will vary based on project, and how smart the creators are. If they're smart, it'll be based on an existing engine, or use a lot of free middle ware. The overly ambitious will embark on a genre defining journey of extrema graphics and the best game play. From my tone and common sense, you can guess why that doesn't work.

2. Large amounts of art assets. With no art assets, theres no game. Now many projects either underestimate the amount of art involved(every crate, every box, every tree), or assume it's as easy as rapidly coding.

3. Creativity and working under constraints. Most game projects are not about "What can we put in to make this game fun." but rather "What can we take out without hurting the core game experience." It is for this reason that most games, and most advice to new game developers, start out with "prototype, make it simple, make it fun, go from there."

Now what typically happens in most projects? A mix of hubris and plain inexperience. Most projects, quite predictably, are ambitious and wish to do all sorts of nice things, like redefine a genre, create the best game ever, etc. Not really that useful, nor really helpful for anyone. Also, most game projects are started out by programmers, or to be more precise, technologically oriented people. These people typically do not really understand the large amounts of time it takes to create art assets for games, never mind 3D games. Also, lack of planning is a key problem. A design goal of "Make this shooting thingy fun," isn't really helpful, not to a programmer, to a developer, or to anyone bar the person who wrote this down.

Furthermore, many programmer focused projects carry around with them a self inflated sense of the programmers ability. Typically, the project will start from scratch, and take many weeks to get to a state where it's playable and where other team members input is relevant(getting the engine up to scratch, etc.), this is a fatal mistake. The mantra of some of the most successful hobbyist projects, among them Red Orchestra and Counter Strike(both successful First Person Shooter mods which turned into retail products) has been to build on and improve existing games, and to use their code base. This significantly reduces early development time, and enables developers to reach the critical stage where game play is actually defined and tested.

The other side of the mountain is where most people(who are not programmers) end up. Full of cool ideas, and lack of ability to implement them. In many projects, there's a will, theres a way, but no one is around with a map, or with hiking shoes. It's not impossible to create worthwhile mods using only provided tools without writing a single line of code, but it's definitely easier if someone's around who can actually understand what all the jib-jab about for loops, function calls and objects is about.

As an additional factor that is usually not considered, is what the Internet provides. The Internet is both a blessing and a curse to all game makers. On one hand, it's an essentially boundary-less expanse of talent and interested people. On the other hand, it's an vast, uncivilized land full of people who will take any excuse to deride your project even if they can't do better. Many a project has failed because it forgot that it's possible to find artists or programmers on line, who might be willing to help, either for fun, or for a small sum of money. But there are also projects which have failed because they took in interest input too readily, ending with a project that is full of compromises which satisfies everybody.

If so, what is recommended for such projects to succeed? There is no ten step plan to gaming success, no pre-defined list of what makes games succeed, or grab awards. Gaming is littered with examples of this, from Sims, to Loco-Roco and DEFCON. However, there are some general hints that help.

1. Plan early, plan often. Not to be confused with setting things down in stone ahead of time, but before you start any journey of magnitude, it helps to have a clear set of goals, and requirements. This is especially true of any group effort, where, needless to say, every person has a different view of what the game will be about, how it will look, or play. A coherent effort will bring results far more readily then a dozen people each dragging the game to their own likes.

2. Re-use. There are three dozen or so open source game engines(OGRE, or such) out there, there are several commercial game engines which enable mod development openly(Source, Unreal Engine being the best examples there.). If you're writing a game from bottom up and all you plan is to create a variation of a known genre, you're better off using pre-made tools, it's unlikely a first attempt will be better.

3. Re-use. Yes, but re-use art assets. Does your box really need to look different then every other box out there? There are several open source texture repositories, more which cost a small sum of money. There's no issue using them. Some of the most wildly known games, among them Half Life 2, S.T.A.L.K.E.R and Doom 3 all use the same common texture repositories, changing them only slightly before using them.

4. Know your limits. Don't set out to create a 150-map First Person Shooter, don't write an RPG to equal Lord of the Rings. Common sense, but it's important not to let delusions of grandeur interfere with common sense and a need to get a product out of the door.

5. Release early, release often. While this maxim is more true of open source application development then game development, it really is relevant. The sooner you can get some sort of working executable(or mod package) out there, the sooner you can get wide interest, offers of help from random third-parties, or the such. Waiting for perfection before release will result in vapor-ware.