I created a simple snake game and all the assets for it in 4 hours. You could "eat" fruit to gain length and points and bad fruit would make you lose points and length. One item would kill you and if you hit your own tail or the wall you would lose. You sped up over time.

Basic concept, and I already had my own little 2d graphics lib set up (since re-written much better) but there you have it.

I also did a missile commander type game in about 8 hours, though I could do it much quicker now. I later expanded upon it and in all spent probably about 30 hours.

i just did a space invaders clone which took about 6 hours but the majority of that was creating a spritesheet and defining the co-ords of the sprites

since art isnt part of a coders job i'd personally allow myself to create the assets before the flight as if it was handed to you on the flight by an artist

I've done pong in 27 minutes. I've done tetris in 2 hours, 34 minutes. Most programming contests are 3 hours long, having you trying to solve 6 to 8 reasonably difficult mathematical problems. We've had four or five 3 Hour Game Design contests that proved to be quite successful with a number of users producing very impressive results (in a couple of cases, I actually ended up playing the games for hours after the contest had ended).

I think 3 hours is plenty of time to do a reasonable amount of coding. You just have to know exactly what you're going for. You need to have a single crystal of an idea. You can't expect to make a fully featured role playing game with magic and combat and items, etc, etc. You *can* expect to make platformers, puzzles, racing games, shooters, and even some simple hack-and-slash games with some minor RPG elements. Use the 3 hours to make a prototype of a game feature that you want to test, an idea of how a certain mechanic would work in a given concept, NOT to make a full game.

After a while of doing this, you're going to notice that a lot of your code is *exactly the same* in each game. This is an excellent opportunity to write your own simple game shell to handle all the boiler plate stuff (setting up your graphics, wrangling menus and game states, etc). You extract it directly out of *your* real game code, so it is fundamentally grown from how *you* work, and you know it inside-out from the beginning. It doesn't have extraneous "might need this" code. It might have a few design flaws here and there, but you'll be intimately aware of them, you'll know how to get around them, and you'll know how to avoid repeating them on your next iteration of your game shell. You'll find that your 3 hours goes a lot further.

And don't forget to let us know how it goes.

Thanks for all the feedback. I did not get to finish the game, battery died 1 hour 45 minutes, don't know why. I will finish the game sometime tomorrow. I accomplished quite a lot in 1 hr 45 min. I think the 3 hour contest was a good idea, I would volunteer my time to help out, and I am also willing to sponsor a price.