How to finish a game.

I wonder how many articles are written with exactly this title. I've just noticed one. It's quite good actually, but here's a few alternative ideas. Do you know how many games I'm making at the moment? Three. A real time worms with grizzly bears, a realistic stealth shooter, a Zelda game, a Stick-man version of the 100 Agent Smiths fight (yes another one) -Hey wait, that's four games. Oh and my Stick-man point and click adventure. So five. Oh, and another shooting one.
Does this have a point? Kind of. The trouble with Klik is that it is so easy to start a game that you tend to end up starting something every time you feel inspired, which can be quite often. You see an action film, so you make an action game. Deus Ex has some real cool stuff, so I thing I'll make a futuristic RPG/shooter with some really cool gadgets. The result is that none of these ideas get passed level one.
But that's the crucial point: level one. Normally I make a game in this way: I start on the first level (naturally). I do the character, terrain, enemies and so on. Then I add something else, and keep adding until I'm happy (never) or take a temporary break to work on something new (often). Since I was never happy, I never got to the stage where I was ready to copy the frame and alter it to make level 2. It's not just me either; I've seen people who can barely manage five events give up on their hard work because they've had a new idea. There are a few solutions.
1: Have one level. No, really. How many levels are there in Sim City, or the Sims? One. Kind of. Or infinite, if you look at it another way. Grizzlies 3, my real-time worms (and most complete game) consists of three frames, an intro (where options are selected), a post game summary (where the score is presented) and the actual game, where every level is randomly generated (possible on TGF and even KnP) and all the options effect the game. Inspiration isn't a problem, as every time I felt inspired, I didn't start a new game, but simply added something new in, like the different skins (that's why there are 2 Neo skins and an Agent Smith… and a monkey).
2: Know when to stop. Limit yourself to something, (perhaps four or five hundred events, for an experienced kliker, maybe less for people with social lives) and don't go passed it. Keep a game simple and you'll get a lot further. Obvious, but oh so true.
3: Really be inspired. Really inspired, not just some spur of the moment thing. If you are making your fifth game of the same very specific type then perhaps it's an idea that you really like. Decide to do it perfectly this time, the definitive version. Then, know when to stop, and play about with the idea a little further on level 2. My stealth game is now on the sixth unique level, each with different objectives, some different terrain, special animated briefings and five different potential outcomes, complete with experience points to spend on abilities for the next mission. It's my third game along almost identical lines, but it's better in every way, and it's the only one to get passed level 1.
4: Don't plan. No, really. I planned out the whole of my Zelda game and guess how much you can do? You can roll, pick up rocks, bottle bugs and use bombs. There aren't even any enemies. Make something first, and don't be afraid to start again if it doesn't seem to be working out. It's fine to start something new if you've barely done anything.
5:Try making a game where the character and enemies aren't important. Give the guy one weapon and don't even bother programming enemies, and make Half-life on TGF. No-one will care about limited weapons when their being guided through the most atmospheric interactive levels they've ever seen, where shadows zip through air vents, valves gush steam, and platforms collapse behind them. Where voices scream but it's too late…Bwa ha ha…Ahem.
6: Don't get bored. Can't help you here. Maybe you're just lucky.

Hope that helps some people. Frankly, there really is no easy option. Well maybe there is. If anyone knows it, tell the world. Please, we need to know.