The only thing I would add, as a precursor to Raven's points, is possibly:
0) Make sure you know what you are doing. Go by an A4 pad and write your ideas for the project in it. Plan everything - that way you can spot if you need to learn how to do something first or if you need to mock-up how something will work in a blank project.
If you're making an RTS like Warzone 2100 where you can swap the chasses, turrets and motive types around then you're going to want to know how to do that. Looking through the manual you'd spot handy commands like "GLUE OBJECT TO LIMB" and "LINK LIMB" and stuff like that so you might want to go off and practise glueing, linking and so on, to make sure these commands actually do what you want them to. Once you have some code, if you get stuck, you're more likely to get help on this forum. I've noticed that people respond in a more helpful manner if you can show them some code.
The other reason I emphasise planning is that changing your mind about how things work when you already have even a thousand lines of code is a total pain in the arse as you typically have to go back and rewrite large chunks.
Ending a sentence with a French word is so passé
