Leaderboard

I generally find the best thing to do is just create your tables and start working. You'll probably make a mistake/forget something in your initial table design but that's fine. You catch it while you're writing the code and then update your database as needed. There's a point at which trying to plan ahead become counter-productive. You spend far too much time trying to anticipate everything you'll need/want and either make things overly complicate or still miss something anyway. So instead just put in the basic stuff you know for sure you'll need and then let the design evolve as you work. I generally don't even bother with indexes outside of the primary key / foreign keys until I have something mostly developed, tested and working. That way I have a better understanding of what my actual queries are rather than trying to guess what they will be early on. Pretty much every mistake can be fixed, some may just take longer or be more complicated than others. Most of your mistakes will probably be minor like "Oh, I should save this too" or "This should go in a different table" or "Maybe I should have an index here". You'd be hard pressed i think to make any mistake that I'd classify as "fatal". Bottom line, don't worry about mistakes and trying to get things right the first time. Spend more time actually working the problem so you know what you need rather than trying to think of everything first. Don't be afraid of re-factoring things later on, possibly multiple times as your understanding of things improves.