Axial tilt: this can make planets very different. We have a range in the solar system from 0 to 177 degrees. Planets with near 90 degree tilt have warmer equators than poles, and quite bizarre weather (Darren M. Williams has a simulation paper about terrestrial tilted planets too). I would guess tilt can just be calculated as (say) 1d6 x 1d6 x 1d6 degrees to get a nicely skewed distribution.
<snip>
I like using just formulas (one reason I loved 2300AD was that it was the only game with a cube root formula in it), but many gamers do not. Tables take a lot of space. Maybe just distribute the formulas but make a simple on-line generation system?
Some problems with multiplying 3 d6:
1. Max is 216. Max tilt is 180. 181 is the same as 179. 1d6 * 1d6 * 1d5 would handily keep it to 180, with an average of 36.75.
2. Either your roll or mine gives big gaps of values you can't get, interspersed with spikes of values. For example, my roll give 1/180 chance of 108 and of 125, 4/180 of 120, and *no* chance of anything in between. The worst gap is from 150 to 180!
3. There should actually be a clump near 0 and a smaller clump near 180, as tidal forces are likely to push other values towards either 0 or 180.
4. Tilt should be more random further out from the star and more likely near vertical close in, for the same tidal reasons.
I'm ok with complex formulae myself, and would suggest something like generating a number between 0 and 1, with at least 3 digits of precision. Then I'd raise that number to a power depending on how close to the star. Far out, a power of 1 or slightly more gives very random, while close in, a power like 3 or so squeezes most values down near 0 while allowing the rare high value. Once you've exponented it, multiply by 90, and on a rare random, subtract that from 180 for upside down ones.
I know, it's more complex than most gamers want to do. But as I said, I go for complex.
