I'd go 100-90-80, overheads 80 through 80, everything else from there.
But panning doesn't do donkey cock until you've found everything its own 'place' in the mix. Pan, compress, and twiddle volumes all you want... you're not getting shit until things fit together. My preferred solution... mix mono, motherfucker. It's got an acronym *and* it helps you find what parts of the spectrum are overloaded, what parts are empty, and how you can mash bits from the overloaded parts of the spectrum into the empty parts so you can actually *hear* shit. Pretty nifty, I'd say. Also, keep in mind that chances are that a huge portion of your audience will be hearing stuff in poor quality, likely in mono or with a mushed up stereo spectrum (Youtube-chu, I choose you! Use your shitty codec attack!), so if things sound mushy and indistinct in mono they're going to stay that way for a lot of the audience *AND* be improved very little in the stereo mix. Compression and panning are like fiddling puzzle pieces around (yes, I've used this analogy before... no, that's not cheating... it works) - you can rearrange mash things together all you want, but you're much more likely to get somewhere by looking at the pieces and finding where they all fit together than by mashing random combinations all over the place and hoping for the best.
Okay... that should be broken into paragraphs, but right now I don't know where to break that up so just bear with me if it's too messy.
Jeff
