I don't have a D2X, but if I did, I'd probably do like Paul says.
This question about compressed vs uncompressed NEF was talked about quite a bit when the D70 came out (even though the D70 doesn't offer a choice). Basically, tonal values in the highlight regions are being thrown out (or more precisely, mapped out) in the lossy compression w/ the assumption that the losses there are least visible. If you don't need to correct for overexposure (or use the "expose to the right" trick, which essentially means needing to correct for overexposure), you're probably ok most of the time w/ compressed NEF.
Of course, I'm speaking mainly from what I know of the compressed NEF from the older cameras, not the D2X. But it's most likely the same, and one can probably verify it by just taking a D2X NEF and looking at the curve mapping (for tonal values) that's being used for the D2X. You can use Dave Coffin's dcraw for that -- which I did for the D70 when it first came out.
Also, do note that even if you need to correct for overexposure, it's not like there's no highlight details to be gained. It's just that there's less of it, or more precisely, you'll have more likelihood to see posterization/color aliasing/etc. there when correcting for overexposure.
_Man_