Read Filip's comment (about power) for the earlier post on weight, and you may be wondering how in the world can an RCX bot compete with an NXT bot?
From a power and weight perspective, one of the 'balancing factors' I can imagine (for now) is the number of extra sensors and more powerful motors that can be made/purchased for the RCX. If you are talking about an 'anything goes' competition, the NXT is at a disadvantage due to lack of additional sensors and motors. But we can't expect this to last too long...
Another 'balancing factor' could be the experience factor. There's a lot more trial-and-error testing with the RCX, and it could be argued that inexperience with the NXT system could help the RCX designer(s). If LEGO Education NXT releases its version in August 2006, the experience factor will last only as long as the time needed for the NXT developers to catch up - could be months... or a year or more.
If you examine the last set of challenges for the FLL, it may or may not be difficult to imagine performing the tasks with the NXT. Lack of 'hands-on' with the NXT is what keeps me from writing 'no problem' - who knows, some of the tasks for RCX may truly be difficult or impossible (?) for the NXT - and by that, I mean the basic retail kit and/or Education version.
I have no doubts the NXT is going to be a powerful robotic construction system... my question is simply whether we should even be comparing the NXT with the RCX? I still tend to think of the two as apples and oranges and, therefore, comparisons cannot be fair to either product.