Thanks Matt. I do think that this is a different side of the same problem.
My theory at the moment is that there is some kind of data race involved when doing certain refactoring actions. 99% of the time, everything is fine and nothing goes wrong.. but on occasion, the data is being represented to NCrunch in an incomplete state.
It's extremely difficult to track and resolve an issue such as this, as it's likely to be occurring beneath the synchronisation layer (i.e. in VS itself), and the inconsistency stops it from being pinned down.
There may be changes to the way NCrunch synchronises with code in future.. for now, if you see this happen, I recommend just going in and touching the source file that seems to be out of sync with the rest of the codebase. This will trigger the event that will make NCrunch refresh its knowledge of the file, and is a much faster way to solve the problem than a full engine reset.