To achieve natural longitudinal stability of the conventional aircraft there needs to be a download on the tail.
Sorry to put people through this again, but that statement is simply not true.
Natural longitudinal stability depends on the variation of pitching moment with angle of attack - if an increase in angle of attack results in a tendency to pitch down again, you're stable. If instead you pitch up more, you're unstable.
Working through the equations to see how the overall aircraft pitching moment GRADIENT is affected by the tailplane/horizontal stabilizer, you'll find that what matters is the lift-curve-slope of the tailplane, and the relationship between aircraft AoA and tail AoA. The actual amount of download on the tail does not affect that calculation.
Where this fallacy comes from, I believe, is the fact that moving the cg forward increases both the download required for TRIM and also makes the aircraft more stable.
But consider this: if I magically move the tail further AFT, the download for trim reduces (due to increased tail arm) but the aircraft actually gets MORE stable.