On the thermal paste issue -- a matter of personal preference. This is not something I made controlled tests for. On other cooling matters, such as choice of "best" coolers, lapping IHS and cooler-base to bare copper, and the choice of thermal paste -- I have hour-long CoreTemp log-file samples of 8-second-interval sample core-specific temperature values loaded into Excel, analyzed with statistical functions and used to create Excel bar-graphs with different colors for each core.
If the IHS and cooler-base are not near-perfectly flat, spreading around the paste on areas that do not make contact wouldn't help much. That's why I recommend lapping unless a very straight metal ruler applied in various positions and angles across the IHS and cooler-base shows they are flat as can be . . . .
After verifying that situation for yourself or lapping the surfaces, I believe that covering both surfaces completely with a thin layer of paste will help at the margin.
You are better -- even with a perfectly flat IHS and heatsink base -- to lap off the nickel-plating on both (unless the heatsink base comes stock as bare copper.) The thermal resistance of nickel is significantly higher than that of copper, or conversely, the thermal conductivity of nickel is lower. Thus, removing the nickel should result in a measurable reduction in both idle and load temperatures.