I would say the best situation is where you exhaust the hot air as directly as possible. Good examples of this are video cards that exhuast to the outside as well as PSUs. The problem is that CPUs tend not to provide this option unless you do some custom duct work.
Barring that, I would say that it's a very good idea to have good exhaust removing heat from the top-rear section of the case. The PSU tends to be in this area, so it nicely fits this bill. Cases that put the PSU on the bottom tend to have exhaust fans at the top and/or rear.
This heat extraction requires a thermal convection to bring the hot air up and away from all passively-cooled components. This is because suction is not a very directional process. When you breathe in the air comes from all directions around your mouth/nose. Blowing, however, is very directional. This is why you can blow a candle out from far away but you can't suck hard enough to blow a candle out. For this reason, if you have known hotspots that don't have active cooling of their own, it's a good idea to have case fans blowing cool air over these. This works for HDD, RAM, and NB/SB/VRM components on the motherboard.
I don't understand the proposition that negative-pressure setups would avoid dead air pockets while positive-pressure setups would allow these. So long as your case had good open vents/screens at the top/rear section, plenty of intake air would flush the heat out of the case as well as an equivalent system of exhaust fans at the top/rear with passive vents/screens at the bottom/front. Inside the case, the only difference would be that the intake-fan system will cause higher local velocities at the intakes.
I prefer exhaust fans - negative pressure - mostly because the hottest components of the computer have been designed to be towards the rear/top of the case. This design makes sense because, for example, HDDs don't have to be cooled by air that's already warmed by the CPU. This is why the intakes aren't placed at the rear/bottom with exhausts at the top/front. I preffer exhaust fans because they are closer to the hotspots. An intake fan that is far away from the electronic hotspots would have to overcome all the leakage through the case to guarantee good flow out the vents in the hotter sections of the case.
Unfortunately, this means that you will be having dust come in through all the cracks in your case, but c'est la vie.
Also, in a negative-pressure arrangement, you can provide extra cooling to an area such as the rear of the CPU socket area on the motherboard by making a vent-hole in this location. Cool air will be drawn in through this hole and cool the motherboard. I don't know how effective that could be, but it's an idea.
Side note: you will never have, in a steady-state situation (so, other than the first fraction of a second of operation), more mass flow in that mass flow out. At the speeds we're talking about, air is also functionally incompressible, however the heat will cause it to expand somewhat. You will have a slightly higher volume flow rate out than in.
Also relevant is that fan's rated CFM doesn't mean that that fan will produce that much flow when installed. Restrictions and adverse pressure gradients reduce the total volume flow rate. Consider a simple situation: a tube with one high-CFM-rated fan at the exhaust with a low-CFM fan at the intake. The total mass rate through each fan is going to be the same (and in this case, also the volume flow rate since heating is negligible). The same happens in your computer case, though you will likely have many more fans involved as well as passive vents and leaks.