When it comes to air cooling, it's often best to test ideas yourself because there are so many inter-dependent factors that have to work together. In my experience, it's really the only way to prove theories. Often, changing one parameter influences everything else in the system. In biology, it's called symbiosis.
Here's an example. There are basically three philosophies when it comes to air cooling and air volume moving through the case: 1) More CFM on intake (to generate positive pressure); 2) More CFM on exhaust (to generate negative pressure); and 3) Equal CFM on both exhaust and intake. Philosophy 3 is usually touted as ideal, but honestly, it's almost impossible to do this, in my opinion. So really you have to choose between generating more positive pressure (more intake, less exhaust) or generating more negative pressure (more exhaust, less intake).
My gaming rig is in a Silverstone TJ09 (just so we know what we're working with); it has mounting points for four fans: 1 definitely for intake down in front (the middle, actually, as it breathes through gills in the sides), an exhaust fan mounted in the back and high up, and two additional mounting points above the motherboard chamber. All fan mounts are for 120mm fans.
I spent weeks searching for the optimum setup for the fans. The only fan I didn't change is the bottom intake fan; I played with all possible combinations with the three other fans. I tried the rear fan as an intake and upper two fans as exhaust; rear fan as exhaust, both upper fans as intake; rear fan as exhaust, either one of the upper two fans as intake with the other as exhaust. I tried all possible combinations. In the end, my temperature data and some visual aerodynamic cues (with incense smoke and, at higher airflow and pressure levels, strips of plastic taped in strategic places), I found that the best temperatures for everything in my system (not just the CPU, but for the chipsets and graphics cards too) resulted from generating negative pressure.
I'm not suggesting that negative pressure is the best solution for you. But the lesson here is: Don't make any assumptions. Test everything you can. The best answers are the ones you find yourself with your own equipment.
It's time-consuming and detailed work, sure. But at least you can trust your own work.
Good luck! And hope this (really long post) helps!