Not sure I buy your argument...I don't understand how the side scoop flow is not just going to collide with the roof flow? What is stopping the side scoop air pressure from going out the roof and/or the roof air from going out the side scoops? Both will be paths of least resistance vs. flowing thru the IC itself which has a very high resistance to flow?
To make your side scoop design work efficiently, you'd need a three chamber (isolated) IC shroud (think exhaust headers). What you've done here is most likely got side scoop air flow with higher volume pushing air out the roof scoop (some will go thru the IC, but it might be worse than what the roof scoop alone could do).
I think your design would work MUCH better if you ran the side scoop ducting directly into the IC shroud with 3 isolated chambers in the shroud - left chamber for left side scoop, center chamber for roof scoop, and right chamber for right side scoop. Isolate the chambers on the IC surface using rubber/foam molding (you would need to figure out how to mold plastic for each internal shroud chamber). This way, the only path for the air to flow in all 3 cases is thru the IC regardless in the differences of pressures between all 3 paths of air flow.
Not trying to knock your design because heat soak is a problem and the more air one can flow over the IC the better. I think you're on the right track, just needs some refinement.
the flow from the roof scoop will also be under pressure from airflow being forced in at speed.
Total effective outlet area of the IC (taking out all the area blocked by the bars and fins, etc.) is still going to be greater than the area of the roof scoop.
This could further be assisted by some fans pulling through the IC