When I hold my hand over the keyboard of my MacBook, I sense a nice, cold breeeze.
Surely the machine sucks in air from the keyboard area, and dispels it through the vents in the back?
In that case, using the machine with the lid closed seems like a bad idea.
We don't need to guess. We can measure.
You are right. Air enters through the keyboard area and exits though the rear. If the CPU is working hard and the fan is running at high speed a piece of paper will not slide off the keyboard if the MacBook is oriented vertically. When cool with the fan running slowly, the paper slides off.
I measured steady-state temperatures using iStat with the CPU taxed to an average of a few percent and to an average of about 74% (this is a very high average - I ran two simultaneous simulations in AstroGrav to get it this high). It was difficult to get a good average reading because the temperature and fan speed would very slowly oscillate up and down. The MacBook does appear to run a little cooler with the lid open but not by a large margin. There is almost no difference when the CPU is not heavily loaded. These tests were done running on AC with a fully charged battery.
It is worth noting that even with the lid closed, there is still a good air inlet area. By design, the lid does not close tight - it maintains about an 0.025" gap all around. This doesn't sound like much, but the length of the gap is almost the entire perimeter of the unit and the air outlet area it needs to match isn't very big either. I should measure this.
To summarize, there appears to be no advantage in running the MacBook open unless the CPU will be highly taxed. Based on the ventillation scheme of the MacBook, it is probably worse to run the MacBook with papers laying on the keyboard than with the lid closed. This is an experiment for another day...
Rob