Consider also this quote from his site:

"I have heard some skeptics state that the engines force would have been dispersed mainly sideways, but if this is so, what actually held up the 2,300lbs of lunar lander when it was on its descent to the Lunar surface?"

This too appears to argue that the characteristics (or even the presence) of an underlying surface is a key factor in how much thrust a rocket develops. Now there is a sort of ground effect when one is very near the surface, but that's not the essential component of thrust. It's merely tacked onto the basic thrust developed through Newtonian conservation of momentum.

And plume dispersal does affect the amount of thrust developed, but not the dispersal that occurs when the plume encounters a surface some distance away. In convergent-divergent nozzle designs, the nozzle converts pressure to velocity. Ideally the path of all exhaust molecules at the exit plane will be collinear and parallel to the axis of thrust, and the pressure of the exhaust gas will be equal to the ambient outside pressure. But it's practically impossible to achieve the ideal nozzle design for a vacuum (exit plane pressure of zero), and there will always be some residual pressure at the exit plane that will cause the plume to disperse and "waste" some thrust nonpropulsively. And it's even more difficult to design the ideal nozzle for a variable-thrust engine.

But this doesn't appear to be what Mr. Cosnette is talking about. A common, though inaccurate, abstraction of the principle of propulsion is that the LM "balances" on a "column" of exhaust. Balance is indeed important; the thrust vector must point through the center of mass. And there is indeed a column of exhaust -- the more columnar the better. But this is not the same physical arrangement as a static load balanced on an actual solid column. The density of the exhaust plume is, in fact, almost entirely irrelevant to the amount of thrust that engine produces.

Apparently Mr. Cosnette's common sense has failed him again.