I didn’t exactly understand the take away from this video. In the first test he does, the river sand concrete is three times stronger than the desert sand concrete, but in the second test, the desert sand concrete is 10% stronger than the river sand concrete. What exactly is going on?
He explained it in the video. The amount of water was adjusted so the viscosity (well, workability) was identical between the samples in the second test. Less water in concrete has a huge positive influence on strength, and rough sand needs more water to be workable. Thus, for comparably workable samples, rounded sand concrete with less water was actually stronger.
It's normalising for different things.
If you normalise for the the same ratio of cement / water / gravel / sand then sharp-sand (crushed rocks) is stronger.
If you normalise for workability (i.e. the concrete flows and can be worked easily) you need to add additional water to sharp-sand, which means that the smooth-sand (e.g. river-rocks) end up stronger.