Assuming the infection vector is a function of virus laden aerosol, then masks reduce the range of a droplet cloud by reducing velocity. If you drew circles on a floor plan with radius R, the infectious area is proportional to the square of the range.

Masks are not a substitute for good tactical movement through a human environment, but they may be useful against far, effective aerosol discharge. Against a point blank breather not so much.

The body armor metaphor is surprisingly useful. Masks don't make you proof against near effective "fire". But they may be quite useful at reducing the velocity of said fire so your statistical safe zone is bigger.

It's possible that wearing masks had no statistical effect on reducing infection because they gave mask wearers a false sense of security. When the steel helmet was first issued to Tommies in WW1 head wounds immediately rose because soldiers assumed helmets were bulletproof.

It took a while before Great War soldiers realized the helmet only stopped low velocity projectiles. It was a supplement, not a substitute for good tactical movement. Wearing a mask in a mall might help, if you still keep your distance. Wearing a mask on an empty beach is crazy.

Regulations mandating masks at all times destroys the public understanding of how it works. Masks become magic objects, like amulets or talismans. The public complies mechanically, mindlessly. This voodoo mentality can cause more harm then good.

At close enough ranges, as the good doctor's video illustrates, masks may make no difference at all. If we shoot each other at arm's length with .308 it makes no difference if some are wearing level 1 vests. So in tight crowds it's possible masks may be ineffective..