DIN does not have the entanglement hazard that a yoke knob presents. And DIN does not have a built in lever (the knob) that could sustain an impact and provide a bending moment to the assembly as does yoke. DIN has a captured O-ring to make the seal whereas the yoke interface is not fully captured (IMO) and can thus extrude if given a chance (asymmetric loading or damage of the sealing interface). DIN is also lighter and more compact and thus advantageous for travel (advantage eliminated by carrying an adapter).
Modern yoke assemblies have been engineered to deal with asymmetric loads by being substantially oversized compared to legacy yoke assemblies and are rated to the working pressures of scuba tanks. Yoke is more damage tolerant of the sealing faces and in assembly is somewhat self adjusting/centering. We have all seen divers (other divers?) whose yoke connection bubbled happily away not just for one dive but an entire trip with no further issue than the annoying stream of bubbles.
And then there is the adapters, most are not heavily built, they increase the leverage potential by pushing the first stage further from the sealing interface, they have two sealing surface, both DIN and yoke, and the knob is there to entangle and to bang into cave/wreck ceilings and become entangled in jump lines (breaking them) thus again, combining the worst of both and the good of neither.
Think about a cylinder in service at a Caribbean resort or dive shop. It gets filled often at an off site fill facility. It gets transported to the dock/boat/shop and then gets transported again into the boat or Jeep or truck or camel. Then all the reverse back to the fill facility, multiple opportunities to drop the cylinder or impact the valve. Almost all tanks I have seen in rental fleets have at least some damage to the valve post. And yet a yoke can often be installed and will seal. Whereas a DIN regulator may not thread in if the broach is deformed or corroded and even if it threads in the sealing surface is not symmetric and thus no longer a fully captured O-ring with now the possibility of extrusion. Now hidden deep inside the threaded broach rather than being clearly visible that something is amiss as with a yoke. Installing a yoke on a damaged valve would not damage the yoke assembly, it just would not seal viably. Running a DIN assembly into a deformed or corroded or otherwise damaged valve could damage the regulator DIN assembly but fortunately usually is apparent such would be the case.