mkay, so the bandwidth usage is shifted from block propagation to mempool syncing; it would be interesting to find out how many fewer bytes get communicated in no-DS attempt scenario.
Its more about spreading the workload. Instead of getting a whole block and validating it ASAP before passing it on, you create the block over time and just send the PoW proof at end
it sounds like competing between miners is slowly eroded with one protocol-change at a time; feels like not long from now WHC-nodes/miners will fully replace BCH-miners