Why wouldn't they bias it this way? I do not see anything wrong there. That's how simple regulated linear PSUs've been made last decades. The schematic is not stupid, but could be made certainly way better.
There is certainly missing a grid-stopper resistor. (gate series resistor would be likely required even when mosfet used instead of tube). Can oscillate, and probably will without it, due to the stupidly placed capacitor from G pin. (making the power tube/mosfet work as a grounded grid/common-gate RF amplifier/oscillator)
The NPN needs to be a high voltage type, but, more intelligent designer would use a cascode configuration with a high gain low voltage one, or stick in there even a TL431 to get very good results.
To get better voltage regulation, one could supply the reference (75V zener in this case) from the output of the supply itself. (stick a resistor from the zener cathode to the output of the PSU) - that way it self-regulates the zener current. The more current from the output to the zener compared to the current through the NPN, the better the result.
The circuit is lacking any over-current and short protection. Implementing fold-back current limiter may also be handy.