The other things to consider are tube-to-tube variations, and aging.
You dont want to exceed the maximum-rated current, because the wearout becomes exponentially faster.
And you dont want to be too low, because you will get cathode poisoning, though it's reversible.
You basically have 2 options for an optimum design
#1 - Use a high anode voltage. My first nixie clock is mains-operated (no transformer) and doubles the voltage to +340 VDC. I use a 75K anode resistor and get 2.4mA of current, which is between the datasheet limits of 1.5 to 3.0mA. This keeps the tube current safe if the tube and/or line-voltage varies by +/- 15 VRMS, and tube-aging down to 145 volts. The downside is that this wastes energy as heat, roughly an extra 1/2 watt per tube. However, with no transformer I believe that is a zero-sum tradeoff. I have 3 clocks with this design that have been running 24/7 since 2012, with zero failures across 18 tubes. These are Burroughs 5092's, which are abnormally durable.
#2 - Use a current regulator. As long as the voltage doesn't go too low, you will always have the same current. As above, the downside is wasted energy (heat), but that is probably acceptable if you are protecting a valuable tube. This is what I've been doing on all of my subsequent designs.
If you stick with a +170 VDC supply, be sure to check how much margin you have for variations in line-voltage and tube-voltage (aging). You're better off boosting it to around 200V.