I machined a couple of alloy spacers to pull the bearings in with a long bolt.
The hub was warmed in the oven.
The threaded collar side pulled in until seated followed by the offside.
I measured around the same as the OldBritts website at 0.056" difference at the spacer top and bearing shoulder in the hub but elected to take 0.055'" off the spacer tube.
When the double row bearing is pulled in with that alloy spacer it is contacting the inner and outer race of the bearing and will stop when it is in full contact with the hub spacer even if there was a 0.001" air gap under the outer race.
I personally think that is the safer option and of course the stock 4203 bearing comes no where near seated in the hub.
I double checked by fitting the front axle and a spacer, torqued to check the bearing rotated smoothly.
It is possible to fit wheel bearings (not taking into account possible hub expansion with braking heat

) and they go from smooth running to notchy when the axle torque loads the inner races and the outer is retained in the hub bore.
Perhaps there is some merit in running C3's for the extra 0.0005" clearance for that same reason.