Richard. This is a classic case of a failure by the editors of the service sheets to update the copy in the light of engine development. This was written when there were two similar but different engine types in production. The confusion continues with the description of the cam followers and rocker gear.
The A7 model mentioned here is the early Longstroke design. The main bearing is a deep groove ball race, and as the race settles against the crank under the tightening of the crank nut the only float will be that present in the bearing by its manufacturing tolerance, in a word, zilch. Originally published in 1948, looks like no one thought to check the copy was relevant in later years.
For your engine read the service sheet as for the A10, and it makes sense. The difficulty with the later design with a roller race is the need to remove the race from the crank to add or subtract the shims, tighten everything up and check again. The inner race and its cage are awkward to remove where bearing inner is a tight interference fit on the crank, many bearings have been wrecked taking off and putting on/off/on!!!! Aim for Musky's figures, but be prepared for a bit of frustration, the theory is easy, doing it ain't. If the bearing is good I'd be tempted to leave alone unless its clearly wrongly set and at least it ran OK before. Even the thickness of any jointing must be taken into account, so make sure there are no lumps of old jointing holding the cases apart.
This topic is well served, with numerous enlightened approaches to the problem. Pour yourself a relaxing drink and peruse the Forum. You'll nod off in next to no time.....
RD.. All short stroke engines are roller. It's the service sheet that's wrong. They went back to the ball race on the early A50/A65, then returned to a roller as the power (and no doubt failures) went up.
Swarfy.