I'm paid by the DTI for my CEN meetings and I don't know why this mistake has propagated, but I have an idea.
LPG from the refinery is guaranteed to be a minimum of 90 Octane and is generally around 91 to 92 octane.
The Octane number of a fuel is determined in a "standard" engine, which is actually a special engine that can have it's timing altered until it causes the combustion to knock.
Pure octane (a linear hydrocarbon with 8 carbon atoms) has a reference octane number of 0, and the 100 point is represented by 2,2,4 tri-methylypentane (a version of octane where the 8 carbon atoms are arranged in a highly branched manner).
A fuel with an Octane number of 50 would "knock" half way between the 0 and 100 reference points, and this could be made from a 50:50 mixture of the linear and branched octanes.
Getting closer to LPG:
If you ran the engine with pure Propane, which has 3 carbon atoms in a linear chain, the octane number would be 79. Butane (4 carbon atoms) is 67, and Pentane (5 carbon atoms) is 59
The more carbon atoms in a linear chain - the lower the octane number.
The more branching - the higher the octane number.
So if instead of the c-c-c-c-c arrangement of Pentane (Octane number 59) you used a fuel made from 2,2 dimethylpropane - which looks like the letter X with a carbon atom at the end of each arm and one in the centre - this has an octane number of 93.
LPG is a mixture of the linear and branched hydrocarbons, and the average value is around 90.
However, someone may be using the American "Research Octane Number" which uses different calibration points - and gives a number about 10% higher. Then the "octane" number would be above 100, but that's misleading, and exactly the same as saying 68