Comparing 2-core and 3-core cable with two cores loaded, there are two thermal factors at work. The greater outside diameter of the 3-core increases the surface area in contact with air, which is usually the main barrier to heat dissipation, so the 3-core can dissipate more heat in total. On the other hand the internal construction determines how well heat can flow from the two loaded cores to the outside - I would expect this effect to be much less significant, although if it is a rope-wormed cable rather than extruded solid or sectored, the air pockets might have an effect. Overall, I would expect a 3-core cable to be as good or bettter than 2-core so I use the 2-core figures.
In contrast, a 3-core cable for a 3-phase circuit could have all three cores loaded so the dissipation could be 50% greater. Since the surface area cannot increase by more than the square root of the CSA, and in practice increases less, the CCC must be lower than for a 2-core. This is why I would not apply the 3-core 3-wire 3-phase tabulated rating to a single-phase circuit. The difference between a 3-phase 3-wire and 3-phase 4-wire is less significant. The difference between the cable O/D's is less, and although the heat dissipation is likely to be lower (without harmonics, sharing some current with the neutral decreases the sum of I²R values of the separate cores) there is the possibility of harmonics increasing the dissipation above the 3-core value for the same RMS total load.
This difference is why I would use different figures for 2- and 3-core cables used for a single-phase circuit, but the same figures for a 3- or 4-core cable used for a 3-phase circuit.
The unusual case mentioned above, of a 3-core cable used for L1, L2 and N would seem to fall into the 3-phase scenario.
