Yesterday I experimented the same old 2-line kite of 3.75 m² and 4 m span, wind speed being 5 m/s. As the wind was constant during some minutes and not to high (for me), not too low, the experiments were more precise, although not measured. 4 wingspan loops were required for an optimal traction, more or less being equivalent of what I obtained by using large figure-eight paths.
Picture page - Kitemill (pictures 13 and 15): loops are used.
I think that loops provide more regular power because the wing is always at the same distant of the middle of the flight window. But I don’t think we would save much space by using loops. Indeed the spacing between the units is necessary even if the loops are tight. And by using figure-eight paths, the second loop of one unit could match the first loop of the next unit, even more with large paths.