Opinions on: line as a functional unit

[oshfdk]

9.  Be aware also that if you compute a large number of frequencies (like the frequencies of 20 glyphs), some of them will come out anomalous (above 3σ)  just by chance.  Some of these false anomalies may be detected by splitting the sample text in half and running the analysis on each half separately.  Those false discrepancies are unlikely to occur on both both halves.

Is it important how the split is made? For example, randomly assigning lines to A or B vs assigning the first half of the lines to A and the second to B.