Have a look at our new tests; https://github.com/JuliaDynamics/RecurrenceAnalysis.jl/blob/main/test/rmatrix_analytic.jl#L71-L95
There I have noted that GlobalRecurrenceRate does not give exactly the results expected analytically. Only approximately. Why isn't this exact? What can we do to improve it...?
E.g., a ration of 0.2 should give exactly 0.2NN recurrences, but it doesnt.