Hello,
The default formula for computing atmospheric radiation is Swinbank (1963) because the first integration of the heat exchange with atmosphere module was introduced in V6P2 released in 2012. It was based on internal experience for years (and before 2014).
The other formulae were introduced later, perhaps compared for one specific example but I do not remember exactly if one formula was significantly better or not.
Hope this helps,
Chi-Tuan