Performance and the printed report use the same budget. When the radio has adaptive modes, that selected mode is the only budget Spanline shows. A second single-threshold summary is not added beside it.
For a length in miles, Spanline uses Pathloss’s mile constant, 96.6:
96.6 + 20·log10(GHz) + 20·log10(miles).
The miles are the full Vincenty length. They are not rounded before the logarithm, and the kilometre constant 92.45 is not used on an imperial path.
Received level is transmit power, plus both antenna gains, minus radome, orientation, free-space loss, atmospheric loss, and the feeder, branching, filter, and miscellaneous losses counted once at each end. Thermal fade margin is that receive level minus the receiver threshold.
Atmospheric loss follows ITU-R P.676. That model has not been retuned. On the College & Bethel check the atmospheric term is 0.088 dB.
When a radio has adaptive modes, the operating receive threshold of each mode is the step-down level plus the ATPC step stored in the radio file. If the file does not give a step, the step is 0.5 dB. The 1E-6 threshold belongs to the lowest mode and stays in its own column. A .pl5 import has already applied its own 0.5 dB step, so Spanline does not add a second one.
Multipath outage uses Vigants. Rain attenuation uses Crane and ITU-R P.838. The availability printed on the path worksheet shows five decimal places.
On the College & Bethel file (Byhalia) the expected figures are free-space loss 127.83 dB, net path loss 58.32 dB, thermal fade margin 17.28 dB, and receive level −31.32 dBm. A change that moves those figures is wrong.