With my new antenna setup I was able to measure loss in the cable from inside the house to the base of the tower, about 15' of RG-6 and 450' of Trilogy MC2 1/2" hardline. This wasn't practical before since the cable ran to the top of the 50' guyed tower. It's pretty hard to get a power meter using 120 V to the top of a tower.
Now I'm using RG-6 from under the house to the TV and RG-11 from the base of the tower to the VHF/UHF diplexer on the mast at 65'.
I had calculated how much loss there should be if everything was perfect. Of course I expected more because it's never perfect. But I measured 4 dB more loss than I expected at 700 MHz. This seemed high.
I was using several pieces of RG-6 to get into the house. Under the house to a wall plate with a barrel connector, jumpered to the preamp power inserter, and two more short pieces of RG-6 jumpered so I could connect the spectrum analyzer to the antenna.
I measured most of these pieces and some of them were awful. One 3' piece of RG-6 had 1.3 dB of loss. A barrel connector had 0.4 dB of loss. Yet one 6' RG-6 section was perfect with 0.3 dB loss. The Tin Lee power inserter had no measurable loss.
I got rid of all the crummy jumpers. Now the hardline under the house has a single piece of RG-6 running from there to the power inserter and the good RG-6 jumper runs from the power inserter to the DVR.
I was able to reduce the loss by at least 2 dB. I had to splice the hard line in 2 places and put my own makeshift F connectors on the ends so I'm sure there is loss there but I can't buy real connectors for the hardline so that's the best I can do. There was no rhyme or reason as to what components had how much loss. The expensive compression F connectors were not lower loss than the cheap crimp types.
The bottom line is don't use any more cable jumpers, barrel connectors, or splitters than absolutely necessary as the can add much more loss than expected.