[GroundUrMast]

1.4-2.0dB loss per 100' of RG6. Worst case 0.02 loss per foot = 1.5dB total.

Actually, it's more likely to be about 5.5 to 6.5 dB per 100' in the UHF band.

Example:
#12 has 12' of coax so 0.24dB + 10.5dB + 7dB = 17.74dB lost.
#8 has 24' of coax so 0.48dB + 10.5dB + 7dB + 17.74dB from the unit above it = 35.72dB
#4 has 36' of coax so 0.72db + 10.5dB + 7dB + 35.72dB + 35.72dB = 89.66dB

It's not clear what the last term in the equation for #8 is representing. The same is true for the last two terms in the equation for #4.

Let's start with a simple example...

Per your TVFR, KRIN, real CH-35 is predicted to arrive at your location with a predicted noise margin of +55.5 dB. If you select an antenna that has 13 dB gain at CH-35, your would have a net NM of 55.5 dB NM(air) + 13 dB(antenna gain) = 68.5 dB NM(antenna terminals).

If you place the antenna in an attic with significant penetration loss, let's say 20 dB, the net-NM would then be 68.5 dB - 20 dB = 48.5 dB NM(antenna in attic).

Then let's account for coax loss. Lets say that from the antenna to the farthest unit, you need to run 60' of coax... perhaps it needs to run part way around the building, rather than straight. Using 6.5 dB / 100' as the estimated loss for RG-6, you would budget 3.9 dB loss for the total coax run. (6.5 * 60/100).

So, at the end of the 60' run of coax in this example, your net NM would be 48.5 dB NM(antenna in attic) - 3.9 dB(coax loss) = 44.6 dB NM(end of longest run with no splitter).

A conservative estimate/budget for splitter loss is:

4 dB / 2-way
8 dB / 4-way
12 dB / 8-way
16 dB / 16-way

So, let's insert a 16-way splitter into the equation. With a 44.6 dB NM(end of longest run with no splitter) - 16 dB(splitter loss) you end up with a net noise margin of 28.6 dB at the farthest antenna outlet.

There is no need for an amplifier in this example, absent interference problems or a drastic error in the signal level prediction, you have enough signal to get the job done and still tolerate more than 20 dB of fading. (Which would imply a problem with multipath, a problem not solved by an amplifier.)