In my experience , testing with closed throttle just requires more compression cycles to peak the pressure. So as long as you cranked until you had two consecutive same/similar-reading pulses; then the test should be valid.
With dry cylinders this could be many many pulses, and the numbers are always lower than expected.
When oiling them up, the oil has to be given time to climb up all around the perimeter of the rings. I usually put too much oil in ,let it sit for a bit, and then crank the engine for about 20 seconds to blow out the excess, which makes a bit of a mess. The excess has to come out, else it will skew your numbers higher than what is true.Some oil will always stay in there in the crevice and on the top of the piston, so your results will always be slightly high. This is not a big deal if you know this. Furthermore, you are not really looking for an absolute number, but comparative between cylinders.
In your case, you found 2 bad cylinders so it doesn't make a difference; the zero pressure in the one hole can only be made in one of two ways, and both require the head to come off.
The 105 to 120 is a large spread indicating the engine has other problems. But you rarely see a smog-teen with more than 130/135 wet tested, so the best of your dry test is 10 to 15 psi low, about what I'd expect. But your low cylinders indicate probably bad valve sealing.
So even if your heads were all fixed up, you won't know about the other problems .... because of only having done the dry-ring test.
But if you now do a wet test and all the numbers jump up a similar amount, say 10 psi. then your best holes are looking good. And the others, because they also jumped up 10 psi, then (except for 3/6) one could assume the rings are reasonably good. Leaving 3 and 6 as obviously head problems. So now, one could expect a fair engine after the heads are fixed.
The second witness to that would be a Leakdown test. No point in testing 3/6 but if the other 6 holes all read low leakage and similar numbers (after the oiled compression test) then I would just fix the heads and drive it.
In your case however, you cannot go straight to the LD test without having done an oiled test, or at least oiling the rings, because with dry rings or rings with dryspots , the LD test-results will be ambiguous. When you put 80 psi in there, it will just leak out the dryspots..
In your case, I would remove the rocker gear and do the LD test with the pistons at the bottom of the bores, where the least amount of wear is likely to have occurred. Also, with the pistons at the bottom, you are free to tap the valves and watch the numbers improve. And so, by the time you are done, maybe all the numbers (except 3/6 of course) may be similar, indicating possibly, that at least the short block is usable as-is.
As to the tapping of the valves;
sometimes carbon gets stuck between the valves and the seats, and they don'r close properly. Tapping smashes the carbon and the numbers improve. Same with rust on the seats or sometimes the valves stick in the guides and sometimes the guides are baggy and the valves don't center themselves. Rising numbers with tapping may mean there is something wrong with the heads. In any case if the numbers come up, that might indicate that the rings are ok.
As to the LD results;
The higher the test pressure, in my experience, the larger the percent leakage will be read. IMO this is a good thing, more accurately reflecting the running pressures which could be over 200psi at times. So I test at 100psi with the pistons at the bottom.
The industry standard seems to be 80psi. And so, with a test pressure of 80, I like less than 4%, less than 2% even on a fresh engine. Some will say up to 8% on a used engine will be ok. And on a hi-mileage DD that might be true. But an engine with 8%(or more) LD, will not win any awards for efficiency, nor for power; but it will locomote the vehicle it is in..
Happy HotRodding
