@NSA, I don't think you've missed anything, certainly not the obvious. Co-channel interference and the multitude of other strong local signals is the obvious issue. WNED is predicted to have a NM of +32.2 via a line of sight path. That number means that on paper, there is over 1000 times more signal power available than needed to get a lock if not for the overpowering local signals.
I have to disagree rather strongly with any suggestion of using any amplifier or preamplifier. With over a dozen transmitters less than 3 miles from the OP's location, I'd expect the best high-input amplifiers to overload. The AP8700 will be hopelessly overloaded.
If I were going to go to any length to receive WNED,
I would try the 91XG first. Next I would try a horizontal stack of identical antennas, adjusted to null out reception of the signal from CITY-DT on real CH-44. If needed, I'd then look into a single channel filter tuned to CH-43. There is more than enough signal and antenna gain to drive any passive filter... I'd expect no more than 3 to 6 dB insertion loss through a good filter, but there's enough power there to drive a filter with over 20 DB insertion loss. (Sadly, the limitation of antenna length may force the OP to try shorter Yagi style UHF antennas if they are willing to attempt to stack.
http://forum.tvfool.com/showthread.php?t=1024 The Winegard HD9075 is about 5' long. The Antennas Direct 42XG and SR15 are also less than six feet in length.)
Rule of thumb: "Antenna gain is always better than amplifier gain." (Unless you make you living selling amplifiers.)