The question has been asked and answered already. Are you seriously saying that you don't understand how a directional antenna works? How about every satellite dish you have ever seen? Or every large, outdoor TV antenna ever made? Any DirectTV or Dish Network customer who has had their dish bumped a tenth of a degree and lost their signal can attest to this.
Anyone can detect the direction of a common Wi-Fi signal with a paper plate wrapped in tin-foil positioned behind the antenna as a focusing reflector.
Why is this so mysterious for you? I contend that it isn't and you are just trying to digress the discussion like you always do. I think grasping at straws has degraded into Number 9 on the "Tools of The Debate Impostor" list. Can anyone confirm?
Edited for preciseness, and added that "JulianMartin" is my new hero! Sorry, Tesla.
Jules, (may I call you Jules too?)
Kudos! You have - in short time and few posts - nailed FET to the wall in what I can only describe as a brilliant and elegant display of scientific expertise and debate prowess. This thread stands as a monument to the true dichotomy between pseudoscience and reality, in a field of false idols and rubbish heaps.
If any FE'er were to best you, it would assuredly convince me that the Earth is as flat as the top of Tom's head.
I am proud to know you, Sir.
Ha! Yes yes call me jules. Well glad you have enjoyed my posts so much!
Well well well Tom. Where do I start?
With a Yagi beam, it's top down emission pattern looks like this:
http://hfradio.org/ace-hf/Fig06A_10-m_OptiBeam_Yagi_3-stack_Antenna_at_28.4_MHz.jpgYou can clearly see the greatest level of gain is achieved at 0 degrees. That's proof of a directional antenna. You can have a receiving antenna in the same form that has a higher gain for signals in the target direction. Using a squelch control on a transceiver, it's easy to cut out all the crap and wait for the signal you want to be received.
Factor in the wavelength of the signals we are talking about. 80m for example. That's pretty big. Thus, you would need a substantially large object of the right material and at the right angle to deflect the entirely of a signal backwards. A mountain range for example might just be able to do it if the signal is on a single path. That's one reason why this phenomena isn't a continual occurrence and isn't a simple thing to recreate at any time. An example of this is that it's easier for me to get this effect by transmitting south from the UK to say, South Africa. In that path, there are few mountain ranges etc.
To an extent Tom, you are right in that there will be alterations every time the signal has a deflection. But given the amount of power that is used to transmit to get this effect, these deflections and slight alterations are seriously insignificant. Think of the signal as a tube, bouncing up and down. Every time the tube bounces, a splinter of the tube flies off in a slightly different direction, but in comparison with the whole tube, they are tiny, and would never attain the propagation to give the described effect. On top of that these "splinters" would have to have the same thing happen to them several times before it could come from behind oneself. If you really want to get nitty gritty, the alterations in angular propagation after a deflection could actually assist in preserving the signal. See this pretty picture that I drew:

As you can see, varying deflections could well help avoid obstacles.
You say there are a plethora of other paths - please could you illustrate these as I have been taking the time to do so for you.
Obviously I have only spoken about one plane of deflection alterations. In the other, yes, it is possible that it can change direction. But not a hope in hell of the entire signal taking 90 degree bends or suchlike which is what I guess you are aiming towards in your statements. This effect that I describe doesn't demand that the signal be received 180 degrees opposite to the angle of transmission to prove the concept, but the plausible variation which I would estimate to only be a maximum of + or - 5 to 8 degrees are completely explainable by obstacles in the path of the signal as you have so kindly suggested. More often than not, it is received directly behind oneself though.
Just for completeness, this image shows the other plane of signal propagation from a Yagi Beam antenna:
http://hfradio.org/ace-hf/Fig06B_10-m_OptiBeam_Yagi_3-stack_Antenna_at_28.4_MHz.jpgAs you can see the varying angles in vertical propagation give 2 clear paths for the signal to take, one of which will inevitably get further round the earth as it will (mainly by luck) hit less interfering surfaces, and thus have better propagation.