Every FE'r forgets that at least three things have to match in every observation with the proposed hypothesis: location of the Sun and Moon (and the other celestial objects), brightness and apparent size.
You can find a mathematical formula for a transformation that matches "FE location" with observed location. The proposed solution is wrong, but illustrates how this can be done:

But you have to match the three parameters and it is clear from the picture that the path of the light is greatly increased as dusk gets near. This means apparent size and brightness will decrease dramatically as the FE Sun and Moon get closer to the horizon, and this is never seen in actual observations.
As the path of the light doubles, the apparent size will halve and the brightness will be reduced by 3/4. Have you ever seen that magnitude 4 stars (faint but clearly visible) get totally obscured as they approach the horizon? FE'rs would have you think that these magnitude 4 stars will drop below magnitude 6 and become invisible to the unaided eye when approaching the horizon.
Lets be clear: you do not need to adhere to one "theory" like bendy light or the "sun is spotlight", for example, to be in trouble with the brightness requirement. Every possible model where the Sun, Moon and stars hover above our heads will have the same problem with decreasing brightness.