The result was that Vector's diagonal receiver catches more light than experimental circuit's horizontal one. This means stronger reflected light comes from the front, though nothing reflects light. I wonder what the emitted light process is like.
I completely forgot "diffuse reflection". Since maze wall is not mirror, light does not go like the picture below (sorry I'm not good at painting).
In fact, light goes like this, I think. Here, diffuse reflection occurs and reflected light goes in all direction.
Because of that light process, stronger reflected light comes from front than side. This explaines the experiment result. And parallel arrangement of IR emitter and receiver is the best, since receiver faces to the direction in that light comes.