Hello, I am doing a Ray-tracer and I may be getting an visual artifact in a scene with a sphere projecting shadow on a plane. Here is:

As can be seen in the above screenshot, it is like the shadow is also projected in the bottom of the sphere in a reduced form. I checked all the calculations in a point located in this shadowed region to see if I could locate the source and my conclusions are this: The problem ain't during the first, direct bounce because the dot product between the light vector and normal is negative. This makes sense, since the light source is exactly above the sphere, aligned with it's center, so if the plane below the sphere wasn't reflective, all the bottom part of the sphere would be in the dark. Ok, then in the first reflection, the ray intersects the plane as expected, but the ray-sphere intersection algorithm says that a intersection happens, I checked the values one by one and everything seems fine. So where can be the error? After some thought I am starting to think that may not be one, that what this intersection means is that the ray hits the plane in a point where it is in the shadow due to the sphere, so none of it's red color hits the bottom of the sphere. But the image seems so unnatural that I am in doubt. What you guys think? Thanks for any suggestions.