This is the reason sight come ups must be made after estimating a trajectory to get on target. The ballistics calculation is an estimate, not gospel no matter how well someone will try to convince us that their program is exact. It's not - analysis is risk reduction, and that is all. Testing confirms the analysis.
Run a few variations of temperature and altitude on the ballistics computer you're using to find the reason. It matters, even at 100 yards when you have a rifle capable of discriminating the difference.
The slope problem is one of the distance the bullet flies through gravity. Let's say the straight line distance to the target is 200 yards at an angle of 5 degrees out of horizontal (the plane of gravity, we're using a flat Earth simplification for this short distance).
The horizontal component of the distance to the target is 200cos5 = 0.996(200) yards. A negligible error at that slope.
But what about 15 degrees? The distance is 200cos15 = 0.966(200) = 193 yards; maybe time to start paying attention, but probably not critical in the field. A 15 degree slope is noticeable and most of us are terrible at estimating angles in the terrain.