There is not enough information to conclude that there is or isn't a problem. There are too many less than ideal variable compromises. 1. Focus sensor area is larger than the target so it might just as easily decide that an edge at a different focal plane is what you intended. 2. Way too little light for meaningful AF tests. 3. multiple focal planes within the target focus point with a very narrow depth of field at f/1.4 4. The shots not shown could have easily been the ones closest to the actual preferred choice by the camera AF. 5. The additional images at each adjustment value, if much different shows that the ambiguity of the tests are producing a range of values that prevent any conclusions at any one adjustment that was selected as the the closest.
Try increasing light, use an external flash patterned assist light to aid the AF to be more consistent. Close down the aperture for a base line set ofimages. With deeper depth of field you can determine if the lens can resolve fine detail or there is an optics problem. Create a flat target with distinct edge contrast. Use enough light for a good shutter speed that eliminates tripod shake. Make sure the focal plane and target are parallel. Shoot a series and evaluate them all, and average them
Your 18-200 at its widest aperture has such a deeper DOF that even moderate AF misalignment would be in focus. The f/1.4 50mm wide open is being asked to perform at a much higher level of precision to be in focus. I do not know the distance from camera to target so can't tell you what DOF you should expect. As a test, try taking set with both lenses at the same aperture. The 50 will probably be better even if it runs into softness at 1.4. Good luck... Stan St Petersburg Russia