I'm having the same problem. I've tried validating the 2018 module against the Annex F example in API 610, but to get the example's answer I have to use exactly the same numbers as the Annex F example. However, the API 610 axes and CII axes don't align, and assuming the X axes align for the Caesar II global axis and pump local axis, the global +Y translates to local +z and the local +y translates to global -Z. When the loads / dimensions are corrected accordingly the results don't align with the Annex F example for the F4 and F5 equations (where moments are resolved to the pump centreline). Is this a bug?