
A stability curve can look fine while its area is wrong by 57.3x
I was wiring a browser-side coursework checker and realized this one is easy to miss: Simpson's rule will integrate whatever step you give it.
If the heel step is 10 degrees and you use h = 10 instead of h = 0.1745 rad, every area comes out 57.3 times too large. The GZ curve still looks normal, so the mistake may only show up in the criteria table.
I put the hand-calculation chain and a small cross-checker on the MetaCAD site I'm building: ship stability coursework workflow.
It's for coursework and preliminary checks, not vessel operations. The page shows the substitutions instead of just returning pass/fail.
What other stability-assignment errors have you seen survive a quick visual check?