The optimal observable tables have slightly ill-defined exp covmats, i.e. when the covmat is built from systematics, there are tiny negative eigenvalues. It would be nice to fix this and properly regularise them so that we are safe from numerical shenanigans.
The optimal observable tables have slightly ill-defined exp covmats, i.e. when the covmat is built from systematics, there are tiny negative eigenvalues. It would be nice to fix this and properly regularise them so that we are safe from numerical shenanigans.