Error Calculation from Expansion

• When we are at a minimum, and Equation (29) becomes

  

• or remembering that the likelihood

  

• Equation (31) is expected Gaussian form for L.

• According to Section 3, M is just inverse of error matrix, i.e.,

  

  where estimate from minimizing ,

  true parameters.

• In particular, ``error'' in is

  

  and the other elements in are the much forgotten correlation parameters.

• Note that , all other , produces a change of 1.