Interpretation of Value of

• Thus, if value of is many (say >3) standard deviations away from its mean, then the theoretical form is in doubt.

• The usual maximum likelihood method does not have this advantage: remember L was unnormalized and so its value had no significance for the goodness of fit.

• This is quite a difficult problem and an elementary discussion is given in MacDonald notes, page 14 onwards.

• The only way to get a feeling for is to use it: so I won't dwell here on the important techniques used to minimize to get good theoretical parameters and estimate errors in the same. We will come back to this later in the notes.

• Mathews and Walker solve the simplest problem, i.e., are linear in theoretical parameters.

• We do an example of this below.

• Bevington or similar book should be read before tackling a real problem.