**fit**command. The most visible is an additional column in the output during the fit. |beta|/N is the norm of the vector of derivatives of the statistic with respect to the parameters divided by the number of parameters. At the best fit this should be zero so provides another measure of how well the fit is converging. |beta|/N can also be used as the criterion to stop the fit instead of the statistic delta although this is still considered experimental. This feature is activated by an additional argument on the

**fit**command. For instance,

*fit 50 1.0e-3 20.0*will run the fit until either 50 iterations are completed, the statistic difference between successive iterations is less than 1.0e-3 or |beta|/N is less than 20.0. Setting either the statistic difference or |beta|/N stopping criteria to a negative number means they will not be used. These stopping criteria can also be set using the

**method**command.

Another internal change is the option to use delayed gratification. This is turned on by giving the argument

*delay*to the

**fit**or

**method leven**commands and turned off by giving the argument

*nodelay*. Delayed gratification is a modification to the choice of the Levenberg-Marquardt damping factor and has been shown to speed convergence in many cases. The default is

*nodelay*.

One other change included in the release was for the first iteration to only modify normalization parameters. We have subsequently determined that this introduces instabilities in some cases so have removed it in patch 12.8.0c. This patch also fixes an infinite loop case which has shown up in a few cases when running the error command.