confidence intervall changes during PLE

[cataschlott]

When i run the PLE analysis I am currently running into the following problem:
I execute the commands

arPLEInit
ple 

and it calculates the profile likelihoods and the confidence intervals for the paramters. However, as soon as it finishes the confidence interval changes and I am very confused why this happens.
I ran the ple analysis on different parameter settings and this did not happen. Do you have any explanation for it?

[clemenskreutz]

There is one point for parameter #34 that has smaller chi2. This is why the y-axes changes.
If I should check, why this parameter has such a small "chi2", I need (at least) the workspace saved with arSave.