Change Details

First reported in the original trac ticket: https://laura.hepforge.org/trac/ticket/64 Generating and fitting toy experiments where the PDF contains Gaussian constraints is a slightly tricky business. This is discussed in ​[arXiv:1210.7141](https://arxiv.org/abs/1210.7141) and we should try and follow the guidelines therein. It would seem that the missing piece (wrt what we currently do) is to generate a different value for the constrained parameter (from the Gaussian constraint PDF) for each toy experiment. It would then be necessary to store a separate ntuple with rows for each generated experiment and columns for each parameter of the model (or perhaps only those that vary from experiment to experiment). It would make sense for this ntuple to be stored in the same file as the ntuple containing the generated data. This ntuple of parameter values would then need to be read by the fitting process before saving the results for each experiment, in particular, the `genValue` of each parameter would have to be set to the value stored, perhaps even before the fit is performed but definitely before the pull is calculated and the various results saved to the fit result ntuple.

First reported in the original trac ticket: https://laura.hepforge.org/trac/ticket/64 Generating and fitting toy experiments where the PDF contains Gaussian constraints is a slightly tricky business. This is discussed in ​[arXiv:1210.7141](https://arxiv.org/abs/1210.7141) and we should try and follow the guidelines therein. It would seem that the missing piece (wrt what we currently do) is to generate a different value for the mean of the constraint (or means in the case of multi-dimensional constraints) from the constraint PDF for each toy experiment being fitted. We need to do this only when fitting toy, so need a mechanism to distinguish that case from a fit to data.

First reported in the original trac ticket: https://laura.hepforge.org/trac/ticket/64 Generating and fitting toy experiments where the PDF contains Gaussian constraints is a slightly tricky business. This is discussed in ​[arXiv:1210.7141](https://arxiv.org/abs/1210.7141) and we should try and follow the guidelines therein. It would seem that the missing piece (wrt what we currently do) is to generate a different value for the constrained parameter (from the Gaussian constraint PDF) for each toy experimentmean of the constraint (or means in the case of multi-dimensional constraints) from the constraint PDF for each toy experiment being fitted. It would then be necessary to store a separate ntuple with rows for each generated experiment and columns for each parameter of the model (or perhaps only those that vary from experiment to experiment). It would make sense for this ntuple to be stored in the same file as the ntuple containing the generated data. This ntuple of parameter values would then need to be read by the fitting process before saving the results for each experiment, in particular, the `genValue` of each parameter would have to be set to the value storedWe need to do this only when fitting toy, perhaps even before the fit is performed but definitely before the pull is calculated and the various results saved to the fit result ntupleso need a mechanism to distinguish that case from a fit to data.