Agreed. There are situations where frequentist guarantees is what you want rather than optimal estimates of parameters. Frequentists consider the model paramet

Agreed. There are situations where frequentist guarantees is what you want rather than optimal estimates of parameters. Frequentists consider the model parameters fixed and allow the data to vary. Bayesian consider the data fixed and allow parameters to vary. For example many software engineering, drug testing, and large scale pipelines want frequentist guarantees, because your system will have varying input data and you want theoretical bounds on what inferences you can make.