How can we make a decision on parameter spacing?

[galaxyumi]

When creating a physics model grid, first thing to do is to decide the step size of your parameters in the beast_settings.txt. Depending on your data (e.g., depth and number of filters), you might get tighter constraints on the fitting parameters with a finer grid. But this might not always be the case, and the finer the physic model grid the slower SED fitting.

If the size of parameter uncertainties are comparable to your parameter step sizes, it is one obvious indication that you under-resolve. We need to think about what is a more objective way to determine the step size for each parameters and/or each data properties.

[karllark]

One idea would be to fit a small portion of the full catalog suitably picked to be representative of the full catalog (maybe 1%?) with physics grids of different resolutions. Use these fits to test the coarsest grid that supports the science you are doing. Then you can run the rest of the catalog with that grid, potentially saving lots of time.

Of course, if you only have a small number of sources, this step would not necessarily save you time. But it may be useful to quantify why the particular grid coarseness.