Cobaya: Fitting BAO-only likelihood

[Graeme Addison]

Hi,

I'm trying to figure out how to perform a fit to some BAO measurements alone (so, measurements of DM/rdrag, DH/rdrag and similar). How do I set up the parameters correctly for this in Cobaya (calling CAMB for theory)? Something like the background parameterization in CosmoMC. For example, in LCDM I want to sample Omega_m and the product rdrag*h, instead of Omega_c*h^2, Omega_b*h^2, etc.. I'm sure there's a simple way to do this, although I tried searching both the Cobaya documentation and this forum and couldn't quickly find the answer.

Thanks for any help.

[Antony Lewis]

You can use derived parameters to sample Omega_m etc in the same way that As and logA are handled in the presents you can see in "cobaya-cosmo-generator" GUI. However, since rdrag is theory model dependent, that cannot be handled directly, so probably best to run in standard parameters and output those as derived parameters.

(as of today DESI is also available in Cobaya master [not the same as parameter table in papers due to rounding in the tables])

[Graeme Addison]

Great, thanks Antony!