### Putting a hard prior on w0+wa

Posted:

**June 29 2017**Dear Anthony et al.,

I am running chains with the CPL parametrization w(z)=w0+wa*(1-a).

However, I am thinking of this parametrization in the context of quintessence (scalar field) dark energy models which mean w(z)>-1 for all z, which translates to w0+wa>-1. I want to impose this condition as a hard prior.

In
which is supposed to return logZero=10^30 (i.e. a very high chi^2) if w0+wa<-1

However, I noticed that my chains are still accepting points with w0+wa<-1 (not much smaller than -1, like -1.02 or so, but nonetheless the hard prior should cut these points out). Do you have any idea if and where my mistake is? The above is the only change I made in the source folder.

Thanks a lot! Cheers,

Sunny

I am running chains with the CPL parametrization w(z)=w0+wa*(1-a).

However, I am thinking of this parametrization in the context of quintessence (scalar field) dark energy models which mean w(z)>-1 for all z, which translates to w0+wa>-1. I want to impose this condition as a hard prior.

In

**CosmologyParametrizations.f90**, within the function**TP_NonBaseParameterPriors**, I wrote the following:Code: Select all

`if (CMB%wa + CMB%w < -1.0) return`

However, I noticed that my chains are still accepting points with w0+wa<-1 (not much smaller than -1, like -1.02 or so, but nonetheless the hard prior should cut these points out). Do you have any idea if and where my mistake is? The above is the only change I made in the source folder.

Thanks a lot! Cheers,

Sunny