If I use -noinit while mcmc sampling, do you think my samples will be more correlated? I'm worried because my CI seem unrealistically narrow after doing samples with -noinit. I'm going to try doing them without -noinit, but I wondered if anyone had some insight. thanks, Mollie