Page 1 of 1

convergence issue

Posted: January 03 2018
by Nishanth Sasankan
Hi ,
I been running CosmoMC and R-1 was about 0.078 the next time it was updated R-1 was 0.059 the next time it was 0.045. Is it usual to see that the drops in R-1 slow down like this between updates. The reason I ask is because the first time it was about 15, then it became 7 and then it was 1.7.

Is it Possible that some chain is just taking its time while others have converged.

Thanks

Re: convergence issue

Posted: January 04 2018
by Antony Lewis
You mean the values in the .log file? Doesn't sound too bad, won't converge so dramatically once the chains are all exploring the same area.