GROMACS version: Any version
GROMACS modification: No
Hi everyone,
I’ve seen that by default gmx bar considers 5 blocks for the block average used to compute the DeltaG_i of a lambda step. What about raising it to 30 (or more), if I have a 20 ns production of a small solute in a solvent? I ask this so to be sure the DeltaG_i is likely more representative of the ““real”” mean of the population, but I assume the default in gromacs is 5 blocks for some other reasons I am missing. Is that only due to the problem of each block putably suffering correlation in its estimate of dH/dl value (i.e. correlation of the frames within each single block)?
Wish you a good continuation, thank you.
Bests,
Jacopo