[gmx-users] error range of surface tension

[Mark Abraham]

Hongyan Xiao wrote:
> Hi all,
>  
> I simulated the water/air system and calculated the surface tension of 
> water by using g_energy. The following is the result:
> 
> Statistics over 500001 steps [ 1000.0001 thru 2000.0001 ps ], 1 data sets
> All averages are exact over 500001 steps
> Energy                      Average       RMSD     Fluct.      Drift  
> Tot-Drift
> -------------------------------------------------------------------------------
> #Surf*SurfTen               1417.01    1894.45    1893.74  -0.179542   
> -179.543
> Executing 'xmgrace -nxy surface-tension.xvg &'
>  
>  From the result, I can obtain the water surface tension is 
> 141.7/2=70.85 mN/m. But I want to know the error range of surface 
> tension I calculated. Can you help me? For example in one paper, the 
> simulated  water surface tension is 70.94±2.25 mN/m.  Can you tell me 
> how to obtain the error range (as ±2.25) in the gromacs? Thank you very 
> much!

What do you want that "error range" to measure? You've done a single 
numerical experiment with an exact result. You couldn't report the 
reproducibility or accuracy range from a single physical experiment either.

Mark