Table 1. Average binding free energy predictions, computational cost, and relative efficiencies of all methods.

Final average binding free energy predictions in kcal/mol computed from the five independent replicate calculations with 95% t-based confidence intervals. The computational cost is reported in millions of force and energy evaluations per replicate calculation. Relative efficiencies of a method X are reported with respect to OpenMM/HREX as e_{err,X/OpenMM/HREX} as defined by Eq. (5). The lower and upper bound of the 95% confidence intervals bootstrap estimates for the relative efficiencies are reported as subscript and superscript respectively.

	CB8-G3					OA-G3					OA-G6
Method	Δ G [kcal/mol]	neval [×106]	estd	e|bias|	eRMSE	Δ G [kcal/mol]	neval [×106]	estd	e|bias|	eRMSE	Δ G [kcal/mol]	neval [×106]	estd	e|bias|	eRMSE
AMBER/APR	−10.5 ± 0.6	2135	$-{0.6}_{-0.9}^{-0.4}$	$-{0.4}_{-0.8}^{0.0}$	$-{0.5}_{-0.6}^{-0.1}$	−6.3 ± 0.1	458	$-{0.1}_{-0.3}^{0.1}$	${0.7}_{0.5}^{0.9}$	${0.0}_{-0.2}^{0.2}$	−6.8 ± 0.1	305	${0.1}_{0.0}^{0.3}$	${0.35}_{0.28}^{0.47}$	${0.2}_{0.1}^{0.3}$
GROMACS/EE						−6.6 ± 0.1	210	${0.2}_{0.0}^{0.8}$	${0.5}_{0.2}^{0.7}$	${0.3}_{0.1}^{0.5}$	−7.0 ± 0.1	212	$-{0.1}_{-0.2}^{0.1}$	${0.32}_{0.27}^{0.39}$	${0.01}_{0.02}^{0.09}$
GROMACS/EE-fullequil						−6.6 ± 0.1	261	${0.05}_{0.04}^{0.52}$	${0.5}_{0.3}^{0.7}$	${0.2}_{-0.1}^{0.4}$	−7.0 ± 0.1	271	$-{0.2}_{-0.3}^{0.2}$	${0.0}_{-0.2}^{0.4}$	$-{0.1}_{-0.3}^{0.1}$
GROMACS/NS-DS/SB	−11.4 ± 0.4	1202	$-{0.1}_{-0.3}^{0.1}$	${0.5}_{0.2}^{0.8}$	${0.2}_{0.0}^{0.3}$	−6.4 ± 0.2	450	$-{0.1}_{-0.2}^{0.0}$	${0.1}_{-0.2}^{0.3}$	${0.06}_{-0.17}^{0.00}$	−7.1 ± 0.2	450	$-{0.1}_{0.3}^{0.2}$	$-{0.2}_{0.5}^{0.1}$	$-{0.1}_{-0.3}^{0.1}$
GROMACS/NS-DS/SB-long	−11.3 ± 0.2	2202	${0.1}_{-0.1}^{0.2}$	${0.5}_{0.4}^{0.8}$	${0.25}_{0.21}^{0.32}$
NAMD/BAR	−13.0 ± 1.0	657	$-{0.8}_{-0.9}^{-0.2}$	${0.2}_{-0.3}^{0.7}$	$-{0.18}_{-0.40}^{0.08}$	−6.8 ± 0.07	657	${0.2}_{0.0}^{0.5}$	${0.9}_{0.7}^{1.1}$	${0.4}_{0.1}^{0.5}$	−7.28 ± 0.08	657	${0.1}_{-0.4}^{0.3}$	${0.3}_{-0.3}^{0.6}$	${0.1}_{-0.2}^{0.3}$
OpenMM/REVO	−16.0 ± 1.0	1920	$-{1.1}_{-1.3}^{-0.9}$	$-{0.9}_{-1.3}^{-0.6}$	$-{1.0}_{-1.3}^{-0.7}$	−11.0 ± 2.0	1920	$-{1.3}_{-1.5}^{-1.1}$	$-{1.4}_{-1.8}^{-0.8}$	$-{1.4}_{-1.6}^{-1.0}$	−12.0 ± 1.0	1920	$-{1.3}_{-1.7}^{-1.1}$	$-{1.9}_{-2.1}^{-1.7}$	$-{1.50}_{-1.66}^{-1.42}$
OpenMM/SOMD	−14.0 ± 2.0	460	$-{0.8}_{-1.0}^{-0.2}$	$-{0.5}_{0.3}^{0.8}$	$-{0.3}_{-0.5}^{0.0}$	−5.7 ± 0.1	420	$-{0.2}_{-0.5}^{0.0}$	${0.2}_{-0.1}^{0.6}$	$-{0.1}_{-0.3}^{0.1}$	−7.0 ± 0.3	420	$-{0.3}_{-0.5}^{0.0}$	$-{0.1}_{-0.6}^{0.3}$	$-{0.3}_{-0.4}^{-0.1}$
OpenMM/HREX	−10.8 ± 0.2	3327	0.0	0.0	0.0	−6.71 ± 0.05	2789	0.0	0.0	0.0	−7.18 ± 0.06	2615	0.0	0.0	0.0