Table 3. Best Agreement between Extrapolated GTO and PW Interaction Energies of the S22* Test Set for Different Fitting Points^a.

points	set	scheme	BSSE	MAE	max dev.	worth?
HF
DTQ	non-aug	Helgaker	CP	0.03	–0.10	≃
	aug	Helgaker	CP	0.03	–0.07	≃
TQ5	non-aug	Helgaker	CP	0.02	–0.07	≃
	aug	Helgaker	CP	0.02	0.05	≃
DTQ5	non-aug	Helgaker	CP	0.02	–0.05	≃
	aug	Helgaker	CP	0.02	0.04	≃
MP2
DT	non-aug	Helgaker	half-CP	0.16	–0.38	√
	aug	Helgaker	CP	0.07	0.19	√
TQ	non-aug	Martin4	CP	0.07	0.19	√
	aug	Martin4	CP	0.06	–0.15	√
Q5	non-aug	Martin4	half-CP	0.06	–0.20	≃
	aug	Martin4	CP	0.06	–0.18	≃
DTQ	non-aug	Peterson	CP	0.07	0.19	√
	aug	Helg. (Pet.)	CP	0.05	–0.15	√
TQ5	non-aug	Mart.4 (Pet.)	CP	0.06	–0.16	√
	aug	Mart.4 (Pet./Fel.)	CP	0.06	–0.16	≃
DTQ5	non-aug	Helg. (Pet.)	CP	0.07	0.18	≃
	aug	Helgaker	CP	0.05	–0.16	≃

^aMean absolute errors (MAE) and maximum deviations are in kcal/mol. Worth indicates that the extrapolated energies are statistically closer to the CBS PW values than the corresponding direct results obtained at the highest extrapolation point (cf. Table 2).