Table 2.

Power for the approximate randomisation test from 5000 replicates of a 2‐arm trial of size T=150 using an FLGI with block size b=30 (J=5) under a case of scenario 1 with a treatment effect of 0.40

(1−β F)(s.e.)	(1−β RT)(s.e.)	p ∗(s.e.)	ΔENS	D
0.8086 (0.39)	0.6057 (0.48)	0.871 (0.09)	22.04	0
0.8972 (0.30)	0.6080 (0.49)	0.881 (0.04)	24.01	0.08
0.9524 (0.21)	0.6021 (0.50)	0.878 (0.05)	23.99	0.16
0.9802 (0.14)	0.5851 (0.48)	0.882 (0.03)	23.73	0.24

Abbreviation: FLGI, forward‐looking Gittins index rule.