Table 6.

Simulation results for designs with no control arm in the LFC scenario p₁ = p₂ = p₃ = p₄ = 0.20 and p₅ = 0.40, for N=250. n̄ = mean per-arm sample size. Each η_m = Pr(N_E1 > N_Ek + m), the probability that the number of patients randomized to arm E₁ is at least m larger than the number randomized to arm E_k. Values in the row E₁ – E₄ are per-arm.

Method	Arm	Pr(Select)	Pr(Stop)	n̄(95% CI)	η10, η20, η30
AR(1, 0)	E1–E4	0	0.58	24 (10, 72)	0.27, 0.15, 0.08
	E5	0.78	0.02	141 (11, 199)	0.03, 0.02,0.02
			Total	236 (90, 250)
$\text{AR}(\frac{1}{2},0)$	E1–E4	0	0.74	29 (10, 74)	0.32, 0.21, 0.13
	E5	0.81	0.02	102 (14, 164)	0.03, 0.02,0.02
			Total	217 (80, 250)
$\text{AR}(\frac{n}{2N},0)$	E1–E4	0	0.71	27 (10, 69)	0.31, 0.19, 0.1
	E5	0.82	0.02	107 (13, 169)	0.02, 0.02,0.01
			Total	216 (70, 250)
AR(1, 0.1)	E1–E4	0	0.68	26 (10, 71)	0.30, 0.17, 0.08
	E5	0.80	0.02	123 (17, 184)	0.03, 0.02,0.02
			Total	228 (80, 250)
ER	E1–E4	0	0.79	36 (10, 89)	0.34, 0.26, 0.19
	E5	0.75	0.02	61 (13, 103)	0.07, 0.02,0.01
			Total	205 (70, 250)