Table 2.

AICc analysis. The most negative AICc value indicates the best model fit. Δ_AICc < 2 indicate that the current model has a high likelihood of being the best model under different circumstances, i.e., a new dataset.

Model (est. parms)	Fixed parms	AICc	Δi	wi
$\mathit{Const}.\mathit{Confidence}\mathit{mean}\left({\overline{X}}_1^i\right)$	k	−70.6447	24.3397	0
$\mathit{Const}.\mathit{Confidence}[\mathit{mean}\left({\overline{X}}_2^i,{\overline{X}}_3^i,{\overline{X}}_4^i\right)]$	k	−72.1602	22.8242	0
Within-week const. confidence	k1 k2	−71.7531	23.2313	0
Polynomial 1st deg. (α1, α2)	k	−70.5479	24.4365	0
Polynomial 2nd deg. (α1, β1, α2, β2)	k1 k2	−61.8877	33.0967	0
Sub-model (Δ)	α=1 ε=0 γ=1	−81.6244	13.3600	0.0005
Sub-model (Δ, ε)	α=1 γ=1	−91.5062	3.4782	0.0637
Sub-model (Δ, ε, α)	γ=1	−89.0275	5.9569	0.0185
Sub-model (Δ, ε, γ)	α=0	−77.3589	17.6255	0.0001
Sub-model (Δ, γ, α)	ε=0	−91.4285	3.5559	0.0613
Sub-model (Δ, ε, γ)	α=1	−94.9402	0.0442	0.3549
Sub-model (Δ, ε, γ, α)		−93.0529	1.9315	0.1381
Sub-model (Δ, ε, α)	γ=0	−94.9844	0	0.3629