Table 4. Percentages of correct matchings of the data-generating model by the AIC-best LLM population size estimation models based on overlap of 3–5 lists for each of four per-list expected probabilities of encounter.

See text for explanation of the data-generating models.

N	Expected Pr(Encounter)1	Lists	Data-generating Model
			${\mathcal{M}}_b$	${\mathcal{M}}_{bh}$	${\mathcal{M}}_h$	${\mathcal{M}}_{th}$	${\mathcal{M}}_t$
1,000	0.025	3	55.7	19.6	14.5	0.0	52.8
		4	42.1	22.2	9.9	0.0	78.7
		5	51.8	24.8	13.4	0.0	90.2
	0.050	3	51.1	12.3	30.0	5.5	59.9
		4	36.9	21.5	40.0	1.7	82.1
		5	55.9	19.8	39.4	3.3	86.9
	0.100	3	64.1	10.6	53.9	3.7	69.2
		4	45.3	10.5	65.9	2.7	83.0
		5	56.2	9.7	79.4	4.2	87.4
	0.150	3	59.9	8.2	66.8	3.8	71.8
		4	69.1	10.8	82.9	3.3	78.8
		5	74.2	13.4	90.5	1.5	84.7
	0.200	3	64.0	7.1	75.6	8.1	73.7
		4	73.9	10.8	87.5	4.5	84.9
		5	77.6	10.2	91.8	6.0	83.9
20,000	0.025	3	51.1	15.8	53.5	3.7	76.7
		4	52.9	14.8	47.9	5.2	83.8
		5	50.5	12.9	65.9	5.3	83.5
	0.050	3	67.0	7.3	63.6	6.3	81.2
		4	64.6	12.0	78.0	5.0	82.5
		5	71.9	12.3	89.4	2.0	84.0
	0.100	3	75.5	4.0	88.5	5.0	79.7
		4	73.4	4.8	89.0	2.0	85.2
		5	75.8	1.5	91.0	2.6	85.0
	0.150	3	70.2	3.5	88.5	6.5	81.5
		4	76.5	1.2	93.2	1.9	84.2
		5	80.0	0.0	95.2	3.3	84.2
	0.200	3	73.8	2.5	87.0	3.0	82.5
		4	82.0	0.0	88.0	5.1	82.8
		5	83.5	0.0	91.8	4.8	83.0

¹For first encounters in ${\mathcal{M}}_b$, ${\mathcal{M}}_{bh}$ and ${\mathcal{M}}_{tbh}$.