Table 1. Descriptive statistics of the data.

Note that the number of perfect animats in both tasks are substantially larger than in Albantakis et al. [47]. This is likely due to optimization of the MABE framework, used for the evolutionary simulation, between then and now.

	Easy task	Hard task
Total number of LODs	50	100
Number of LODs with perfect fitness on final generation	34	8
Number of LODs with mean Φ = 0 on final generation	10	6
Number of LODs with at least 1 trial of mean Φ = 0 on final generation	20	31
Percentage of trials with mean Φ = 0 on final generation (out of 128 total trials)	22%	8.3%
Quartiles of Φ values on final generation	min: 0 Q1: 0 Q2: 0.09 Q3: 0.69 max: 2.49	min: 0 Q1: 0 Q2: 0.18 Q3: 0.56 max: 4.11
Quartiles of surprisal values on final generation	min: 0.09 Q1: 0.09 Q2: 0.22 Q3: 0.54 max: 3.58	min: 0.09 Q1: 0.36 Q2: 0.69 Q3: 1.97 max: 3.58