| . | Fitness scores averaged over all generations . | |
|---|---|---|
| External fitness based on comparison with past winners . | External fitness based on comparison with random seeds . | |
| Layer 1 | 2.3 | 0.739 |
| Layer 2 | 3.2 | 0.814 |
| Layer 3 | 2.4 | 0.799 |
| Layer 4 | 11.7 | 0.892 |
| Layer 4 Shuffled | 10.8 | 0.871 |
| Layer 4 Mutualism | 12.0 | 0.897 |
| Correlation | 0.767 | |
| p-value of correlation | 4.15E−15 | |
| p-value < 0.05 | Yes | |
| . | Fitness scores averaged over all generations . | |
|---|---|---|
| External fitness based on comparison with past winners . | External fitness based on comparison with random seeds . | |
| Layer 1 | 2.3 | 0.739 |
| Layer 2 | 3.2 | 0.814 |
| Layer 3 | 2.4 | 0.799 |
| Layer 4 | 11.7 | 0.892 |
| Layer 4 Shuffled | 10.8 | 0.871 |
| Layer 4 Mutualism | 12.0 | 0.897 |
| Correlation | 0.767 | |
| p-value of correlation | 4.15E−15 | |
| p-value < 0.05 | Yes | |
Notes. The table shows the correlation between the two measures of external fitness. The correlation is calculated from 72 random seed fitness scores and 72 past winner fitness scores (6 layers with 12 fitness values for each layer yields 72 fitness scores). Each fitness score in this table is an average over all the generations for a given layer (an average of 12 fitness values, one for each run, each of which is an average over the 100 generations). We evaluate the statistical significance of the correlations using a two-tailed Student t-test for Pearson correlations. The correlation is high (0.767), and it is statistically significant.