Table 1.
Teams composed entirely of leads (L=M) are significantly more successful (MWU test on S) than teams of the same size with one lead (L=1), regardless of team size M. Teams above the 99th percentile in S were excluded to ensure the differences were not due to outliers.
| no. teams N | mean success S | ||||
|---|---|---|---|---|---|
| M | L=1 | L=M | L=1 | L=M | MWU p-value |
| 2 | 14 823 | 8894 | 18.9 | 42.5 | <10−213 |
| 3 | 6171 | 2261 | 14.5 | 58.3 | <10−210 |
| 4 | 3063 | 717 | 12.8 | 62.1 | <10−112 |
| 5 | 1489 | 289 | 12.1 | 94.5 | <10−55 |
| 6 | 740 | 124 | 12.3 | 85.0 | <10−36 |
| 7 | 350 | 46 | 9.8 | 120.5 | <10−15 |
| 8 | 179 | 19a | 7.5 | 224.1 | <10−8 |
| 9 | 125 | 9 | 22.2 | 316.8 | <0.008 |
| 10 | 66 | 6 | 17.8 | 163.5 | <0.005 |
aWhen M≥8, the number of teams with L=M is too small (N<20) for us to reasonably conclude the difference in S is significant, despite the small p-values.