Thanks for posting the data. This information is very handy to know.
Just a quick thought regarding the 4 CIM + 2 Mini CIM test outperforming the 6 CIM test:
When I look at your battery voltage plots on pages 8 (single robot) and 14 (pushing), I see that your battery voltage on the 4+2 starts out higher than the 6, and the 4+2 voltage does not droop as low as the 6 initially, when the motors are in a stall situation. This is happening despite the higher measured current draw of the 4+2 system than the 6 system. I do see that you listed battery voltage as a variable that changed with each test. Do you think it is possible that the battery could have been in a higher charge state (and thus able to supply more power to the system) in the 4+2 test than in the 6 test? Is this difference significant enough that it could cause the 4+2 test to outperform the 6 test?