Do negative start times express relative order of arrivals in the 2020 packing trace?

Hello, I'd like to use the 2020 packing trace for a research project.

I care about the order of arrival of VMs.

VMs that were running before the start of the 14 days window for the trace have negative start times. However, not all such VMs have the same negative start time.

I'd assume that if starttime(VM1) > starttime(VM2) then VM1 arrived to the cluster after VM2, just like for VMs with positive start times. Is this assumption correct?

What makes me question it is that I see ~3.5k VMs with a start time of 20 years ago (as pointed out in #14 ) => so I don't know how meaningful the negative start times actually are.