Question
Juncture studies what happens to queue simulations when time is quantised. The aim is to distinguish collision pressure from the loss metrics that users actually experience.
Analysis correction
The final aggregation unit is the mean of AF and DF together, not separate AF and DF summaries. The 38,880 observations are duplicated across scheduler architectures, so they are not treated as independent observations when interpreting the result.
Current interpretation
At fine resolutions, collision scaling is close to linear. Critical collisions have a stronger relationship with AF/DF loss gaps than total collision count, which makes severity more informative than raw collision volume for this experiment.
Boundary
These are simulation findings under the modelled workloads and scheduler conditions. They identify a useful analysis path and a corrected statistic; they do not claim to describe every production queue.