Agile systems are inherently unstable where the outflow of work rarely matches the inflow. In a stable system Little's Law, (the relationship between throughput, cycle time and work in progress)
applies.
This presentation looks at systems where Little's Law does not apply and uses Monte Carlo techniques to map the relationship between arrival rates, departure rates, cycle time and total work in progress. Unlike Little's Law this technique is non-linear and stochastic in nature. However, it does not require the imposition of false stability through very high (40%-60%) of customer requests.
This presentation shows how Monte Carlo may be used
to tackle questions like "what is the correct size for a backlog?" and "should we add more team members / a new team" inexpensively to help Agile management.
It was presented as part of ALI2018