The Gain/Loss experiment isolates the impact of interruptions on overall system performance, helping you understand which downtime sources most strongly affect throughput and efficiency. By running this experiment, you can see how removing or reducing specific interrupts changes model behavior, often revealing improvement opportunities that are not immediately obvious from day-to-day operation. The table below details the results viewable with a gain/loss experiment.
| Result | Description |
|---|---|
| Summary | Shows run data for each run, including efficiency and any interrupts dropped for that run. |
| Efficiency Loss | Charts the impact of each interrupt on efficiency. |
| Efficiency Gain/Loss | Charts the impact of each interrupt on efficiency, as well as the gain, or efficiency savings, that occur with that interrupt active. |
| Efficiency Gain/Loss Box and Whisker | Shows a box and whisker plot representing the range of runs across each interrupt condition. This shows how these changes might impact run variability. |
Of important note with gain/loss experiments is that the "gain" in questions is a real life net positive. If two interrupts share a reset condition, and one interrupt often prevents a much more disruptive interrupt from occurring, that firs interrupt improves system efficiency overall. A real work example of this might be something like preventative maintenance to avoid a machine breakdown.