Statistical context
Average FPS describes overall throughput; 1% Low and, with enough samples, 0.1% Low describe slower performance. Duration, scene and calculation conventions affect comparisons, so align conditions first.
Monitoring & reports
Game performance reports
Analyze complete sessions using average performance, low-frame behavior and valid sample duration. Reports support repeatable comparisons and attribution, not simply a contest for the highest FPS.
Mechanisms and technical considerations
Sessions and event markers
Stutter markers and segments connect perceived problems to a time context. They identify intervals for investigation, not proof that a coincident hardware reading caused the issue.
Controlled comparisons
Keep resolution, graphics settings, frame cap, route and shader preparation consistent. Changing several settings at once evaluates the group, not the contribution of one switch.
Validation method
- Record a sufficiently long baseline session together with scene and settings.
- Change one variable for a specific hypothesis and record a comparable session.
- Evaluate averages, lows, valid duration and marked intervals, repeating the test to confirm the trend.