Locate abnormal intervals
An average can conceal brief fluctuations. Locate slow-frame clusters and their duration to distinguish sustained limitations from intermittent stutter and define the next test.
Monitoring & reports
Deep game reports
Move from session summaries to time-series analysis of slow frames, load changes and event markers. Deep reports narrow an investigation without presenting correlation as a definitive diagnosis.
Mechanisms and technical considerations
Relate surrounding state
Read hardware trends and scene context before and after an event rather than isolating the peak. Transitions, loading and shader preparation may require separate interpretation; spikes need not share a cause.
Evidence versus causation
Several slow frames may belong to one stutter, so slow-frame counts are not independent-event counts. Validate clues with repeated scenes, isolated changes or diagnostics rather than declaring a component faulty.
Validation method
- Start from a marker or abnormal interval and retain the scene, settings and timing context.
- Compare frame times with available hardware trends to form a testable hypothesis.
- Use a narrowly scoped retest or rollback and check whether the abnormal pattern changes with that variable.