Quick Answer
FPS performance analysis interprets average, minimum, maximum, and percentile lows together to judge smoothness and headroom rather than relying on a single peak number.
Formula
Smoothness Risk ↑ when (Avg FPS − 1% Low FPS) is large
Introduction
This guide is part of the FPS Test knowledge base. Use the FPS test tool on the run page for live browser measurement, then work through the sections below to interpret frame delivery quality beyond a single average number.
Average FPS is a summary, not a verdict. This article explains how to read the full distribution of frame delivery, including lows that define stutter during combat, driving, or camera motion.
Core FPS Metrics
Average FPS is the mean over the test window. It is easy to compare but blind to spikes.
Minimum FPS is the worst single-frame rate during the session. One deep dip can cause a visible hitch.
1% low and 0.1% low approximate the slowest 1% and 0.1% of frames, better correlating with perceived stutter than absolute minimum.
Before analyzing spreads, confirm your counters measure the same pipeline stage. Our article on how FPS is measured explains why mismatched tools make percentile comparisons meaningless.
Think of average FPS as top-line revenue and 1% low as worst-case cash flow: both matter, but the second tells you whether you survive stress moments.
- Average FPS for headline throughput
- Minimum FPS for worst-case hitch detection
- Maximum FPS for uncapped peak (context only)
- 1% low FPS for stutter-sensitive analysis
- 0.1% low FPS for competitive smoothness
Headroom and Spread
Large gap between average and 1% low signals inconsistent pacing or thermal throttling.
Target 1% low near your refresh rate for competitive play; single-player may tolerate lower lows with adaptive sync.
Genre changes the weight you assign each metric. Competitive shooters punish low 1% lows harshly; slow-paced RPGs may tolerate wider spreads. See FPS test for gaming for genre-specific targets and scene selection.
Frame spread shrinking after a settings change is stronger evidence of improvement than average FPS alone rising a few points.
Frame Spread = Average FPS − 1% Low FPS
- Log all metrics from the same session file
- Compare spread before and after tuning
- Weight 1% low heavier for multiplayer titles
- Ignore max FPS unless testing uncapped latency
Analysis Workflow
Turn raw numbers into decisions. Analysis without a fixed test scene is guesswork dressed in spreadsheets.
-
Collect three runs
Same scene, same settings, note ambient temperature. Three runs reveal variance; one run is anecdote.
-
Median the averages
Discard outliers from background apps or notifications. Median resists one bad run skewing conclusions.
-
Inspect lows
If 1% low falls below half of average, investigate pacing or thermals before buying new hardware.
-
Map to refresh
Compare 1% low to monitor Hz for visible smoothness. A 1% low of 55 on a 144 Hz panel will feel uneven in motion-heavy scenes.
-
Document changes
One variable at a time: drivers, settings, or hardware. Multi-change tuning makes attribution impossible.
-
Split by scene type
Log travel, combat, and menu scenes separately in open worlds. Averages blended across unlike scenes hide problem areas.
Analysis Scenarios
Avg 95 FPS, 1% low 42 FPS: playable average but noticeable stutter in firefights; reduce shadow quality or cap FPS before chasing higher averages.
Avg 72 FPS, 1% low 68 FPS on 75 Hz panel: excellent pacing; prioritize latency tuning next instead of more graphics reductions.
Avg 120 FPS, 1% low 118 FPS after driver rollback: spread collapsed versus broken driver, confirming software not hardware was the culprit.
Ray tracing on: average acceptable but 0.1% low crashes in rain scenes; analysis points to RT reflections, not general GPU weakness.
- Open-world travel vs arena combat splits
- Ray tracing on vs off percentile comparison
- DLSS quality mode impact on lows
FAQ
- Which metric matters most?
- For feel, 1% low and frame time variance often matter more than average FPS.
- What is a good 1% low?
- Near your target refresh rate with small spread from average. Competitive setups often want 1% low within 20% of average.
- Can browser tests report 1% low?
- Our tool reports min FPS and stability; native tools like PresentMon add percentile lows for in-game titles.
- Should I optimize for average or 1% low first?
- Fix 1% low and pacing first. Raising averages while lows stay poor often wastes settings headroom on numbers you will not feel.
Conclusion
Analyze averages with minimum and percentile lows. Spread between them reveals stutter risk average FPS hides.
Run FPS Test