FPS Stability Test

What Is an FPS Stability Test?

An FPS stability test evaluates consistency across many frame intervals. A stable sequence delivers frames at nearly even spacing, whether its typical rate is 60 FPS, 120 FPS, or another cadence. This tool samples browser animation timestamps, reports average and median equivalent FPS, and then measures interval spread with standard deviation, coefficient of variation, median absolute deviation, percentiles, and a bounded stability score.

The emphasis is consistency rather than maximum speed. A perfectly even 60 FPS sequence should not receive a worse stability result merely because an even 144 FPS sequence is faster. That is why the score normalizes timing deviation against the median interval. Speed and stability remain separate questions: average FPS answers how fast the cadence is, while the stability score describes how evenly those opportunities arrive.

Why Stable FPS Matters

A high average can coexist with noticeable hitches. Consider a sequence that alternates several fast frames with a few very slow frames: its average may look acceptable, yet motion can feel uneven because delivery is irregular. Stable frame pacing makes motion easier to follow and input feedback more predictable. In browser interfaces, it can also reveal when animation work intermittently misses its expected scheduling window.

Stability should still be interpreted in context. One timing spike during a thirty-second run may come from an operating-system task, garbage collection, a notification, thermal management, or unrelated tab activity. A repeated wide band across controlled runs is more informative than one minimum value. The timeline helps separate isolated events from continuous variation.

How to Get Started

Keep the page on the intended display, set the desired Stress Level, and close workloads you do not want included in the environment. Start the FPS stability test and wait for its warm-up. Avoid switching tabs; a hidden tab stops the run because throttled background cadence would not represent normal foreground behavior.

1. Let a baseline run for roughly 30 seconds in the laptop or desktop’s normal test power mode, then select Stop Test.
2. Review median interval, jitter, variation, the timeline, and spike count before focusing on the score.
3. Repeat twice, change one setting or workload, and compare the shape of the results rather than only the average.

Understanding Variance, Jitter, and Consistency

Standard deviation measures the overall spread around mean frame time and reacts strongly to outliers. The coefficient of variation divides that deviation by mean interval, allowing cadence variability to be expressed as a percentage. Jitter on this page uses median absolute deviation from median frame time, a robust measure that is less dominated by one unusually long callback. P99 shows an interval near the slow edge of the distribution.

The stability score is 100 × exp(-8 × relative MAD), clamped from zero to one hundred. Relative MAD is median absolute deviation divided by the median interval. A narrow, even distribution approaches 100. The formula is a convenient summary invented for this site, not an industry certification or universal grade. The methodology page documents it so repeat runs and future implementations use the same definition.

FPS Stability vs Average FPS

Average FPS is derived from total mean frame time, so it describes throughput over the retained sample. Median FPS describes the typical interval. A gap between mean and median can indicate a skewed distribution with slower events. FPS range compares equivalent rates from the fastest and slowest intervals, but range is sensitive to a single extreme. Stability analysis combines robust and conventional metrics to keep that extreme in perspective.

Suppose two sessions average close to 120 FPS. The first remains around 8.3 milliseconds; the second frequently oscillates between 5 and 12 milliseconds. They have similar throughput, but the first has tighter pacing. The stability test should show lower jitter, variation, and interval spread for the first session even though average FPS barely changes.

Common Reasons for Unstable Browser FPS

Main-thread JavaScript, layout work, image decoding, compositing, background applications, extensions, video playback, and CPU or GPU contention can disrupt callback timing. Power saving may reduce available performance. A browser window split across displays can be scheduled in an unexpected way. Variable refresh behavior, remote desktops, virtualization, recording software, and thermal throttling can also change the interval pattern.

The test cannot identify which cause is responsible. Use it to establish that a pattern exists, then isolate variables. Browser developer tools can attribute long tasks on a page you control, while operating-system monitors can show system load. If the concern is a native game, collect the game’s own frame-time capture rather than applying a browser result to it.

When the Test Helps—and Its Limits

This FPS consistency test is useful when average motion looks smooth but occasional hitches remain, when comparing a browser change, or when checking whether a display-related setting produces steadier browser cadence. It also helps explain why 1% low and frame time matter. Use a longer run when irregularities are infrequent, but keep conditions stable so the result remains interpretable.

The page does not directly measure panel scanout, native-game presentation, hardware faults, or network latency. Browser timestamps have finite precision and are subject to scheduling policy. A stability score is only meaningful alongside its interval data and under comparable conditions. No threshold proves a component is good or bad, and one clean run cannot guarantee that every application will be stable.