Methodology

Browser-Based Measurement

MetricRally runs its tests in your browser. Timing, scoring, and personal result storage happen locally on your device. This means the accuracy of your results depends on your device's hardware, your browser's rendering engine, and your input method.

High-Resolution Timing

For timing-sensitive tests such as Reaction Time and Aim Trainer, MetricRally uses performance.now() instead of Date.now() for short elapsed-time measurements. performance.now() provides high-resolution, monotonic timestamps designed for performance measurement.

The measured value still includes latency from the display, browser event processing, input device, and operating system. A reaction-time result therefore represents browser-observed elapsed time, not a direct measurement of nervous-system latency.

Device, Browser, and Input Latency

Several hardware and software factors affect measured results:

• ▸ Display latency: Monitor refresh rate affects when a stimulus becomes visible.
• ▸ Input latency: Mouse polling rate, wireless transport, keyboard hardware, and touchscreen processing can add delay.
• ▸ Browser overhead: Browsers can process rendering and events differently.
• ▸ System load: Background activity can introduce timing jitter.

For the most useful self-comparison, repeat tests using the same device, browser, input method, and similar conditions.

Mouse vs Touchscreen

Mouse and touch input have different latency and precision characteristics. Aim Trainer and CPS support touch input, but touch and mouse results should not be assumed to be directly equivalent. The Spacebar Clicker is designed for a physical keyboard.

Background Tab Handling

Browsers may throttle timers and rendering when a tab is hidden. MetricRally invalidates or pauses timing-sensitive in-flight states where appropriate, but the safest way to obtain consistent results is to keep the test visible and avoid switching apps during a timed run.

Personal Best Methodology

Personal results are stored locally in your browser. For tests where lower is better, such as Reaction Time and Aim Trainer, the lowest score is the personal best. For higher-is-better tests, the highest score is the personal best.

Typing, CPS, and Spacebar personal bests are duration-aware so a short burst is not treated as directly equivalent to a longer endurance run. Quick Benchmark sprint results are kept separate from standalone-test personal-best history.

Why Self-Improvement First

MetricRally emphasizes comparison against your own prior results. Cross-user comparison can be distorted by hardware, software, and input differences, so a stable personal baseline is usually more useful than an unsupported global percentile.