Measure browser-observed latency for a visual prompt and summarize repeated trials.
Inputs and method
Start a trial, wait for the randomized prompt, then respond with the button or keyboard. Valid trials report best, arithmetic mean, median, and spread across all accepted trials. The retained set is limited to 100 completed trials; the five-row Recent Results section is only a view of the newest five. Early responses are discarded, and Reset All clears the complete set.
The randomized wait is scheduling time and is not included in a completed trial. A trial starts when the ready callback records its timestamp and ends when the page’s event handler records the response.
Interpreting the measurement
This measures browser-observed latency from the callback that starts the prompt until your click or key event reaches this page. Display rendering, the event loop, the operating system, and the input device can add delay. It does not isolate or measure neurological reaction time.
The browser can coarsen timer resolution, and scheduled callbacks can run later than requested. Callback execution does not establish when the changed prompt becomes visible. Do not use these values for clinical, neurological, readiness, injury-risk, or athletic-ranking decisions.
Method basis
The displayed statistics are publisher derivations from the locally retained trials. The W3C High Resolution Time specification supports use of a monotonic browser clock for elapsed-time measurement but does not guarantee sub-millisecond resolution. The WHATWG timer specification explains why callbacks can run later than requested and how a pending timeout is canceled. For related measured-performance tools, see the vertical jump calculator, the stride length calculator, and the one rep max calculator.
Keep a session comparison worksheet
Before resetting, write down the device and input method you used and the number of accepted trials. Keep the displayed best, average, and median together; choose one statistic in advance for comparisons rather than selecting whichever looks fastest afterward.
| Session record | What to retain |
|---|---|
| Setup | Device, browser, pointer or keyboard method, and any interruptions you observed |
| Sample | Number of accepted trials; retain the visible trial list before Reset clears it |
| Summary | Best, average, and median from the same set of trials |
| Next session | Repeat the same chosen method, or label a device/method change explicitly |
A session comparison describes performance on this browser task. It does not establish a clinical reaction-time assessment or separate human response time from every device and browser timing effect.
For an arithmetic illustration, accepted trials of 180, 190, and 500 ms have a mean of (180 + 190 + 500) ÷ 3 = 290 ms, a median of 190 ms, and a best time of 180 ms. These describe the same three-trial list in different ways. Record an observed interruption beside the trial record rather than silently comparing one session’s best with another session’s average.