Formal definitions of the terms used in reaction-time measurement. Each entry is a one-sentence formal definition followed by detail.
In calibrated large-scale testing, mean simple reaction time in healthy adults is about 231 milliseconds (about 213 ms after hardware-delay correction; N=1469, ages 18–65; Woods et al. 2015). As that study shows, reaction times vary substantially across labs and devices, partly because of equipment timing delays — which is exactly why comparisons across different setups are rough guides only. AXIOM does not use reaction time to race for speed: measured repeatedly under the same conditions, change relative to your own usual is a proxy indicator of attention and alertness.
→ Measure it: free reaction testReaction-time distributions have a right tail (occasional large delays), so the mean gets dragged by a few slow trials. The median is robust to this skew, which is why reaction-time research prefers medians and distribution parameters over means. Whenever a single "average reaction time" is quoted, that number is heavily influenced by the slowest trials.
It decomposes reaction-time data into three parameters — μ, σ, and τ. Intuitively: your typical speed (μ), the width of the wobble (σ), and occasional large delays (τ). The advantage is keeping the shape of the distribution that a single average would destroy.
The anchor for "I'm fast/slow today." AXIOM compares μ against your baseline (your usual range).
Even the same person wobbles from trial to trial. σ larger than usual reads as: responses are becoming scattered.
Attention research has linked τ to lapses of sustained attention. Looking only at the mean, changes in σ and τ are indistinguishable; the decomposition separates "slower overall" from "usual speed, but occasionally dropping out." This is why AXIOM insists on decomposition rather than averages.
One of the most robust cognitive tests, used in sleep research for 40 years, and known for being highly resistant to practice effects (so repeated measurement is distinguishable from "just getting used to it"). AXIOM's measurement stands in this lineage.
Comparing against a population average ("top 22%"-style displays) is interesting at first, but what matters in practice is how you compare against your own usual. The same 250 ms is "slow" for someone whose usual is 230 ms and "fast" for someone whose usual is 280 ms. AXIOM puts this within-person comparison at its core.
What makes tilt nasty is that it is hard to notice from the inside — metacognition (your ability to assess your own state) is said to degrade early. AXIOM does not diagnose tilt. It displays statistical changes in reaction time — a proxy indicator — compared against your own usual.
→ Related articles (English)Lightning (under 200 ms) / Falcon (200–239) / Cheetah (240–279) / Wolf (280–339) / Bear (340–399) / Tortoise (400+). They classify speed only — not ability, state, or worth. Browser timing carries 1–2 frames (~16–33 ms) of uncertainty, so bands are reference displays.
→ Find your band: free reaction testA browser cannot know the exact moment a stimulus actually appeared on screen, nor detect dropped frames or virtualized environments. This is why AXIOM's free test labels results as reference values — and why the desktop app exists.
References: Woods, D. L., Wyma, J. M., Yund, E. W., Herron, T. J., & Reed, B. (2015). Factors influencing the latency of simple reaction time. Frontiers in Human Neuroscience, 9, 131. / Dinges, D. F., & Powell, J. W. (1985). Behavior Research Methods, Instruments, & Computers, 17(6), 652-655.