| Lever | Measured effect | Source |
|---|---|---|
| A poor night's sleep | ~29 ms slower on RT/attention tasks, on average | Wüst et al. (2024) |
| Severe sleep loss (17–19 h awake) | Some tests up to 50% slower — alcohol-limit-equivalent | Williamson & Feyer (2000) |
| A couple of warm-up rounds | ~7–13 ms faster than a cold first attempt | Del Rossi et al. (2014) |
| Moderate caffeine (1–3 mg/kg) | ~10–20 ms faster on a simple RT task | Rogers et al. (2024) |
| Talking on the phone while responding | Twice the missed signals, slower on the rest | Strayer & Johnston (2001) |
You can sharpen it — within real limits
Reaction time has a hard floor: no amount of training gets a human being below roughly 100 ms on a genuine visual reaction, because that is how long it takes a signal to reach the retina, cross the visual cortex, and travel back out to a muscle. What you can change is everything above that floor — the gap between your slow, distracted, cold-start self and your best self.
That gap is real and it is measurable, but it is made of several small, separately-studied effects rather than one big dial. The rest of this page is those effects, each with the study behind it, so you can see exactly what is supported and what is closer to folklore.
The levers, in order of impact
Sleep and alertness. This is the single largest lever with real data behind it. A 2024 systematic review pooling 44 sleep-restriction studies found reaction time on sustained-attention tasks ran about 29 ms slower, on average, after just one night with a reduced (2–6 hour) sleep opportunity compared with a full night's rest — and a study of medical students measured a comparable slowdown, with median reaction time rising from about 320 ms to 338 ms after a single night of partial deprivation. Push sleep loss further and the damage compounds: after 17–19 hours awake, performance on some cognitive and motor tasks was as impaired as the legal driving alcohol limit in many countries, and after 24 hours it matched full legal intoxication. Testing rested is the cheapest, largest gain available to anyone.
Practise the exact task. Reaction training transfers narrowly. A large evaluation of commercial "brain-training" programs found no good evidence that gains on a trained task broadly improve unrelated cognition — the improvement mostly stays inside the game you practised. If you care about a specific skill (a sport, a game, a start), the closest available drill to that skill is a better use of your time than a generic app.
Warm up first. Your first cold attempt is reliably your slowest. Repeated-testing research using the ruler-drop method found scores improved by roughly 7 ms from the first attempt to the second, and by about 13 ms once several more sessions had passed, before performance plateaued — meaning your very first try is not a fair baseline for anyone.
Fitness and hydration, in a more complicated way than people assume. Regular physical activity is associated with better vigilance performance, but the evidence is more nuanced than "fitter is faster": a pooled analysis across several exercise studies found that raw cardiovascular fitness alone did not reliably predict quicker reactions once the type of sport was accounted for — externally-paced sports that demand fast reactions to a moving ball or opponent (football, basketball) showed the benefit, general fitness on its own did not. Hydration matters less here than you might expect, too: a meta-analysis of dehydration and cognition found reaction-time tasks specifically were among the least affected by mild dehydration, compared with attention and motor-coordination tasks — worth staying hydrated for your health, but don't expect it to move your reaction-time score much.
Kill distraction. Divided attention costs real time, and it is well studied in exactly this shape. In simulated-driving research, holding a phone conversation — even hands-free — roughly doubled the rate of missed signals and slowed the reactions that did register, while listening to the radio or an audiobook caused no such penalty. The lesson generalises: a single-task, notification-free environment is not a nicety, it changes your numbers.
Caffeine, in moderation. A moderate dose measurably shortens simple reaction time — in one dose-response study, esports players reacted roughly 11–18 ms faster on caffeine than on water alone, with no extra benefit from the higher of the two doses tested. Daily use blunts the effect over time as tolerance builds, so the gain is real but not permanent leverage.
Fifteen minutes, a few times a week
Reaction time responds to short, focused, frequent practice better than to occasional marathon sessions, and the practice-effect research above backs that shape: gains concentrate early and then flatten. Here is a compact routine you can run in about fifteen minutes.
| Step | What to do | Why |
|---|---|---|
| 01 · Warm up (2 min) | Two or three throwaway rounds on your target test. | Removes the cold-start penalty before you record anything. |
| 02 · Baseline (1 set) | One honest, focused run at full effort, logged on the stats page. | This is the number you track over time — resist cherry-picking a lucky attempt. |
| 03 · Volume (3–5 sets) | Several more sets with short rests between them. | Aim to tighten your spread between rounds, not just chase a single fast one. |
| 04 · Cross-train (1–2 tests) | Finish on a different test — audio or a different input. | Varied stimuli keep practice from becoming rote. |
Where people waste their effort
"Brain games make you faster." The most thorough evaluation of commercial brain-training programs to date found no convincing evidence that they produce broad, real-world improvements beyond the trained task itself. A generic app is unlikely to sharpen your driving or your aim.
"Supplements cut reaction time." Beyond their caffeine content, most marketed reaction-time supplements do not have controlled evidence behind them. Sleep and warm-up, both covered above, do more for a verified reason.
"More practice is always better." The same repeated-testing research that found a 7–13 ms warm-up gain also found performance plateaued after a handful of sessions — further grinding did not keep producing gains, and fatigue works against you past a certain point.
Turn the tests into drills
Because training is so task-specific, the best drill is the live test closest to the skill you care about. Rotate through a few, warm up first, and let the stats page track your medians over weeks.
| Test | What it trains |
|---|---|
| Classic Reaction Time Test | Baseline visual speed |
| Audio Reaction Time Test | The fastest sense — sound |
| F1 Start Lights Test | Launching under anticipation pressure |
| Spacebar Reaction Time Test | Keyboard input speed |
| Virtual Ruler Drop Test | The classic physical method, digitised |
Know when you've hit your floor
Improvement is fast at first and then flattens hard, which matches every study cited on this page: the jump from a cold first attempt to a warmed-up, rested, focused score is the largest gain available, and it happens quickly. Beyond that, you are chasing single-digit milliseconds against a nervous system that has a genuine physical limit. When your median stops falling and your spread is already tight, the better goal is maintaining that number with rest and occasional practice, not squeezing out gains the research does not support.
- 01Wüst, L. N., Capdevila, N. C., Lane, L. T., Reichert, C. F., & Lasauskaite, R. (2024). Impact of one night of sleep restriction on sleepiness and cognitive function: A systematic review and meta-analysis. Sleep Medicine Reviews, 75, 101940. Link
- 02Del Rossi, G., Malaguti, A., & Del Rossi, S. (2014). Practice effects associated with repeated assessment of a clinical test of reaction time. Journal of Athletic Training, 49(3), 356-359. Link
- 03Rogers, E. J., Trotter, M. G., Johnson, D., Desbrow, B., & King, N. (2024). Caffeine improves the shooting performance and reaction time of first-person shooter esports players: a dose-response study. Frontiers in Sports and Active Living, 6, 1437700. Link