How much of your reaction time can you actually change?
Let's start with the honest answer, because most articles on this topic overpromise.
Your reaction time has two parts. The first is hardware: how quickly your eyes register a stimulus, how fast that signal reaches your visual cortex, how efficiently your motor cortex fires, and how long the command takes to travel to your finger. This part is strongly influenced by genetics and is largely fixed. It's why some people are simply quicker than others no matter what either of them does.
The second part is software: your level of alertness, how prepared your attention is when the cue arrives, how much decision-making sits between the stimulus and your response, and how much latency is baked into your equipment. This part is very trainable — and it's where essentially all of your realistic improvement will come from.
Put plainly: a committed month of training typically takes an adult averaging around 270 ms down into the 230–250 ms range. Breaking 200 ms takes years of specific practice or an unusually good genetic starting point. Anyone promising you a guaranteed sub-200 ms transformation is selling something. If you want to see where you currently sit, take five minutes with our reaction time test and note the median before you read further.
Simple reaction time vs choice reaction time
This distinction explains most of the confusion about whether reaction time can be trained at all.
- Simple reaction time — one stimulus, one response. "Click the moment it turns green." A single pathway with no decision in it. This is what most online tests measure, including ours, and it averages around 250 ms in adults.
- Choice reaction time — several possible stimuli, and you must pick the right response for each. Because every additional option adds decision time, these are typically 100–200 ms slower. This is what gaming, driving and competitive sport actually run on.
Here's the key point: simple reaction time is the least trainable of the two. There's no decision to optimise, no strategy to learn — just a hard-wired sensory-motor loop. Choice reaction time has far more headroom, which is why the most effective training drills add decisions rather than just asking you to click faster.
This is also why elite gamers aren't impressive because their raw hardware is different. A sprinter's reaction to a starting gun and a fighter pilot's response to an instrument alert are both trained selective responsiveness, not faster nerves.
Seven levers, ranked by how much they actually move your score
Ordered roughly by effect size per unit of effort, with the biggest and laziest wins first.
1. Sleep — the largest effect, and the one nobody wants to hear
Nothing degrades reaction time faster or more reliably than sleep debt. A single night of restricted sleep commonly adds 20–50 ms, and chronic short sleep has a cumulative effect that no amount of practice will out-train. In practical terms, being badly rested can cost you more milliseconds than a month of drills gains you.
Reaction time is one of the first cognitive measures to degrade and one of the last to recover, which makes it a decent proxy for whether you're actually recovered. Aim for 7–8 hours with consistent bedtimes — the consistency matters nearly as much as the total.
2. Remove the latency from your equipment
Unglamorous, and it beats most training outright. Every layer between the cue appearing and your click registering adds delay. Typical costs:
- 60 Hz display vs 144 Hz: up to ~16 ms waiting for the next frame to be drawn
- Wireless vs wired mouse: typically 8–20 ms in USB polling
- Background tabs and screen recorders: unpredictable, occasionally much worse
- Trackpad vs mouse: a different mechanical path, generally slower
Stacked together that's routinely 20–40 ms. Someone who "improved" by 30 ms after switching to a wired mouse and closing twenty Chrome tabs didn't get faster — they removed measurement error. Always fix your setup before you start measuring progress, otherwise you'll credit yourself for changes that never happened in your nervous system.
3. Targeted drills with unpredictable timing
Here's the trap: if a stimulus appears on a predictable beat, your brain learns the rhythm rather than the response. You get faster at anticipating and no faster at reacting. This is why drills with randomised delays are the only reaction exercises worth doing.
The Reaction Age drill randomises the delay before green appears precisely so your brain has to stay genuinely ready instead of counting along. Five rounds, roughly three minutes.
Once simple reaction time plateaus, add decision load. Two exercises do this well: Focus Sprint, where you track multiple moving targets and respond only to the right ones, and Speed Scan, which requires finding a target shape among distractors under time pressure. Both train choice reaction time, where your remaining headroom actually lives.
4. Aerobic exercise
Fit adults have measurably faster reaction times, and the effect has two timescales. Chronically, aerobic fitness improves cerebral blood flow and dopamine signalling. Acutely, even a single 20-minute brisk walk improves same-day processing speed — enough to show up in your scores that afternoon.
The most reliable pattern is doing aerobic exercise before cognitive training rather than after; several studies find the ordering produces larger combined gains, likely via BDNF (brain-derived neurotrophic factor) priming the circuits you're about to train.
5. Caffeine, used honestly
100–200 mg of caffeine reliably shaves 10–20 ms off simple reaction time in most people. It's a real, repeatable effect — and it's temporary, and tolerance builds fast.
Treat it as a tool for testing on equal terms rather than as training. If you normally test after your morning coffee, keep doing that, because stopping will make you look like you regressed. Just don't confuse a caffeinated score with a trained one.
6. Alcohol, blood sugar and medication
Alcohol slows reaction time in a dose-dependent way that persists well past the point you feel sober — next-day effects are measurable even after blood alcohol returns to zero. Skipping meals produces a smaller but real degradation, and a range of common medications (particularly sedating antihistamines and some anxiolytics) carry reaction-time warnings for good reason.
None of this is advice to change medication. It's a reminder that if your scores swing unexpectedly, look at the last 24 hours before you look at your training.
7. Fatigue and decision fatigue
Reaction time degrades measurably after extended demanding mental work. A genuinely useful move for testing purposes is to test consistently — same time of day, similar energy level, before rather than after a hard cognitive day. This alone will tighten your measurement more than any drill.
Short breaks, hydration and stepping away from the screen all help your baseline recover. Breathing drills such as box breathing are worth trying here — they mainly work by pulling you out of the depleted state rather than by training reflexes directly.
A four-week protocol
If you want structure rather than a list of ideas, this is a reasonable template. The whole thing is 10–15 minutes a day.
| Week | Focus | Daily work |
|---|---|---|
| Week 1 | Clean baseline | Fix the setup (wired mouse, high refresh display if you have one, no extra tabs). Test at the same time each day, five trials, record the median. Don't train yet — you're establishing where you actually started. |
| Week 2 | Simple reaction drills | Sleep first. Then one daily Reaction Age session. Still no chasing your best score — record the median. |
| Week 3 | Add decision load | Alternate Reaction Age with Focus Sprint or Speed Scan. This is where most of the real improvement happens. |
| Week 4 | Consolidate and verify | Keep the routine, add the aerobic-before-training pattern, then compare week-4 rolling averages against week 1 on the same device. |
Most people see a genuine 15–30 ms improvement over four weeks. Some see nothing measurable, which is usually a sign that week 1 was done on equipment different from week 4.
How to track progress without fooling yourself
Reaction time is noisy. Day-to-day swings of ±20 ms are completely normal, so a single session tells you almost nothing. These four rules do most of the work:
- Record your median, never your best. Averaging over five trials removes most single-trial noise; your best score is just the luckiest trial.
- Compare weekly rolling averages, not daily numbers. This is the difference between seeing a trend and seeing static.
- Use the same device and roughly the same time of day. Otherwise you are measuring your hardware and your circadian rhythm.
- Ten unlucky minutes don't mean you regressed. Fatigue, distraction and caffeine all move you more than training does over a week.
If you want absolute numbers rather than just your own trend, our average reaction time benchmarks break the distribution down by age and percentile.
What doesn't work
- Generic "brain training" subscriptions. Gains on trained tasks rarely transfer to untrained ones. Practising a reaction test makes you better at that reaction test.
- Supplements beyond caffeine. Effects are small, inconsistent, and mostly disappear once you control for sleep.
- More repetitions instead of harder decisions. Volume of simple clicking plateaued early. Adding decision complexity is what extends the curve.
- Testing every hour. Practice effects from repeated testing mostly reflect learning the interface, not getting faster.
Reaction time also declines steadily with age — roughly 1–1.5 ms per year from your mid-twenties. That's worth knowing so you don't read a normal slow drift as a failure of your training. For more on why reflexes track cognitive ageing generally, see our piece on the reaction time and brain age connection.
Where to start today
Do these three things in order, and you'll get most of the available benefit:
- Get a clean baseline. Fix your setup, then take the reaction time test and write down your median.
- Sleep for a week. It costs nothing and outperforms everything else on this page per unit of effort.
- Train once a day with randomised timing. Start with Reaction Age, and move to the choice-reaction drills when simple scores plateau.
Come back in four weeks and compare medians on the same machine. That number — not today's single attempt — is the only one worth caring about.