← All guides · Fundamentals

Why Left-Hand Speed Matters (And What Reaction Training Can Actually Fix)

There are two claims floating around gaming forums that sound similar and aren't. The first is "you can train your reaction time." The second is "you can train the delay between deciding to act and the game receiving your input." The first is mostly false. The second is emphatically true, it's worth more milliseconds than the first would be anyway, and almost nobody practises it deliberately.

This guide is about the second one — where those milliseconds actually live, what the published research supports, and what it doesn't.

Your right hand gets all the practice

Consider how lopsided PC gaming practice is. Aim trainers are a genuine industry: Aim Lab and KovaaK's have millions of users, all drilling the right hand's fine motor control in isolation, with detailed analytics.

The left hand gets nothing. It handles four directions of movement, a stack of ability keys, weapon swaps on the number row, crouch, sprint, walk, interact, and whatever your game calls "the button you press when things go wrong" — and the standard practice regimen for all of that is "play more games." The closest available tool is a typing test, which trains prose across the whole keyboard at a rhythm nothing like a game.

That gap is the whole argument. Not that left-hand speed matters more than aim — it doesn't — but that it's the one with no training infrastructure, which makes it the cheapest place to find improvement.

Where your milliseconds actually go

When something happens on screen and you respond, the delay breaks into parts that behave very differently:

The part you can't train (much)

Simple visual reaction time — a light appears, you press a button — has been measured in laboratories for over a century, and healthy young adults typically land somewhere in the region of 180–250 ms. Some of that is pure physics: light hits the retina, a signal is transduced and relayed through the visual system, a motor command travels out to the hand. You are not going to negotiate that down.

Two things worth knowing before you compare yourself to a number you saw online:

• Browser tests read higher than lab figures, and that's not you. Any reaction test running in a web browser measures your display's latency, the browser's frame timing and your keyboard's polling and debounce alongside your biology. That's why online benchmarks tend to report noticeably slower medians than laboratory studies of the same task. Compare your browser number to your own browser number over time, not to a published lab figure.
• The "elite players react in 150ms" claim is not well supported. You'll see it repeated constantly. I removed it from an earlier version of this guide because I couldn't find peer-reviewed data behind it. Simple visual reaction times below roughly 150 ms sit close to the physiological floor, and published work on action gamers describes advantages measured in tens of milliseconds, not a different tier of human.³⁴

What the research does support is real but modest: people with extensive action-game experience respond faster on a range of perceptual tasks without a corresponding loss of accuracy — that is, they're genuinely quicker rather than just guessing sooner.⁴ The broader claim that action games durably improve general cognition has had a rockier time, with meta-analyses and replication attempts producing weaker effects than the early studies. Treat "games make you faster at games" as well supported and "games make you faster at everything" as unsettled.

The part you can

Now the useful half. After your brain has decided to press Q, something still has to happen: your finger must get to Q and press it. If Q is under a resting finger, that's fast and automatic. If your hand has drifted, or the bind is a stretch, or you're not quite sure where it is, you've added a search-and-travel step that can dwarf your reaction time.

Two well-established findings explain why this part behaves so differently:

Fitts's law describes movement time as a function of how far you must travel and how small the target is — further and smaller means slower, in a predictable relationship.¹ It was formulated for aimed movements of the arm rather than fingers on keycaps, so treat it as directional rather than a formula you can plug your keyboard into. The direction is unambiguous, and it's the entire case for keeping reaction-critical keys close to your home position.

Hick's law describes how choice reaction time grows as you add alternatives to choose between — roughly logarithmically with the number of equally likely options.² This is why a 30-bind MMO rotation feels mentally slower than a 6-bind shooter. The important caveat is the interesting bit: the effect shrinks substantially with practice. A heavily drilled response stops behaving like a deliberate choice at all. That is precisely what you are buying with reps.

What to realistically expect

Honest targets, because inflated ones just make people quit:

• Execution lag on your core keys: reducible to near zero. This is the big one, it's the one the trainer measures per key, and it's mostly a matter of reps. Two weeks of short daily sessions is usually enough to flatten the outliers.
• Your average measured reaction: a modest improvement, then a plateau. Some of the early gain is genuine learning and some is familiarity with the task itself. Both are real for your purposes; neither continues indefinitely.
• Your raw neural latency: essentially unchanged. Nothing on this site will alter it, and be sceptical of anything claiming otherwise.

The reason this is still worth doing is that the trainable component is typically the larger one, and it's the one nobody is working on.

The boring inputs matter more than the interesting ones

Two variables move reaction time more than any drill will, and both are unglamorous.

Sleep. This is the best-evidenced item on the list. A meta-analysis of short-term sleep deprivation found simple attention and vigilance tasks — exactly the kind of alertness that reaction speed depends on — among the most strongly affected domains.⁵ Sleep also does something more specific for practice: motor skills consolidate between sessions, and a night of sleep after learning a motor sequence produces measurable improvement without any additional practice.⁶ Practising then sleeping beats practising for twice as long.

Caffeine. Moderate doses reliably improve vigilance and reaction time, particularly when you're fatigued or sleep-restricted.⁷ It's a real effect and a small one, it doesn't substitute for sleep, and it comes with the obvious tolerance and jitter trade-offs. It's a top-up, not a strategy.

Warming up is worth two minutes as well. Cold, stiff hands are slower hands, and starting a ranked session on your first inputs of the day means playing your worst round when it counts.

How to actually work on this

Concretely, in order of what returns the most per minute spent:

1. Find your outliers. Run a few rounds and look at the reaction heatmap rather than the score. Nearly everyone has one or two keys that are dramatically slower than the rest — usually a number key or a bind just outside comfortable reach.
2. Fix them at the source first. If a key is slow because it's a stretch, rebinding beats drilling. See the keybinds guide for the framework.
3. Then drill what's left. Short daily sessions, at a difficulty where you miss occasionally. The muscle memory guide has a ten-minute routine and the evidence behind its structure.
4. Train the keys your game uses. Not a generic set — the preset that matches your genre.

▶ Find your slowest key