Designing for One-Handed Mobile Play

When designers talk about accessibility in mobile games, the conversation usually starts with screen readers, color-blind modes, and font scaling. Those are all important — we ship all of them — but they are not the most important accessibility consideration in mobile puzzle design. That distinction belongs to one-handed play. Not because one-handed play is exotic; because it is the default. The majority of mobile puzzle sessions are played with a single thumb, on a held-not-rested device, in a context that demands short bursts of attention.

This article is about how we design for that reality. It is drawn from shipping titles like Block Puzzle: Save Girl, Puzzle Hex, and Puzzle: Water Sort on the HKQS Games platform, and from the instrumentation we run on roughly 1.4 million puzzle sessions per week. The thesis is simple: if a puzzle is not comfortably playable with one thumb, it is not comfortably playable. Every other accessibility consideration is layered on top of that foundation.

The Data: 65 Percent of Sessions Are Commute-Shaped

The most important number in mobile puzzle design is not a device spec or a market share figure. It is the share of sessions that take place in a context where the player has one hand free, a held device, and short bursts of attention. We measure this indirectly — through session length distributions, time-of-day patterns, and device-motion telemetry that signals whether the phone is being held or rested on a surface.

The commute shape is not the only context we design for — there is also the evening wind-down (longer sessions, often two-handed on a tablet), the lunch-break burst (similar to commute shape but with more flexibility), and the dedicated gaming session (rare in puzzles, but real). But the commute shape dominates the volume, and it dominates it increasingly as our catalog grows in Hong Kong and Southeast Asia, where the commute-share is even higher than in our English-market builds. Designing for the commute is not a niche consideration; it is designing for the modal player.

The question is not "can this game be played one-handed?" The question is "is one-handed play the path of least resistance, or is it a struggle against the design?" — David Tanaka, accessibility lead, HKQS Games

What the commute shape rules out

The commute shape rules out, by default, a number of design patterns that look reasonable on paper. Pinch-to-zoom is essentially impossible one-handed — it requires two fingers, which means two hands. Long-press menus that demand a held finger while the other hand taps an option are similarly excluded. Any control scheme that requires the player to reach the top of the screen with their thumb — a thumb that, on most phones, comfortably reaches only the bottom two-thirds — is asking the player to either re-grip or use a second hand. None of these are deal-breakers in absolute terms; all of them are deal-breakers in the modal context.

The Thumb Zone: Designing for Anatomy

The single most useful concept in one-handed design is the thumb zone — the region of the screen a player's thumb can comfortably reach without re-gripping the phone. The concept was popularized by Steven Hoober's research on mobile device grip, summarized well by UX Matters' coverage of mobile grip studies. The shape of the thumb zone depends on hand size, hand used (left vs. right), and device size, but the rough pattern is consistent: a comfortable arc near the bottom of the screen, tapering to unreachable corners at the top.

On a modern 6.1-inch phone held in the right hand, the thumb zone covers roughly the bottom 60 percent of the screen on the right side, tapering to the bottom 30 percent on the left side. The top 20 percent of the screen — the area where many games put their primary controls — is essentially unreachable without re-gripping. This is the anatomy that decides our layout decisions.

There is a useful refinement to the thumb-zone rule that we apply: the controls that the player uses most often should be closest to the thumb's natural rest position, not just inside the thumb zone. On a right-handed grip, the thumb naturally rests near the bottom-right of the screen; the most-used controls go there. Less-used but still primary controls — undo, restart — go elsewhere in the thumb zone, where they are reachable but not accidentally tappable. The pattern is borrowed from gamepad design, where the most-used buttons sit under the most natural finger positions.

Touch Target Sizing: The Minimum Viable Tap

The thumb zone answers the question of where to put controls. The second question is how big to make them. The standard reference for touch target sizing is the WCAG 2.1 target size guidance, which recommends a minimum of 44 by 44 CSS pixels for any tappable element. We treat 44 pixels as a floor, not a target — our puzzle pieces, undo buttons, and primary controls are sized at a minimum of 56 pixels, with the most frequently used controls sized at 64 to 72 pixels.

The reason for the larger sizing is the asymmetry between comfortable taps and strained taps. A 44-pixel target in the center of the thumb zone is comfortable. The same target at the edge of the thumb zone — where the thumb is stretched, less precise, and more prone to overshoot — becomes a source of mis-taps. Our internal telemetry shows that mis-tap rate rises sharply for targets under 56 pixels placed at the thumb-zone edge, and rises even more sharply when the player is in motion (on a moving train, for instance, which is most of our commute cohort).

• 56 px minimum for any tappable element inside the thumb zone.
• 64 to 72 px for the most frequently used controls (undo, primary piece placement).
• 8 px minimum spacing between tappable elements, to prevent edge mis-taps.
• Visual size ≠ tappable size: the hit area is always larger than the visual element, extending invisibly into the surrounding space.
• Motion compensation: targets in motion-context sessions (detected via accelerometer) get an additional 10 percent sizing buffer.

The last point — motion compensation — is the one most teams miss. A target that is comfortable to tap when the player is still becomes uncomfortable when the player is on a moving train. Our instrumentation detects the motion context through the device accelerometer (with explicit permission) and silently enlarges hit areas by roughly 10 percent during motion sessions. The player does not see the change. They just experience fewer mis-taps. The effect on session completion is small but real — roughly 3 percent on motion-context sessions — and it costs us nothing in visual design.

A mis-tap is not a player mistake. It is a design mistake. The player tapped where their thumb told them to tap. The design failed to put a large enough target there. — David Tanaka, HKQS Games

Gestures: Which Ones Survive One-Handed Play

Not all gestures are created equal under one-handed constraints. Below is our internal ranking of gesture types by their one-handed survivability, drawn from three years of instrumentation across the catalog.

Tap — survives, dominates

The humble tap is the one-handed champion. It requires a single finger, a single moment of contact, and no precision of motion. Every primary interaction in our puzzle titles is reducible to a tap, and where it is not, we have made it so. Block placement is tap-to-select, tap-to-place. Liquid pouring in water sort is tap-source, tap-destination. Even undo is a single tap. We have deliberately rejected designs that required more complex gestures as their primary interaction, because those designs would not survive the commute cohort.

Swipe — survives, with constraints

Swipe is one-handed viable but more demanding than tap. It requires a held finger and a directed motion, which means it is more affected by motion context. We use swipe for one purpose: scrolling the level-select carousel. We do not use it for primary puzzle interactions, because the precision of a swipe — the start point, the end point, the direction — is harder to deliver reliably when the player is on a moving train. Where a swipe would be the natural gesture, we look for a tap-equivalent.

Drag — fragile, used sparingly

Drag-and-drop is the most natural gesture for many puzzle mechanics — pick up a piece, move it, release it. It is also the most fragile under one-handed constraints, because it requires a held finger for the duration of the motion, and the precision of the release point is what decides whether the move is valid. We use drag in Block Puzzle: Save Girl — there is no clean tap-equivalent for "place this block here, not there" — but we wrap it in forgiveness: the drop target is large, the snap-to-grid is generous, and an accidental release does not commit the move. A drag that releases mid-motion is treated as a cancelled drag, not a misplaced piece.

Pinch, two-finger, long-press menus — excluded

Pinch-to-zoom, any two-finger gesture, and long-press-then-tap menus are excluded from primary interactions entirely. They require either two hands or a level of dexterity that the commute cohort cannot reliably deliver. We use long-press only for non-primary actions — "long-press to access settings" is fine, because the player can do it when they are not in a puzzle-solving moment. Long-press as a primary interaction is, for our catalog, banned.

Case Study: One-Handed Optimization in Block Puzzle

Block puzzle is the format where one-handed design choices are most consequential, because the primary interaction — pick up a piece, place it on the board — is the most demanding gesture in our catalog. The piece has to be picked up, moved, and released at a specific location, and the move is committed on release. Below are the specific optimizations we shipped on Block Puzzle: Save Girl, in roughly the order they moved the metrics.

1. The piece tray lives in the bottom-right thumb rest

The three pieces available for placement sit in the bottom-right of the screen, in the thumb's natural rest position. The player can pick up a piece with a single, short motion. We tested a centered tray — visually balanced but requiring a longer reach — and the centered version's first-session completion dropped by 8 percent. The thumb-rest position is the single most consequential layout decision in the title.

2. Drop targets snap generously

The block grid is 8 by 8, but the drop target for a piece is not a single cell. It is a generous snap zone that extends roughly 30 percent beyond the visible cell boundary, in every direction. A player who releases a piece slightly off-center still lands it on the intended cell. We measure "intended-but-misplaced" releases — where the player's release point is closer to one cell but the piece lands on another — and the snap-zone change reduced these by 71 percent.

3. Accidental releases do not commit the move

If a player picks up a piece and releases it without moving it meaningfully — defined as less than 20 pixels of motion — the piece returns to the tray. The move is not committed. This is the forgiveness that makes one-handed drag viable on a moving train: an accidental release, which would otherwise waste a piece or break a strategy, simply does not happen.

4. Undo is a single tap, in the thumb zone

The undo button is in the bottom-right, immediately adjacent to the piece tray, sized at 72 pixels. It is the most-used control in the game — players undo roughly 1.4 times per completed level — and it sits exactly where the thumb can reach it without travel. The button's hit area extends beyond its visual boundary, so even an imprecise tap lands.

The pattern across these optimizations is consistent: each one removes a small failure mode that one-handed play introduced, without changing the underlying mechanic. The puzzle is the same puzzle. The skill required is the same skill. What changed is that the design stopped punishing the player for the physical constraints of the context they were playing in. That is the whole discipline of one-handed design in a sentence.

One-handed design is not a separate accessibility workstream. It is the baseline. Two-handed play is the optimization; one-handed play is the constraint that decides whether the game gets played at all. — David Tanaka, HKQS Games

Left-Handed Players: The Forgotten Half

Roughly 10 percent of players are left-handed, and a meaningful share of right-handed players also play left-handed when their right hand is occupied — holding a transit pole, carrying a bag, drinking coffee. Designing for right-handed-only is designing for one cohort and excluding another. Our approach is hand-detection by default — the game asks the player, on first session, which hand they use, and the layout mirrors accordingly. The piece tray moves to the bottom-left for left-handed players; the undo button mirrors; the entire primary control surface is reflected.

The mirroring is not free — it requires every layout decision to be expressed in terms of "thumb-side" and "non-thumb-side" rather than "left" and "right" — but it is the only honest way to serve both cohorts. The alternative, a "left-handed mode" buried in settings, is a setting that most players will never find. We have measurably better retention among left-handed players since shipping the mirror-by-default behavior, and the cost has been entirely in engineering, not in visual design.

The Future of One-Handed Design

Looking forward, three trends are likely to shape how we think about one-handed play over the next eighteen months. None of them are revolutionary; all of them are extensions of the principles above into new contexts.

1. Foldable and large-screen devices

Foldable phones and large-screen devices (tablets, foldables in unfolded mode) break the assumption that the screen is roughly 6 inches. On a 7.5-inch unfolded phone, the thumb zone is a smaller fraction of the screen, and controls that were comfortably reachable on a 6-inch device become unreachable. Our current approach is to detect the screen class and reflow the layout — placing primary controls in the bottom 50 percent of large screens, rather than the bottom 60 percent of standard phones. The foldable share of our traffic is still small (under 4 percent in 2026) but rising fastest among premium-device cohorts, and we expect it to be a meaningful share by 2028.

2. Haptic feedback as a confirmation layer

The second trend is the use of haptic feedback — device vibration — as a confirmation layer for one-handed interactions. When a player cannot reliably see their thumb (because it is covering part of the screen), haptic feedback can confirm a successful tap, a valid drop, an undo. We are experimenting with a "haptic-first" feedback layer on a 2026 prototype of Block Puzzle: Save Girl, and early results show a 4 percent reduction in repeated taps (players tapping twice because they were unsure the first tap registered). The pattern is borrowed from native iOS and Android apps, where haptic confirmation is now standard for primary interactions.

3. Voice as an auxiliary input

The longest-horizon trend is voice as an auxiliary input for non-primary actions. "Undo," "restart," "next level" — these are commands a player can issue without taking their thumb off the screen. The technology is mature; the constraint is social (players do not want to talk to their phone on a quiet train). We expect voice to remain a niche input for puzzle play, useful primarily for players with motor impairments that make thumb play difficult, but worth shipping as an option.

What This Adds Up To

The discipline of one-handed design is, in the end, the discipline of respecting context. A puzzle that is brilliant on a desktop monitor with a mouse may be unplayable on a phone on a moving train. The puzzle has not changed; the context has. The work of one-handed design is to make the puzzle playable in the context where the player actually is, not the context where the designer wishes they were.

At HKQS Games, the thumb-zone rule, the touch-target minimums, the motion compensation, and the hand-mirroring are now part of the standard review every title goes through before it ships. None of these are exotic. None of them are expensive. They are simply the discipline of designing for the modal player — a player with one hand free, a held device, and a few minutes to spare. Get that right, and the rest of accessibility is a much easier problem. Get it wrong, and no amount of color-blind modes or font scaling will recover the players who never made it past the first session.

The best one-handed design is invisible. The player never notices that the controls were placed for their thumb, that the targets were sized for their motion, that the layout mirrored for their hand. They just feel like the game was made for them. That is the goal. — David Tanaka, HKQS Games