Cubing Glossary

The whole family of turning puzzles the 3×3 belongs to - cubes of every size plus shapes like the Pyraminx and Megaminx.

A single physical piece of the cube. On a 3×3 the movable cubies are the edges and corners; the centres are fixed to the core.

A piece with two stickers, sitting between two corners. A 3×3 has 12 edges.

A piece with three stickers, at the intersection of three faces. A 3×3 has 8 corners.

The single-sticker piece in the middle of each face. On a 3×3 the six centres never move relative to each other, so they fix each face's colour.

One of the six sides of the cube (Up, Down, Front, Back, Left, Right) - a 3×3 grid of nine stickers.

A slice of nine cubies that turns together - the top layer, middle layer, or bottom layer.

One coloured facelet on a cubie. Modern cubes often use moulded plastic tiles instead of adhesive stickers, but the term stuck.

The internal skeleton the centre pieces screw into. Adjusting the core's screws sets each face's tension.

The configuration where every face shows a single colour. The one arrangement all methods aim for.

The shorthand for writing cube moves - single letters for face turns, with modifiers for direction, width, and rotations.

A 90° turn of one outer face, written with its letter: R, L, U, D, F, B. The move is clockwise looking at that face.

The apostrophe (e.g. R′) meaning turn that face counter-clockwise - the inverse of the plain move.

A 180° turn, written with a 2 (e.g. R2). Direction doesn't matter since it's half a rotation.

A 90° turn of one face. Move-count metrics differ on whether a 180° turn counts as one 'half turn' (HTM) or two quarter turns (QTM).

Turning two layers at once - the face plus the adjacent inner slice - written with a lowercase letter or a 'w' (r or Rw).

A turn of a middle layer only: M (between L and R), E (between U and D), S (between F and B).

Rotating the whole cube in your hands without turning any layer: x (like R), y (like U), z (like F). Re-orients your view.

A memorised sequence of moves that produces a specific known change while leaving the rest of your progress intact.

A short, muscle-memorised move group that flows as one motion - the building blocks algorithms are chunked into, like the Rightie (R U R′ U′).

A move pattern of the form A B A′ - do a setup, do something, then undo the setup. The building block behind setup moves and commutators.

A move (or two) done before an algorithm to position pieces, then undone afterwards. Central to BLD and F2L improvisation.

The first way most people learn - solve the cube one layer at a time using a small set of algorithms.

The first step of most beginner tutorials - form a plus of edges around one centre so each edge also matches its side colour.

The face colour you build your cross on. Beginners fix one; colour-neutral solvers choose whichever is easiest each scramble.

Completing one full face plus its side stickers - the cross edges and the four corners around them.

Inserting the four middle-layer edges to complete the first two layers in the beginner method.

The final face and its side stickers. Beginners solve it in several small steps; CFOP does it in two (OLL then PLL).

Orienting the last-layer edges so the top face shows a plus of yellow - a beginner step and the first half of 2-look OLL.

The four-move trigger R U R′ U′ - the single most common pattern in cubing, threaded through F2L, OLL, and beginner algorithms. We call it the Rightie, and its left-hand mirror L′ U′ L U the Leftie. It's traditionally known as the 'sexy move' - a name coined by Arnaud van Galen in a 2008 beginner tutorial and popular ever since - but that label has downsides: awkward to say aloud and off-putting when teaching kids.

A final U-face turn to line the last layer up with the rest of the cube, before or after an algorithm.

The most popular speedsolving method: Cross, F2L, OLL, PLL. Also called the Fridrich method after Jessica Fridrich.

The first CFOP step: solve the four bottom-layer edges around a centre, usually planned entirely during inspection.

Building the cross on the bottom face rather than the top, so the first two layers are visible and lookahead is easier.

The second CFOP step: pair a corner with its edge and insert both together, filling the bottom two layers one slot at a time.

A corner and its matching edge joined together, ready to drop into a slot as one unit.

One of the four corner-plus-edge spaces between the centres in the bottom two layers, where an F2L pair goes.

The act of bringing an F2L corner and edge together in the top layer before inserting them.

Dropping a joined F2L pair down into its slot, or placing a last-layer corner into position.

The third CFOP step: with one algorithm, orient every last-layer piece so the top face is a solid colour. 57 cases.

An OLL where no last-layer edge is yet oriented, so the top shows just the centre 'dot'. The hardest edge-orientation cases to execute.

OLL split into two easier steps - first make the top edges yellow, then orient the corners - needing only about 10 algorithms.

The corner-orientation half of 2-look OLL: seven cases that turn the top corners yellow once the edges already are.

A famous OLL case (and its algorithm) that cycles three corners' orientation - one of the first last-layer algs most people learn.

Sune's mirror - the reverse corner-orientation case. Sune and antisune together handle the two simplest OCLL patterns.

The trigger R′ F R F′ - a common alternative to the Rightie for pairing F2L and building last-layer cases.

The inverse of the sledgehammer, F R′ F′ R. The name is a playful reversal of 'sledgehammer'.

The final CFOP step: with one algorithm, permute the already-oriented last layer into its solved position. 21 cases.

PLL split into two steps - permute the corners, then the edges - needing only about six algorithms instead of 21.

The four PLL cases that permute only edges (U, H and Z perms). The edge half of 2-look PLL, and a common stepping stone to full PLL.

One of the most-used PLL algorithms - swaps two adjacent corners and two edges. A common first PLL to learn.

A PLL that swaps a diagonal pair of corners and two edges - the usual first diagonal-corner PLL people learn.

A PLL that cycles three corners while leaving the edges solved - one of the shortest and easiest permutation cases.

A PLL that cycles three edges while leaving the corners solved. The Ua and Ub cases are common first edge-PLLs.

A family of four PLLs that cycle three corners and three edges at once - often considered the trickiest to recognise.

A set that solves the last-layer edges in one algorithm when the corners are already done - mostly a curiosity for CFOP solvers.

When the last layer comes out already permuted after OLL, so no PLL is needed - a lucky, time-saving outcome.

When the last layer is already oriented after F2L, so OLL is free. Rarer than a PLL skip.

Reading the pattern of pieces to identify which case you're facing and which algorithm to fire - the skill that gates OLL/PLL speed.

A pair of matching corner stickers side by side on a face, like a car's headlights - a key landmark for spotting OLL and PLL cases.

A last-layer case where the two corners that need trading sit next to each other - recognisable by a single pair of headlights.

A last-layer case where the two corners to trade sit across the diagonal - no headlights anywhere, the tell-tale of cases like the Y-perm.

Solving the cross plus one F2L pair in one go - a way to save moves and get a head start on the first two layers.

An F2L trick that uses an unsolved slot as free space to insert a piece with very few moves, ignoring one pair temporarily.

Working on two F2L slots at the same time - influencing or setting up the next pair while solving the current one.

Efficient, case-specific F2L solutions beyond the basic pair-and-insert - empty-slot tricks, rotationless insertions, and reductions.

Watching the pieces for your next step while your hands finish the current one, so you never stop turning to search.

A practice technique - turning deliberately slowly so you never pause, training lookahead by forcing your eyes ahead of your hands.

Executing moves with quick flicks of individual fingers instead of regripping the whole cube - the basis of fast turning.

Turns per second - a measure of raw turning speed. High TPS only helps when paired with good lookahead.

Repositioning your fingers or the cube mid-solve to reach the next move. Fewer regrips means smoother, faster solves.

Being able to start the cross on any colour rather than always the same one, so you can pick the easiest cross each solve.

The up-to-15 seconds before the timer starts, used to plan the cross (and ideally a first pair) without turning.

Solving F2L with slice and wide moves instead of cube rotations, to keep pieces in view and improve lookahead.

Solving F2L pairs into the wrong-but-offset slots, then fixing the offset with a single turn later - an efficiency trick for tricky cases.

When the end of one algorithm and the start of the next overlap, letting you drop redundant moves and save turns.

A sequence that uses only two faces (classically R and U). 2-gen algorithms are very fast to fingertrick since the hands barely move.

How comfortably an algorithm flows for the hands. Cubers often pick a slightly longer alg if it fingertricks better.

A popular alternative to CFOP: build two side blocks, solve the remaining corners, then finish the last six edges - low move count, few rotations.

The Roux step that solves all last-layer corners in one algorithm, ignoring the six remaining edges - Roux's answer to CFOP's corner work.

The final Roux step: solve the last six edges (and the M-slice centres) using only M and U moves.

Assembling solved chunks of the cube intuitively rather than piece by piece - the core idea behind Roux and Petrus.

A method that orients all edges first (EOLine), so the rest of the solve needs no F or B moves - enabling ergonomic, rotationless F2L.

Edge orientation - whether an edge is 'good' (solvable without F/B moves) or 'bad'. The founding idea of ZZ and a key BLD concept.

The first ZZ step: orient every edge while placing two bottom edges, leaving a solve that avoids F and B moves entirely.

A block-building method: grow a 2×2×2 into a 2×2×3, orient the edges, finish the two layers, then solve the last layer.

Solving the entire last layer with a single algorithm - hundreds of cases, the theoretical extreme of last-layer study.

A fast, low-memory 2×2 method: build a face, orient the opposite face, then permute both layers at once.

An advanced 2×2 method that solves the last layer in one algorithm even when the first face isn't fully permuted.

A large last-layer set (~500 algs) that solves OLL and PLL in one look whenever the last-layer edges are already oriented.

Solving the last F2L pair while orienting all last-layer edges at once, so you reach OLL already edge-oriented (setting up ZBLL).

A last-layer set solvable with only R and U moves, used when the corners are already permuted - fast and ergonomic to execute.

A last-layer set that orients and permutes the corners in one algorithm when the edges are already oriented, leaving only an edge PLL.

An extension of OLL that also fixes corner permutation, so the last layer finishes with only an edge permutation left.

A technique that orients the last-layer corners while inserting the final F2L pair, skipping straight to a PLL.

Winter Variation's sibling for the front-slot insert - orients the last-layer corners while inserting the final pair, again skipping OLL.

Inserting the final F2L pair while orienting the last-layer corners, skipping OLL for a large speed gain on the right cases.

Solving the final F2L slot and the whole last-layer orientation in one step, so OLL is skipped - the umbrella set that VLS and Winter/Summer Variation belong to.

Inserting the last F2L pair while orienting the last-layer corners, assuming the last-layer edges are already oriented - leaves a corner-permutation-only last layer.

Solving more than one F2L pair with a single sequence - a double insert, keyhole, or pseudoslotting - so two pairs go home at once instead of one at a time.

Blindfolded solving - memorise the whole scramble, then solve without looking. A separate discipline with its own methods.

A beginner blindfolded method that shoots one piece at a time to a fixed spot using a single setup-and-swap algorithm - slow but easy to learn.

The standard A–X scheme that gives every corner and edge sticker a letter, so a scramble becomes a string of letters to memorise.

The memorisation phase of a blindfolded solve - turning the scramble into letter pairs or images you can recall while solving.

Two Speffz letters combined into a memorable word or image, so a long memo becomes a short story instead of raw letters.

The piece you cycle every other piece through during a BLD solve - the fixed 'home base' a tracing algorithm returns to.

A popular intermediate BLD method for edges - pieces are cycled through the buffer with the simple M2 move plus small setups.

A move sequence of the form A B A′ B′ that cycles just three pieces while leaving everything else untouched - the engine of advanced BLD.

The advanced blindfolded method - solve three pieces at a time with custom commutators, the fastest and most flexible approach.

Multi-Blindfolded - memorise several cubes, then solve them all in one sitting while blindfolded. A separate WCA event.

The standard big-cube approach - solve the centres and pair up the edges so the puzzle 'reduces' to a 3×3 you finish normally.

A big-cube method that solves most of the cross early, between centres and edge pairing, to speed up the final 3×3 stage.

A cube built for speed - corner-cutting, smooth turning, and adjustable tensions - as opposed to a stiff original Rubik's-brand cube.

A cube whose colours come from the coloured plastic itself rather than applied stickers, so they can't peel, fade, or chip.

A cube with small magnets inside that help each layer click into place, adding stability and a crisp feel. Now standard on speedcubes.

A spring-free design where repelling magnets, not steel springs, set each face's tension - for a smoother, quieter feel.

A modern core with magnets between the centres and a central ball, adding a second layer of alignment and a distinctive 'thock'.

A cube with internal sensors that report every turn over Bluetooth, so an app can track your moves live and mirror them in 3D.

Any change made to a cube to improve it - sanding, lubing, magnetising, or swapping parts between cubes.

A cube's ability to complete a turn even when the adjacent layer isn't lined up, forgiving imperfect turning at high speed.

How tightly each face is set on the core. Looser is faster but pops more; tighter is stable but stiffer - a personal-preference tune.

Silicone lubricant applied inside the cube to adjust speed and feel - thin lube for speed, thick lube for control.

When a piece flies out of the cube mid-solve, usually from loose tensions. A pop mid-solve almost always ruins the time.

Small plastic tabs on the edge pieces that stop them from popping out, trading a little corner-cutting for stability.

When a single corner rotates in place, usually from fast loose turning - it leaves the cube unsolvable until twisted back.

When the cube jams because a layer is slightly misaligned, stopping a turn dead. The opposite failure mode to a pop.

The World Cube Association - the body that runs official competitions, sets the rules, and keeps world records.

World Record - the fastest official result for an event, either a single solve or an average, recognised by the WCA.

National Record - the fastest official result within a single country. Continental (CR) records work the same way per continent.

One of the official disciplines the World Cube Association holds records in - 3×3, the other cubes, one-handed, blindfolded, FMC, and more.

Solving with a single hand - its own WCA event, using tweaked, regrip-light algorithms suited to one-handed turning.

Fewest Moves Challenge - a WCA event where you have an hour to write the shortest solution to one scramble, on paper.

The touch-pad timer used in competition - you hold both sensors to start, and lift your hands to run the clock.

A random sequence of moves that mixes the cube. Competitions use official computer-generated scrambles so everyone solves the same case.

The time for one solve, as opposed to an average. Your fastest is your 'PB single' (personal best).

A named batch of solves in your timer, with its own puzzle, settings, averages and personal bests. You pick one to solve into and can keep several side by side, say one per puzzle or one for warmups.

One continuous stretch of solving. A break of 30 minutes or more starts a new sitting, so a single session builds up from many sittings over weeks or months.

Average of 5 solves with the best and worst dropped, then the middle three averaged - the standard round format for 3×3.

A solve that actually counts toward an average - the middle results left after the best and worst are dropped.

Average of 12 solves with the best and worst dropped. A steadier gauge of your typical speed than an Ao5.

The average of every solve in a session or history, as opposed to a rolling window like an Ao5 or Ao12.

Mean of 3 - the plain average of three solves with nothing dropped. Used as the round format for some events.

Did Not Finish - a solve that doesn't count, e.g. it ended unsolved or broke a rule. Counts as the worst result in an average.

Did Not Start - a solve a competitor skips entirely, e.g. after being eliminated by a cutoff. It counts like a DNF in the average.

A two-second penalty added to a solve, most often for finishing with one face turned up to 45° out of alignment.

A time you must beat within a round's first solves to earn the rest of your attempts - a way to keep competitions running on schedule.

Personal Best - your fastest-ever result for a given statistic, whether a single or an average.

Shorthand for averaging under a time - 'sub-20' means averaging below 20 seconds. The classic milestones are sub-20, sub-15, and sub-10.

A move-counting metric. HTM counts each face or double turn as one; STM also counts a slice as one. Used to compare algorithm and solution length.

The most moves ever needed to solve a 3×3 from any scramble: 20 in the half-turn metric, proven in 2010.

An arrangement of the pieces. The 3×3 has about 43 quintillion reachable permutations, only one of which is solved.

A situation, common on even-layered cubes, where a single pair of pieces is swapped - an arrangement impossible on a 3×3, needing a special algorithm.

The branch of maths describing how the cube's moves combine. Every scramble and algorithm is an element of the 'cube group'.

The scramble where every piece is home but all 12 edges are flipped. A famous position that needs the full 20 moves to solve.

The 2×2×2 cube - all corners, no edges or centres. A fast WCA event where top solvers average around two seconds.

The 4×4×4 cube. Solved by reduction, and the smallest cube where parity - an odd edge or corner swap - can occur.

Any cube larger than 3×3. All are solved by reduction; the bigger they get, the more the centres and edge pairing dominate the solve.

A tetrahedral (triangle-faced) WCA puzzle. Its trivial rotating tips make it one of the quickest events to solve.

A twelve-sided dodecahedral WCA puzzle - like a 3×3 with more faces, solved layer by layer with familiar last-layer ideas.

A corner-turning WCA puzzle whose cuts run through the diagonals, so the whole puzzle shifts with every move.

A shape-shifting WCA puzzle that changes form as you turn it - you must first return it to a cube before permuting the pieces.

A non-cubic WCA event with nine clock dials per side, set with pins and wheels - no algorithms, pure pattern and dexterity.