Speedcubing Glossary

A fixed sequence of moves that transforms a cube from one specific state to another predictable state. In speedcubing, algorithms are memorized for OLL and PLL recognition so the solver can execute them without thinking about individual moves. A good algorithm is efficient (low move count), ergonomic (finger-friendly), and consistent (produces the same result every time regardless of how fast you execute it). Most competitive CFOP solvers have 70–80 algorithms memorized for OLL and PLL combined.

OLL case number 26, and one of the 10 most important OLL algorithms to learn. Anti-Sune orients the last-layer corners (three corners need rotating while one is already correctly oriented) and is the mirror image of the Sune algorithm. Recognized by a single yellow corner on the top layer with the other three corners facing sideways. Algorithm: R U2 R' U' R U' R'. Anti-Sune appears in roughly 1 in 8 solves and is the foundation for many 2-look OLL approaches.

The U (top) face rotation (U, U', U2, or no move) applied before recognizing and executing a PLL algorithm to align the top layer corners with a recognizable PLL pattern. AUF adds zero moves in the best case and three moves in the worst case. For OLL, AUF refers to rotating the top layer to match the correct angle before execution. Advanced solvers learn to pre-recognize which AUF is needed during the final moments of F2L so the transition to OLL is instantaneous.

A WCA event where the solver memorizes the cube's scrambled state, puts on a blindfold, and then solves it entirely from memory without being able to see the cube. BLD is one of the most impressive speedcubing events. The world record for 3x3 BLD is under 13 seconds total time (memorization + execution). Most BLD solvers use a method called M2/OP or 3-style that relies on piece-by-piece memorization using a letter-pair system called "lettering schemes."

A solving approach where the solver builds large blocks of solved pieces rather than solving one layer at a time. Block building is central to methods like Roux and ZZ. Unlike CFOP's step-by-step structure, block building requires the solver to see and connect pieces across multiple layers simultaneously, making it harder to learn but potentially more move-efficient at high levels. Many advanced CFOP solvers also incorporate block-building intuition into their F2L.

The most popular speedcubing method, used by the majority of competitive solvers. CFOP stands for Cross, F2L (First Two Layers), OLL (Orientation of Last Layer), and PLL (Permutation of Last Layer). Developed by Jessica Fridrich in the 1980s and popularized over the internet, CFOP uses 78+ algorithms and is the method that virtually every sub-10-second solver has mastered. The method solves the cube in four distinct stages, each with a well-defined goal. See the CFOP Beginners Guide for a full breakdown.

The ability of a cube's mechanism to complete a layer turn even when the adjacent layers are slightly misaligned. Good corner cutting is the primary quality indicator for a speedcube. Without it, fast solving causes constant lockups. Corner cutting is described in degrees (e.g., "45-degree forward corner cutting") and can be positive (forward cutting, where the cube can turn despite misalignment) or negative (reverse cutting, where it can return from over-rotation). Modern flagship cubes cut 45+ degrees in both directions.

The first step of CFOP. The solver places all four edge pieces of the bottom face to create a "plus sign" or cross shape on the bottom layer, with each edge correctly aligned to its corresponding center. A well-executed cross takes 6–8 moves on average and should be planned entirely during the 15-second inspection period. Advanced solvers plan the first F2L pair as well during inspection. The cross is solved on the bottom face (white for most solvers) so the solver can simultaneously look ahead to the F2L pairs.

A move that rotates the entire cube without turning any individual face. x rotates the cube around the R-L axis (tipping forward), y rotates around the U-D axis (spinning horizontally), and z rotates around the F-B axis (rolling sideways). Cube rotations are denoted in lowercase in cube notation and are generally considered undesirable in algorithms because they interrupt the solving rhythm. Advanced solvers minimize unnecessary rotations, especially during F2L, because each rotation adds dead time. Rotations with primes (x', y', z') rotate in the opposite direction.

A result in a WCA competition where the solve is disqualified. DNF is given when a solver cannot complete the solve within the time limit, makes an illegal move or touches the timer illegally, has pieces that pop out and cannot be reinserted in the same configuration, or has a cube that is more than two moves away from solved at the end. In casual solving, "DNF" is used informally to mean a solve where something went wrong, a pop, a wrong move, or running out of time. DNF results count as infinite time in official averages.

A WCA competition result where the solver elects not to attempt a solve at all. This can happen if the previous result is already good enough, if the solver wants to accept a DNF rather than risk a slower result, or if they are unable to compete due to an injury. DNS is worse than DNF in official WCA results: it counts as worse than any finished time. In the five-solve average format, a single DNS means the round result is a DNS regardless of other solves.

Moves that turn two layers simultaneously. Lowercase letters in cube notation denote double layer moves: r turns both the R face and the middle slice adjacent to it. These moves are common in advanced algorithms because they can solve multiple pieces simultaneously. In 4x4 and larger cubes, double layer moves (written as Rw, Uw, etc.) are used extensively. On a 3x3, double layer moves are equivalent to the corresponding single layer move plus a middle slice move (r = R M').

One of the 21 PLL cases, and one of the most challenging to execute quickly due to its move count and structure. The E-Perm swaps all four edges of the last layer diagonally, while leaving all four corners in place. It is recognized by having no solved edges on any face after OLL but all corners appearing in their correct positions. The most common algorithm for E-Perm is: x' R U' R' D R U R' D' R U R' D R U' R' D' x. Advanced solvers often use a variant that removes the cube rotation.

In speedcubing, ergonomics refers to how well an algorithm or sequence of moves can be executed quickly and smoothly given the physical constraints of human hands. An ergonomic algorithm flows naturally from one move to the next, minimizes awkward finger positions, and can be executed in a single continuous motion rather than in segments. When choosing between two algorithms that solve the same case, experienced solvers often prefer the more ergonomic one even if it has more moves, execution speed matters more than move count at high levels.

The second step in CFOP, where the solver simultaneously inserts the four corner-edge pairs into the four slots of the first two layers. Rather than solving the second layer after the first (as in beginner methods), CFOP F2L pairs a corner with its matching edge piece and inserts them together using intuitive or algorithmic techniques. There are 41 distinct F2L cases, though most experienced solvers handle the majority intuitively. F2L is typically the biggest time sink in a solve, representing 40–50% of total solve time. Look-ahead during F2L is the single most impactful skill for intermediate solvers to develop.

Specific hand positions and movements that allow individual cube faces to be turned using a single finger rather than the entire hand. Mastering finger tricks is essential for achieving fast TPS (turns per second). The most common finger tricks involve the index finger for U moves, the ring finger for U' moves, and the right thumb for D moves. Different algorithms favor different sets of finger tricks, which is one reason why speedcubers often have personal preferences for specific algorithm variants even when move counts are identical.

Solving the OLL step in a single algorithm by recognizing and executing one of the 57 OLL cases. Contrasted with 2-look OLL, where the OLL is solved in two stages (first orienting edges, then orienting corners). Full 1-look OLL requires memorizing all 57 algorithms but eliminates one algorithm execution per solve, a significant time saving at high levels. Most solvers approaching sub-15 seconds begin learning full OLL. See the Top 10 OLL Cases guide for where to start.

A family of four PLL cases (Ga, Gb, Gc, Gd) where three corners and three edges cycle simultaneously. The four G-Perms are among the most complex PLL cases to recognize because they look similar to each other and require checking multiple faces to distinguish. They are also among the more move-heavy PLL algorithms. Many advanced solvers learn all four G-Perm variants but prefer to avoid them through PLL skip or AUF optimization. Together, the four G-Perms account for about 18% of all PLL cases.

The PLL case that swaps two pairs of opposite edges on the last layer while leaving all four corners in place. H-Perm is easily recognized: all four corners appear solved (two bars of matching color on opposite sides) while the two pairs of edges are swapped. It has a clean, fast algorithm (M2 U M2 U2 M2 U M2) and is one of the fastest PLL cases to execute. The name comes from the shape the edge pieces make when viewed from above, which resembles the letter H.

A PLL recognition pattern where two adjacent corner pieces on the same face share the same color, creating a visual "bar" or "headlight" pattern. Headlights are the primary recognition feature for most PLL cases. A face with a single headlight (two matching corners, one different) narrows the possibilities significantly. Two headlights on opposite faces narrow the case to a small family. No headlights (all corners different on every face) is also a recognizable pattern. Learning to spot headlights instantly is the fastest path to sub-second PLL recognition.

In WCA competitions, solvers receive 15 seconds of inspection time before each solve during which they may hold and examine the cube without making any moves. During inspection, experienced solvers plan the entire cross (and often the first F2L pair) before their hands touch the solve timer. Inspection time is measured from when the judge says "go" until the solver lifts their hands from the Stackmat timer to begin solving. Using inspection time efficiently is one of the highest-leverage skills in competitive speedcubing.

The most common beginner method for solving the Rubik's Cube. LBL solves the cube one horizontal layer at a time: first the bottom layer (daisy and cross), then the middle layer (inserting middle-layer edges), then the top layer (OLL + PLL, or a series of beginner algorithms). LBL typically requires 5–7 distinct algorithm sets and produces solve times of 60–120 seconds for most beginners. It is the natural starting point before transitioning to CFOP, where F2L replaces the separate first-and-second-layer steps.

When the cube physically jams during a solve, preventing a face turn from completing. Lockups occur when corner cutting limits are exceeded, when a piece is slightly misaligned, or when cube tension is too tight. Even brief lockups cost significant time: a 0.3-second lockup represents a 1–2% time loss on a 25-second solve. Reducing lockups is one reason top solvers maintain their cubes carefully, adjusting tension and applying lubricant regularly to keep the mechanism smooth.

The practice of tracking pieces for the next step of the solve while executing the current step. During F2L, look-ahead means locating the next corner-edge pair while inserting the current pair, so no time is lost to "searching" between pairs. Look-ahead is the single most important skill separating intermediate from advanced solvers. It cannot be developed by solving faster; it requires deliberately slowing down during practice to train the eyes to track pieces independently from the hands. See the F2L Look-Ahead guide for drills.

A substance applied to a cube's internal mechanism to reduce friction, improve turning speed, and protect plastic parts from wear. Speedcubing lubricants come in two main categories: silicone-based (thin, fast, good for mechanisms) and Teflon/PTFE-based (thick, slow, good for reducing noise). Popular lubricants include DNM-37, Martian, Angstrom Dignitas, and Weight 5. Applying the wrong lubricant or too much of it can ruin a cube's feel for weeks. Most flagship cubes come pre-lubed but benefit from re-lubing every 6–12 months of regular use.

A speedcube that contains small neodymium magnets inside its corner and edge pieces that attract to each other when layers are nearly aligned. Magnets provide tactile feedback, help pieces "snap" into position during fast turning, and reduce overshooting. Virtually all modern flagship cubes are magnetic. Magnet strength is a personal preference: weaker magnets (like GAN's softest setting) feel floaty and fast; stronger magnets feel more controlled and clicky. Most solvers settle on a strength between 3–6 on a 10-point scale.

A dodecahedron-shaped puzzle (12 pentagonal faces) that is a WCA event and popular alternative to the 3x3. The megaminx follows the same basic CFOP principles but on a 12-faced puzzle with 50 movable pieces rather than 20. Megaminx solves typically take 40–90 seconds for intermediate solvers. The WCA world record single is under 25 seconds. Learning megaminx after 3x3 is straightforward for CFOP solvers since the last layer algorithms are nearly identical to 3x3 OLL and PLL.

One thousandth of a second. WCA competition times are officially recorded to the nearest hundredth of a second (two decimal places). Stackmat timers display times in this format. At elite levels, solvers discuss improvements in hundredths of seconds. "I dropped 0.03 from my average this month" is a meaningful and celebrated statement. For intermediate solvers, the goal is typically to improve by whole seconds per improvement cycle.

The standardized system for describing cube moves. The six face moves are R (right), L (left), U (up), D (down), F (front), and B (back). Each can be followed by ' (prime, meaning counterclockwise) or 2 (180 degrees, same in either direction). Middle slice moves are M (between R and L), E (between U and D), and S (between F and B). Cube rotations are x, y, and z. This notation system was developed for CFOP and is used universally in the speedcubing community for sharing algorithms, reconstructions, and method explanations.

A WCA event where the solver completes the entire solve using only one hand. OH requires a different technique from two-handed solving: the thumb does much of the work that two-hand solvers assign to the left hand, and algorithms must be re-learned or adapted for one-hand ergonomics. OH world records are typically 1.5–2x slower than two-handed records. Many competitive solvers practice OH as a supplementary event and find it improves their two-handed look-ahead because it forces slower, more deliberate recognition.

The third step in CFOP, where all pieces on the top face are rotated so that the top color (usually yellow) faces up, without necessarily placing them in the correct positions. There are 57 distinct OLL cases (plus the "skip," where OLL is already solved). Beginners use 2-look OLL (10 algorithms: 3 for edges, 7 for corners). Full OLL requires all 57 algorithms but reduces OLL to a single algorithm per solve. OLL algorithms range from 6 to 18 moves. Recognition involves checking the yellow stickers on the top face and the top row of each of the four side faces.

A solver's fastest single time or best average on a given puzzle. PBs are tracked separately for single solves (the fastest individual solve ever) and averages (typically ao5 or ao12). Most speedcubers are more proud of their average PB than their single PB, since singles can result from lucky scrambles. WCA competitions recognize both single and average results. A "PB single" and "PB average" on the same day ("PB double") is considered an exceptional session.

A speedcubing method developed by Lars Petrus that builds the cube in stages using block building rather than layer-by-layer solving. Petrus starts with a 2x2x2 block, expands it to a 2x2x3 block, corrects edge orientation, and then completes the last layer. The method is highly move-efficient but difficult to learn intuitively, since it requires strong block-building vision. Some solvers find Petrus produces more ergonomic and efficient solutions than CFOP for certain scrambles.

The fourth and final step in CFOP, where the correctly oriented last layer pieces are permuted (moved) into their correct positions to complete the solve. There are 21 distinct PLL cases (plus the "PLL skip," where all pieces are already correctly placed). PLL algorithms range from 6 to 20 moves. Recognition uses patterns on the top and side faces, specifically looking for "headlights" (two matching adjacent corners) and "bars" (two or three matching colors in a row on a side face). Advanced solvers recognize any PLL case in under 0.3 seconds without rotating the cube.

When the last layer pieces are already correctly permuted after OLL, requiring no PLL algorithm. PLL skips occur with a probability of 1/72 on a random scramble. A PLL skip combined with an OLL skip (probability 1/216) produces a "full last layer skip," one of the luckiest events in speedcubing. When a PLL skip occurs, the solver needs only an AUF rotation to complete the cube, potentially saving 1–3 seconds on the solve time.

When a piece (or pieces) physically ejects from the cube during a solve. Pops are rare with modern well-tensioned cubes but can still happen during very fast turning or when magnets are too weak. In WCA competition, if a piece pops during a solve, the solver may re-insert it and continue, but if the cube is more than two moves away from solved at the end, the solve is a DNF. Preventing pops is one reason tension adjustment is important: too loose and the cube pops; too tight and it lockups.

The ability to quickly identify which OLL or PLL case is present on the cube, or which F2L case needs to be solved. Recognition is the cognitive skill that separates fast solvers from slow ones, most intermediate solvers know all the algorithms but pause because they can't identify which case they're looking at. Good recognition is fast (under 0.3 seconds for PLL, under 0.5 seconds for OLL) and accurate. Recognition improves through deliberate repetition drilling (which is what this trainer is designed to support), not through casual solving alone.

A detailed written record of every move made during a specific solve, usually notable because of an exceptionally fast time. Reconstructions allow other solvers to analyze what made a particular solve fast (efficient cross, lucky F2L pairs, OLL/PLL skip) and learn from it. Top solver reconstructions are regularly posted in speedcubing forums and analyzed move-by-move. Reconstructions are created by working backward from a known scramble, or by watching slow-motion video of the solve.

The second most popular competitive speedcubing method after CFOP. Roux solves the cube by building two 1x2x3 blocks on the left and right sides, solving the top corners (CMLL), and then solving the remaining six last-layer edges (LSE) using primarily M-slice moves. Roux requires fewer algorithms than CFOP (only the CMLL set, ~42 cases) and has a lower average move count, but requires strong block-building intuition. Several sub-5-second solvers use Roux, making it a viable method for elite speedcubing.

The sequence of moves applied to a solved cube to put it into a randomized state before a solve attempt. In WCA competitions, scrambles are computer-generated using a WCA-approved scramble program (TNoodle) and applied by a certified scrambler before each solve. All competitors in the same round use identical scrambles, ensuring fairness. Scrambles typically consist of 20–25 moves and are designed to fully randomize the cube from any starting state. Each scramble is printed on a slip of paper and handed to the scrambler to ensure consistency.

A WCA puzzle where the rotation axes pass through the corners of the cube rather than through the face centers. The Skewb has only five distinct piece types and a relatively small number of states (~3.1 million compared to 43 quintillion for a 3x3). Top Skewb solvers regularly sub-2 to sub-3 seconds. The most popular solving method involves building layers using intuitive techniques with a small algorithm set for the last layer. Skewb is considered one of the more beginner-friendly WCA events.

The official timing device used in WCA competitions. A Stackmat timer displays time to the hundredth of a second and requires the solver to place both hands flat on the two sensors to stop it. The timer starts when the solver lifts their hands from the sensor pads and stops when they place their hands back down after completing the solve. The specific Stackmat models (Gen 3, Gen 4, Gen 5) are officially approved by the WCA. Serious solvers typically own a Stackmat for home practice to simulate competition conditions.

A shorthand for averaging below X seconds on the 3x3 (or other puzzle). "Sub-20" means averaging under 20 seconds, "sub-10" means averaging under 10 seconds, and so on. Sub-X milestones are the primary goal markers in speedcubing. Common milestones: sub-60 (beginner competency), sub-30 (intermediate), sub-20 (solid CFOP foundation), sub-15 (full OLL required), sub-10 (full CFOP mastery + advanced F2L), sub-7 (elite level). See the Sub-20 in 30 Days guide for a structured plan.

OLL case number 27, and possibly the single most important OLL algorithm to learn first. Sune is the most common OLL case appearing in real solves, occurring in roughly 1 in 8 scrambles when using 2-look OLL with edge orientation first. Recognized by a single yellow corner on the top-left with a yellow sticker on the front face. Algorithm: R U R' U R U2 R'. Sune also forms the basis of many other OLL algorithms. Its mirror (Anti-Sune) is equally important.

The tightness of a cube's screws, which controls how loose or tight the layers feel when turning. Adjusting cube tension is one of the most impactful hardware optimizations. Tension that is too loose causes the cube to feel mushy and unstable; too tight creates lockups and resistance. Most modern cubes use a dual-adjustment system with inner and outer screw layers. Tension should be adjusted so the cube feels stable at slow speeds but smooth and free-turning at fast speeds. Tension and lubrication work together and should be adjusted simultaneously.

One of the 21 PLL cases, where two adjacent corners swap and two adjacent edges swap simultaneously. T-Perm is one of the most common PLL cases and has a fast, ergonomic algorithm: R U R' U' R' F R2 U' R' U' R U R' F'. It is recognized by a single "bar" (two matching colors in a row) on one face. T-Perm appears in approximately 1 in 12 solves and is typically one of the first PLL algorithms advanced beginners learn after U-perm.

A measure of execution speed: how many individual move turns a solver makes per second on average during a solve. Elite solvers reach 12–15 TPS during peak execution. Average competitive solvers operate at 7–10 TPS. TPS by itself is not a reliable indicator of solving ability; what matters is effective TPS, which accounts for pauses during look-ahead and recognition. A solver doing 5 TPS with no pauses can outperform a solver doing 10 TPS who pauses for 2 seconds after every F2L pair.

A short, fixed sequence of moves (usually 3–5 moves) that is executed as a single fluid motion. The most famous trigger is the Sexy Move: R U R' U'. Triggers appear as sub-sequences within longer algorithms and are trained until they feel like a single gesture rather than a series of individual turns. Recognizing and drilling triggers is fundamental to increasing TPS and reducing the mental overhead of executing long algorithms, since the brain treats each trigger as one unit rather than multiple individual moves.

Two closely related PLL cases (Ua and Ub) where three edges cycle on the top layer while all four corners remain in place. U-perms are among the easiest PLL cases to recognize (the corners all appear correct on every face, and three edges cycle in a visible pattern) and have relatively fast, ergonomic algorithms. Ua algorithm: R U' R U R U R U' R' U' R2. Ub algorithm: R2 U R U R' U' R' U' R' U R'. U-perms are typically the first PLL algorithms beginners learn after T-Perm.

The international governing body for speedcubing competitions. The WCA sets rules, approves competitions, trains judges, and maintains official world records for all 17 WCA events (3x3, 2x2, 4x4, 5x5, 6x6, 7x7, 3x3 OH, Clock, Megaminx, Pyraminx, Skewb, Square-1, 3x3 BLD, 4x4 BLD, 5x5 BLD, 3x3 MBLD, and FMC). The WCA maintains a public database of all official results at worldcubeassociation.org. Competitions are organized by volunteer delegates certified through a formal training process.

The fastest officially recorded time for a given event on a given puzzle, maintained by the WCA. World records are set only at WCA-sanctioned competitions using WCA-certified equipment and judges. The 3x3 single world record has dropped from over 50 seconds in the 1980s to below 3.1 seconds as of 2026. World records exist for both single solves and averages (typically ao5). The WCA also maintains continental records (NR/CR) and national records. Seeing a live world record attempt is considered one of the most exciting moments at a major competition.

A PLL case that swaps two pairs of adjacent edges on the top layer in a "Z" pattern, leaving all four corners in place. Z-Perm is easily identified by two opposite faces each showing a split color pattern (neither matching completely). Like H-Perm, it uses primarily M-slice moves: M2 U M2 U M' U2 M2 U2 M' U2. Z-Perm is less common than T-Perm or U-Perm but has a satisfying algorithm that experienced solvers execute quickly once the M-slice technique is smooth.

A speedcubing method that begins by orienting all edges on the cube (EO, or Edge Orientation) before building blocks, allowing the rest of the solve to be completed without M-slice moves or cube rotations. ZZ proceeds in three phases: EOLine (orienting edges and solving two specific edge pieces), ZZF2L (building both F2L blocks using only R, U, L moves), and ZBLL or 2GLL last layer. ZZ has a very low move count when planned well but requires strong look-ahead and comfort with the EOLine phase, which is less intuitive than CFOP cross planning.