// glossary

Glossary

Overview of Roux related terms.

The method overview

What Roux is, and the four phases of the solve.

# Roux method

also: Roux

The Roux method is a Rubik's Cube speedsolving method that builds two 1×2×3 blocks on opposite sides of the cube, solves the last-layer corners in one step, then finishes the remaining six edges with only middle-slice and U turns.

Invented by Gilles Roux, it is known for a low move count and heavy reliance on M-slice moves rather than rotations. A solve is done in four phases: FB → SB → CMLL → LSE.

See Rouxvolution — guides · Kian Mansour — Roux tutorials ↗

# FB — First Block

also: FB · left block

FB (First Block) is the first phase of Roux: build a 1×2×3 block on the left side of the cube, around the L centre.

The block is made of the L centre, the DLF and DBL corners, and the FL, BL and DL edges. First Block is solved intuitively — there are no algorithms — so it is the step that most rewards planning during inspection.

# SB — Second Block

also: SB · right block

SB (Second Block) is the second phase of Roux: build the matching 1×2×3 block on the right side of the cube, around the R centre, without disturbing the First Block.

The block is the R centre, the DFR and DBR corners, and the DR, FR and BR edges. SB is usually built as two pairs — a back pair, then a front pair. It is solved with the <R, U, M, r> move set.

# CMLL

also: Corners of the Last Layer · CxLL

CMLL (Corners of the Last Layer) is the third phase of Roux: orient and permute all four top-layer corners in a single algorithm, while ignoring the M-slice edges.

Because the M-slice is left free, CMLL has shorter algorithms than a full last-layer corner method. After CMLL only the six LSE edges remain.

See Rouxvolution — CMLL guide · Kian Mansour — CMLL ↗

# LSE — Last Six Edges

also: Last Six Edges · LSE

LSE (Last Six Edges) is the final phase of Roux: solve the remaining six edges — UF, UR, UB, UL, DF and DB — using only the M and U move set.

LSE is traditionally split into three sub-steps: EO (4a) orients all six edges, UL/UR (4b) places the left and right edges, and L4E (4c) finishes the last four M-slice edges and the centre. The combined EOLR / EOLRb approaches merge the first two into one step.

CMLL

There are different ways to solve the CMLL step.

# Two-look CMLL

also: 2-look CMLL

Two-look CMLL solves the last-layer corners in two algorithms: first orient all four corners, then permute them.

It is the usual starting point before learning the full one-look set. See CMLL.

See Rouxvolution — CMLL guide

# One-look CMLL

also: 1-look CMLL · full CMLL

One-look CMLL solves the orientation and permutation of all four last-layer corners with a single algorithm. It's faster than two-look CMLL but you have to memorise more cases.

It is what most solvers graduate to after two-look CMLL. See CMLL.

Last Six Edges sub-steps

The LSE pipeline — EO, UL/UR, L4E — and the combined EOLR variants.

# EO — Edge Orientation

also: 4a · Edge Orientation

EO (Edge Orientation, the 4a sub-step of LSE) is the step that orients all six last-layer edges so their U/D-colour sticker lies on the U or D face, using only M and U moves.

An edge is oriented when its U/D-colour sticker points up or down with the centres in their canonical position; a misoriented edge is a bad edge. Bad edges always come in even counts (0, 2, 4 or 6), so you can read the hidden DB edge from the other five. Every EO case reduces to the arrow.

See Rouxvolution — EO guide

# UL/UR

also: 4b · LR edges · L/R edges

UL/UR (the 4b sub-step of LSE) places the left and right U-layer edges while preserving edge orientation, leaving only the four M-slice edges for L4E.

See Rouxvolution — UL/UR guide

# L4E — Last Four Edges

also: 4c · L4EP · Last Four Edges · Last Four Edge Permutation

L4E (Last Four Edges, the 4c sub-step of LSE) solves the final four M-slice edges — UF, UB, DF, DB — and the M-slice centre, using the M and U2 move set.

Sometimes written L4EP (Last Four Edge Permutation). It is also called 4c, the last positional tag in the EO (4a) / UL-UR (4b) / L4E (4c) numbering.

# EOLR

also: EO + LR

EOLR is a combined LSE step that orients all six edges and moves the UL and UR edges to the bottom layer in a single algorithm.

EOLR fuses EO and UL/UR into one step. UL/UR edges are solved to the bottom layer. This makes the 4b step trivial.

See Kian Mansour — EOLR ↗ · Athefre — EOLR ↗

# EOLRb

also: EOLR-b

EOLRb is a variant of EOLR (its "b" version) that orients all six edges and fully solves the UL and UR edges at their slots in one algorithm, rather than parking them on the bottom layer.

Compared with EOLR, which parks UL and UR on the bottom layer for L4E to finish, EOLRb places them directly at the UR and UL slots.

Concepts

Roux concepts that every solver should know.

# Block (1×2×3)

also: 1x2x3 · block-building

A block in Roux is a solved 1×2×3 cuboid — two corners and three edges around a centre — that the solver builds and then preserves; the First Block and Second Block are both 1×2×3 blocks.

# Edge orientation

also: orientation

Edge orientation describes whether an edge's U/D-colour sticker points toward the U or D face (oriented) or toward a side face (misoriented); orienting all six LSE edges is the EO step.

Orientation in Roux is read relative to the centres, so an M move that shifts the centres changes which edges count as oriented. See EO and bad edge.

# Bad edge

also: misoriented edge

A bad edge is a misoriented edge during the LSE step — one whose U/D-colour sticker points toward a side face instead of up or down (when centres are oriented).

Bad edges always occur in even numbers (0, 2, 4 or 6), which lets you deduce the orientation of the hidden DB edge from the five you can see.

# The arrow

also: arrow case

The arrow is the EO case with three bad edges on top and one on the bottom; every other EO case is solved by first reducing it to the arrow.

See Rouxvolution — EO guide

# AUF

also: Adjust U Face · adjust the U face

AUF (Adjust U Face) is a single U-layer turn done before or after an algorithm to line up the case for recognition or to seat the result — it is not part of the algorithm itself.

# M-slice

also: M layer · middle slice

The M-slice is the middle layer of the cube between the L and R faces; the M move turns it in the same direction as an L turn, and Roux's last phase moves only this slice and the U face.

Roux leans on M-slice turns instead of cube rotations, which is why the method feels different from other methods.

Notation & metrics

Move letters and how solutions are counted.

# Move notation

also: notation · turn notation · U R M r

Move notation uses single letters for face and slice turns: U (top), R (right), M (middle slice, in the L direction) and r (wide R, the two right layers together); a prime (') means counter-clockwise and a 2 means a half turn.

The faces used in Roux SB and LSE are mostly U, R, M and r — L, F, B and D and their variants are rarely needed. A wide r equals R · M'.

# STM — Slice-Turn Metric

also: Slice-Turn Metric · slice turn metric

STM (Slice-Turn Metric) is a move-counting convention: every quarter or half turn of a face, the M-slice, or a wide layer counts as exactly one move.

Under STM a wide or slice move is not penalised against a face turn — so r and M each count 1. This differs from HTM (Half-Turn Metric), where a wide r would count as two moves.