Cross, F2L, OLL, PLL. Four stages that replace the beginner method with one continuous solve. This page is the map; the individual case pages contain the algorithms, recognition notes and 3D execution.
Each stage has a precise goal. If you cannot explain the goal, more algorithms will not help.
Solve four bottom edges during inspection. The cross is a planning problem more than an execution problem; a deliberate eight-move cross beats a lucky five-move cross you cannot repeat.
Solve the four corner–edge pairs together. Learn the three easy insertions intuitively first; the 41-case library is a reference, not a list to memorise from day one.
Orient every last-layer sticker with one algorithm. Start with two-look OLL (edges, then corners) and add cases in order of how often you actually meet them.
Permute the last layer after orientation. Learn two-look PLL first — T, Y, Ua, Ub, H and Z — then replace it with full PLL when recognition is reliable.
Understand pair joining and insertion before memorising 41 cases.
Plan the cross during inspection; stop discovering it during the solve.
Ten algorithms, applied from the correct angle.
Six algorithms, used until recognition is automatic.
21 cases, learned in families, reviewed with spaced repetition.
57 cases, prioritised by frequency and recognition.
Every featured case is engine-verified and linked to a 3D animation from its inverse setup.
Cross, F2L (First Two Layers), OLL (Orientation of the Last Layer) and PLL (Permutation of the Last Layer). It is the most widely used speedcubing method.
Full PLL first for most solvers: 21 cases, frequent use, and a clearer recognition structure. Two-look OLL is enough while you build F2L speed. Learn full OLL when your F2L and PLL are already fluid.
Intuitive F2L and two-look last layer can be learned in weeks. Full PLL and full OLL take months of spaced practice. The timeline depends on deliberate practice, not on how many algorithms you bookmark.
CFOP is the most popular and has produced many world records, but Roux, ZZ and other methods are also competitive. Method choice matters less than solution efficiency, lookahead and practice quality.