OLL library
Start with 2-Look OLL (yellow cross, then yellow corners), then drill all 57 one-look OLL cases. Filter by shape family, read one recognition tip per case, and demo any algorithm in Cube Lab.
Before you start
This page is the OLL part of CFOP. Finish F2L first so the bottom two layers are solved and white stays on the bottom.
2-Look OLL does not try to solve every last-layer pattern in one algorithm. Instead, you do edges first, then corners. Notation refresher: cube notation.
Demo plays the case, then the fix. Cube Lab resets, applies the inverse instantly so you see the pattern, then animates the algorithm.
Open 2-Look OLL in Cube LabPractice random OLL cases in Cube Lab
Why 2-Look then full OLL
Full OLL is 57 algorithms — one per yellow orientation. 2-Look OLL covers every scramble with a small set: make the yellow cross, then orient corners. You trade one extra look for a list you can actually finish learning.
Stay on 2-look until it is automatic. Full OLL is worth it when last-layer recognition — not F2L — is what slows you down.
1. Yellow cross
Read only three edge patterns: dot, L-shape, and line. Your goal is the yellow cross. Side colours do not matter yet.
- Dot: no oriented top edges yet. Use the same edge idea twice (dot → L or line, then finish).
- L-shape: hold the L with its arms pointing up and left.
- Line: hold the yellow line horizontally.
The core edge trigger is short and worth memorising early.
2. Yellow corners
Once the cross is done, orient the four top corners until the entire top face turns yellow. The seven recognition patterns are Sune, Antisune, H, Pi, U, T, and L.
If you are new, learn Sune and Antisune first. Then add the more visual cases: Headlights (U), Chameleon (T), and Bowtie (L).
Do not worry if the cube looks worse halfway through the algorithm. The first two layers often appear broken before they come back at the end.
Full OLL library (57 cases)
After 2-Look OLL, speedcubers learn one algorithm per last-layer orientation pattern. The library below lists all 57 standard cases, grouped by the yellow shape on top: dot, line, L-shape, T-shape, and other.
Learning order for the full set: filter by the shape you recognise from 2-look first (dot / line / L), then expand into T and other families. Each card shows one primary algorithm and a short recognition tip.
Practice path
- Drill yellow-cross shapes until you never pause on hold (L up-left, line horizontal).
- Add Sune and Antisune; do ten random corner cases each in Cube Lab.
- Solve full CFOP with 2-look OLL only — no peeking at the 57 list.
- When 2-look is automatic, pick one shape family from the library and add 3–5 cases a week.
Stuck?
If the yellow cross will not appear, check the hold: L arms up and left, line horizontal. Wrong hold turns the same alg into a different case.
If F2L stays broken after OLL, you stopped mid-algorithm or rotated the cube. Undo to white-down after F2L and retry.
Need the short CFOP recipe? How to solve a 3×3 with CFOP. Still learning pairs? F2L library.
After this page
Open the PLL library and learn 2-look PLL next. Full 57 OLL can wait until 4-look last layer feels easy. Return to the CFOP overview for the whole flow.
FAQ
Do I need all 57 OLL cases?
Not at first. 2-Look OLL covers every scramble with a handful of algorithms. Add full OLL when 2-look feels automatic and you want fewer looks on last layer.
In what order should I learn OLL?
Dot, L-shape, and line for the yellow cross; then Sune and Antisune for corners; then the other 2-look corner shapes; then full 57 by yellow shape family.
Why does the first two layers look broken mid-algorithm?
OLL algorithms open F2L on purpose and close it at the end. Finish the whole sequence before judging the cube.
When is full OLL worth it?
When your F2L is steady and 2-look OLL is no longer the bottleneck. Many cubers sit on 2-look for a long time and still improve with better F2L and PLL.
Should I learn OLL or PLL next?
Stay in order: finish 2-look OLL, then 2-look PLL. Full OLL and full PLL come after 4-look last layer is comfortable.