CFOP stage 3
OLL makes every last-layer sticker face upward without requiring the pieces to be in their final positions. Full OLL has 57 cases, but a smaller two-look path solves every orientation first.
Direct answer
Begin with 2-look OLL so every last layer can be oriented with a small set. Learn Sune and Anti-Sune early, then add full OLL cases in recognizable shape families. For each case, learn multiple viewing angles, a consistent starting grip and a setup move for isolated execution practice.
Complete guide
The top-face shape narrows the case, but side sticker bars and headlights distinguish cases that share a similar outline. Good recognition works from front, side and back angles without rotating the cube first.
A case is learned when you can recognize it, begin from a stable grip, execute accurately and predict the resulting PLL view. Merely reciting the moves is not enough.
Gain complete coverage with 2-look OLL, then group full cases by shape: Sune families, T shapes, squares, lightning, P shapes, C shapes and awkward cases. Add a few per week and keep older cases in review.
Use setup moves to create one case repeatedly, switch the viewer angle between attempts and time recognition separately from execution. Mix the case into random OLL training only after isolated accuracy is high.
Track recognition time, execution time and failure rate per case. Community averages can give context, but compare your own trend and remove lockups before chasing maximum turn speed.
In full solves, watch the pause after last-pair insertion. If execution is fast but the pause is long, the next gain comes from transition recognition rather than a different algorithm.
Do not learn only the top pattern, use an algorithm with a hidden regrip you cannot reproduce, or add too many cases before spaced review. AUF habits should be separated from the OLL itself.
Start with the Sune, Anti-Sune and T-shape pages, then use the OLL hub to compare recognition families and recommended order.
Quick answers
Full OLL has 57 cases; a common two-look approach uses about 10 algorithms for complete coverage.
Many solvers learn full PLL first because it has fewer cases and is easier to complete.
Side stickers and corner orientation can distinguish cases with similar top-face shapes.
Sune, Anti-Sune and common simple shapes are strong early additions after two-look coverage.
Practice with feedback
SupaKewber connects timing, smart-cube telemetry, training and online competition in one browser workflow.
Start OLL training