See 2-look OLL: three edge shapes, then the corners
Full OLL is 57 cases, but you can orient the whole last layer in two looks with about 10 algorithms. Look 1 — make the yellow cross. Read only the four edge stickers on top; you'll see one of three shapes:
Dot — no edges oriented. Do the edge alg twice.FRUR'U'F'fRUR'U'f'L-shape — two adjacent edges. Hold them at back-left.FRUR'U'F'Line — two opposite edges. Hold the bar horizontally.fRUR'U'f'
Look 2 — orient the corners. With the cross done you're in one of 7 corner cases. Two algorithms cover all of them by repetition: SuneR U R' U R U2 R' and Anti-SuneR U2 R' U' R U' R'. Learn the 7 dedicated cases later for speed; Sune-spam works today. That's the on-ramp before the full 57.
Raw algorithm knowledge only takes you so far — these techniques, drawn from Dan Harris's Speedsolving the Cube, are what turn a correct solve into a fast one.
You do not have to learn 57 OLL algorithms before you can use full PLL. The last layer is best climbed as a ladder — each rung is usable on its own, and you upgrade over months, not in one sitting.
1
4-look (where you start)
Orient edges → orient corners → permute corners → permute edges. This is the beginner last layer — four small algorithm families, four separate looks. It works; it's just slow.
2
3-look (the smart stepping stone)
Learn the 7 corner-only OLL cases to orient the top in one look on top of the yellow cross, then do full PLL (one look). Cross → cross-edges → 7-case OLL → PLL. Harris's own recommendation: front-load PLL — 21 cases beats 57, and it's the part of the last layer most worth learning early.
3
2-look (full CFOP)
Full 57-case OLL (one look) + full 21-case PLL (one look). The expert finish: see the case, fire the algorithm, done. Beyond it lie VHF2L, COLL and ZBLL — last-layer influence you start during F2L — but those are years away and optional.
Getting faster: the honest progression
Speed is not faster fingers — it's fluency: a solve with no pauses. New solvers turn at full tilt, then freeze after every step to hunt for the next piece, and those freezes dominate the clock. The fix feels backwards: turn a little slower, slow enough that your eyes track the pieces, run the algorithm on autopilot, and use the freed attention to find the next F2L pair before your hands finish the current one. Creep the speed back up only once solving stays continuous.
The order that works: (1) bottom cross and intuitive F2L; (2) 3-look last layer; (3) drop OLL to 2-look, then learn full PLL; (4) chase fluency — muscle-memory algorithms free your eyes for look-ahead. What separates a sub-20 solver from a 40-second one is rarely more algorithms; it's an efficient, rotation-light F2L and a pause-free solve. F2L is where the time is — the last layer is just recognition speed.
