How to Get Sub-20 on the 3x3
Reviewed by CuberPal Editorial Team · Updated 2026-07-07 · Get faster at 3x3 learning path
Quick answer
To average sub-20 on 3x3 you need: (1) the full cross planned during inspection, (2) F2L with no pauses between pairs — around 10–11 seconds, (3) full PLL and at least 2-look OLL, and (4) lookahead trained through slow solves. You do not need ZBLL, a new cube, or faster hands.
Sub-20 is the milestone that separates casual solvers from real speedcubers — and it is achievable by almost anyone with a plan. No advanced algorithm sets required.
Here is the exact skill checklist, in the order that gives you the most seconds per hour of practice.
The sub-20 checklist
A sub-20 average breaks down roughly like this: Cross ~2s, F2L ~10–11s, last layer ~6s. Work backwards from those numbers:
- Cross in 8 moves or fewer, fully planned during the 15-second inspection — your first pair should start with zero hesitation.
- F2L pairs solved without pausing between them. This is worth more than everything else combined.
- Full PLL (21 algorithms) so every permutation is one fast algorithm instead of two slow ones.
- 2-look OLL minimum (10 algorithms). Full OLL helps but is not required for sub-20.
- Fingertricks: algorithms executed with wrist-and-finger flow, not hand regrips.
The one skill that matters most: pause-free F2L
Film yourself solving and count the pauses during F2L. Each visible hesitation is 0.3–1 second. Six pauses is six seconds — the entire gap between a 25 and a 19.
The fix is slow solves: turn at 50–70% speed and never stop moving. If you have to pause to find the next pair, you were turning too fast. Over 2–3 weeks this rewires you to track the next pair while inserting the current one — that skill is called lookahead, and we cover it in depth in our F2L lookahead guide.
Learn full PLL before full OLL
PLL is only 21 algorithms and you use one every solve. Going from 2-look to 1-look PLL saves 1.5–3 seconds per solve — the best value in cubing. Learn T, U, and J perms first (they appear most often), then A, R, G perms, and finish with the rare cases like E and N.
Drill new algorithms daily with spaced repetition so they survive past next week. Recognition matters as much as execution: learn to identify each case from headlights and blocks in one glance.
Measure your progress by phase, not just by average
Your ao12 can plateau while your F2L is quietly improving. Track phase times so you can see which part of the checklist is done and which still bleeds time. When F2L is consistently near 10 seconds with a planned cross and full PLL, sub-20 follows automatically.
How CuberPal gets you to sub-20
CuberPal is effectively a sub-20 trainer: it measures each checklist item and tells you which one is holding you back.
- 1Record a solve and let the AI break it into Cross, F2L, OLL, and PLL timings — instantly see whether your F2L is at the 10–11s target.
- 2Check the F2L pair analysis for detected pauses and rotations; each flagged hesitation is a concrete thing to fix.
- 3Learn full PLL in the algorithm catalogue: all 21 cases with 3D previews, alternative algorithms, and mark-as-learned tracking.
- 4Drill the PLLs you fumble with daily spaced repetition sessions tuned to your weak cases.
- 5Track ao5, ao12, and ao100 on the WCA-style timer and watch the average slide under 20.
Frequently asked questions
How long does it take to get sub-20 on 3x3?▾
From a 30-second average, typically 3–6 months of consistent practice (30–60 minutes a day). The biggest time sink is building pause-free F2L; full PLL takes only a few weeks with spaced repetition.
Do I need full OLL for sub-20?▾
No. Many cubers reach sub-20 with 2-look OLL. Full OLL saves around a second and is worth learning on the way to sub-15, but pause-free F2L and full PLL matter far more at this stage.
Do I need an expensive cube to get sub-20?▾
No. Any modern magnetic speedcube from the last few years is more than capable of sub-20 — and well beyond. Skill, not hardware, is the bottleneck at this level.