Lookahead
Knowing where the next pair is while your hands are still on this one — the largest saving left after F2L, and the one nobody can hand you.
By the end you will be able to
- Work out how much of your solve is turning and how much is looking
- Practise deliberately slowly without it feeling like a waste of time
- Track a piece through turns you are making with your hands
- Recognise which pauses are a lookahead problem and which are a recognition problem
Two people can turn at the same speed and be twenty seconds apart. The difference is not in their hands. One of them knows what they are doing next while the current pair is going in; the other stops, looks, finds, and starts again, four times a solve.
The arithmetic
Take a seventy-turn solve at four turns a second. That is under twenty seconds of turning. If your average is forty seconds, then twenty seconds of every solve is your hands sitting still. No amount of finger training touches that half.
Turning slower to go faster
The drill is uncomfortable and it works. Solve at about two turns a second — slow enough to feel silly — with one rule: you may not pause. Not between pairs, not before the last layer. If you have to stop, you were going too fast; go slower.
Your first slow solves will be no faster than your normal ones, which is the point being made rather clearly. After a fortnight of ten a day they will be faster, and then your normal solves will follow.
Tracking a piece
While your hands do the current insertion, your eyes should be somewhere else. Pick one piece — the corner of the next pair, not both pieces — and follow it through the turns you are making. It will move; you know exactly where it goes, because you are the one turning the face it is on.
- Start by tracking a piece through a single insertion. Three turns, one piece.
- Then track it through a whole pair, including the setup turns.
- Then track both pieces of the next pair.
- Then stop tracking and notice you can already see where they are.
Where the pauses are
| You stop here | It is really this | What fixes it |
|---|---|---|
| Before the first turn | The cross was not planned | Inspection drills — cross, then cross plus one pair |
| Between F2L pairs | Lookahead | Slow solves with no pauses |
| After the last pair | OLL recognition | Drill the ten cases until the shape names itself |
| After OLL | PLL recognition | Learn to read headlights before the algorithm ends |
| Halfway through a pair | You had no plan when you started it | Look before the first turn, not during the third |
The plateau
Nearly everyone stalls at around the same place: F2L learnt, two-look last layer learnt, times sitting between forty seconds and a minute and refusing to move for weeks. It is not a sign that you need more algorithms. It is a sign that the algorithms are ahead of the eyes, which is the normal order for it to happen in.
The way through is boring and reliable: slow solves, no pauses, every day, for a fortnight. Nothing new to learn. It is the least glamorous advice in cubing and the most consistently effective.
Practice notes and drills Short, specific things to try, sorted by what you are working on. Averages, not personal bests An average of twelve tells you whether the practice is working. A single fast solve tells you nothing.