CFOP F2L Tutorial

Insert directly once aligned, without first splitting or extracting.

Shares the "D to a side · edge adjacent · corner-side sticker up" branch with Split Pair by Going Over: an edge position on the opposite side is Basic Inserts, on the corner side is Split Pair by Going Over.

Keep the corner's favourable orientation, move the edge into a pairable relationship first, then insert.

Distinguish from Reposition Edge and Flip Corner by the up sticker: a corner-side sticker up only moves the edge, an other-side sticker up also requires flipping the corner.

Change the corner orientation while repositioning the edge, reducing to a pairable relationship.

When the edge is far and the other-side sticker is up, an edge position on the opposite side is Basic Inserts, on the corner side requires flipping the corner; distinguish from Reposition Edge by whether the up sticker is the corner-side or the other-side sticker.

The two pieces are adjacent but the relationship is wrong; split them by going over the top layer first, then pair correctly.

When D is up and the edge is adjacent it sits next to Weird (opposite side / same side); when D is to a side, the edge is adjacent and the corner-side sticker is up, it sits next to Basic Inserts (edge position on the corner side / opposite side).

Use the side to build the correct Pair, then send it into the target Slot.

Reduce to a familiar Family first, or use a dedicated solution; do not force a conventional adjacency.

Shares the "D up · edge adjacent" branch with Split Pair by Going Over: an up sticker on the same side as the edge position is Weird, on the opposite side is Split Pair by Going Over.

Temporarily move away or make use of the corner, bring the edge into the pairing process, and restore and solve the Slot at the end.

Temporarily move away or make use of the edge, bring the corner into the pairing process, and restore and solve the Slot at the end.

If Solved, stop; otherwise both pieces are already "stuck" in the Slot — first extract one piece to the U layer, then reduce and pair via the corner-in-place / edge-in-place idea, and finally insert and restore. Use an alternative solution only when its preconditions hold.

Both pieces in their target positions does not mean solved: first check F2L 37 (stop if solved), then whether orientation and relative relationship match the Slot.

Definition — During inspection — after the Cross is planned but before you start — use the fixed lookup method to fix at least the first target Pair and Slot, and pre-pick a clear, preservable solution; once fluent, extend it to the second Pair.

Benefit — You start with a target already in mind, removing the "stop to find" pause and leaving a continuous field of view for lookahead.

Practice — In inspection, first locate a target corner, read its two side stickers to pin the Slot, then find the edge carrying the same side-sticker pair; aim to choose the first Pair's solution before the timer starts.

When to learn — You can already execute the basic relationships across all four Slots on a random case.

Risk — Inspection time is limited; the goal is "a clear plan for the first Pair", not planning all four. When the prediction is uncertain, prefer a clear, stable solution over the theoretical optimum.

Unlocks — X-cross

Definition — When a Pair can be handled in the current viewing direction, do not habitually rotate just to bring the Slot to F-R; gradually migrate to side and back-slot execution.

Benefit — Fewer regrips and observation breaks, keeps unsolved pieces visible, and leaves a stable field of view for lookahead.

Practice — Place the same topology randomly across the four Slots; before each rotation, state the specific problem it solves; when a sensible solution exists for the current view, do not force F-R execution.

When to learn — You can already transfer the core relationships to a random Slot.

Risk — Zero rotation is not the goal; an awkward move, a poor fingertrick or a blocked view can be slower than one sensible rotation. The standard observation frame (F-R) is only a recognition coordinate, not a mandatory starting move.

Unlocks — Back-slot

Definition — Once the current solution is decided, the hand keeps executing the current move while the eyes start locating the next target corner or the next Pair.

Benefit — Fewer pauses between Pairs, turning four separate searches into a continuous flow.

Practice — Lower your turn speed, prioritising continuity; once the solution is decided, let your eyes leave the current Pair.

When to learn — Family/solution decisions and the current execution no longer need step-by-step watching.

Risk — Raising TPS is not lookahead; if execution is not yet automatic, forcing your eyes away harms both the current Pair and the next recognition.

Unlocks — Multislotting

Definition — Pre-pick a clear, preservable efficient solution for the Cases where the intuitive path is long or error-prone, rather than re-deriving it every time. Algorithmic does not mean memorising all 41 as formulas — most stay intuitive; the point is to fill in the ones you handle inefficiently.

Benefit — Fewer pauses and misjudgements per Case, giving lookahead and the advanced techniques a stable execution foundation.

Practice — Identify weak Cases from a random pool and note a clear, fingertrick-friendly solution for each; you need not chase the theoretical minimum.

When to learn — You can already handle all 41 Cases intuitively and want to cut pauses on specific Cases.

Risk — You need not memorise every formula; prioritise filling in the Cases where intuition is inefficient, and keep the solutions transferable.

Unlocks — the execution foundation for all advanced techniques

Definition — Treat an unsolved Slot as temporary working space, used to extract, hide, reposition or build a Pair.

Benefit — Less needless protection and detouring, and a foundation for Keyhole, Multislotting and Pseudo-slotting.

Practice — Before executing, point out which Slots are open; after executing, check the solution-boundary invariant of the Cross, prior Slots and the target Slot.

When to learn — You can accurately distinguish open Slots from solved Slots and reliably restore temporarily moved structure.

Risk — Never treat a solved Slot as a workspace; "may be disturbed mid-way" does not mean "no need to restore at the end".

Unlocks — Keyhole, Multislotting, Pseudo-slotting

Definition — When several Pairs are visible at once, choose the order by solution clarity, rotation need, visibility, open Slots and the transition to the next Pair, rather than a fixed Slot order.

Benefit — Fewer rotations and re-searches, keep the easier-to-track next Pair, and avoid breaking another good Pair.

Practice — Before starting, compare at least two candidates' field of view, solution and Slot impact, then pick the stable option.

When to learn — You can handle random Slots and identify candidate Pairs without a Family filter.

Risk — Do not pause too long hunting the theoretically optimal Pair; when the prediction is uncertain, a stable solution beats the theoretically shortest solution; during practice you still need to cover weak spots.

Unlocks — Multislotting, Last-slot influence

Below is the advanced-techniques landscape — methods beyond intuitive F2L. Each answers "what it is, what problem it solves, when not to use it, and what it depends on"; no algorithm tables are provided. Most learners need not go deep into these before the fluency core is solid.

Back-slot

Definition — Insert a Pair into a back Slot without rotating the cube to a new front.

Benefit — Fewer rotations, keeping the front unsolved pieces visible.

When to learn — After the basic relationships can transfer across the four Slots.

Risk — A back Slot is hard to view directly; do not force a poor move or blindly damage a prior Slot for the sake of rotationless execution.

Prerequisite — Slot transfer, rotation-reduction awareness, boundary invariant.

Keyhole

Definition — Treat an unsolved Slot as a working hole, often placing target pieces in separately via a temporary D-layer alignment, then restoring the D-layer relationship.

Benefit — On a suitable open Slot, you can avoid first building a conventional Paired Pair.

When to learn — You can track D-layer changes and reliably judge which Slots remain unsolved.

Risk — Highly context-dependent, not a universal solution for every Case; skipping the restore breaks Cross side alignment or another Slot.

Prerequisite — Empty-Slot use, D-layer tracking, solution applicability.

Multislotting

Definition — When choosing the current Pair's solution, also consider where it sends another set of F2L pieces; a stronger form can advance or solve two Pairs in one continuous plan.

Benefit — Improves the next Pair's state, reducing re-search and extra starting move.

When to learn — You already have stable lookahead and can track the current solution's effect on other pieces.

Risk — Observation load rises; failing the current Pair also breaks the next Pair's plan.

Prerequisite — Lookahead, Pair-order, Empty-Slot use.

Pseudo-slotting

Definition — In a context with at least two usable Slots, temporarily offset the D layer so a corner and edge belonging to different Slots form a pseudo pair; after insertion, restore the D layer so both Slots land at once.

Benefit — Certain last-two-Slot cases can merge two sets of work.

When to learn — You have mastered Keyhole and multislotting and can track the ownership of two Slots.

Risk — A pseudo pair is not a real Pair of the same Slot; a wrong restore swaps or breaks Slots.

Prerequisite — Keyhole + multislotting.

Last-slot influence

Definition — When solving the fourth Pair, among several legal solutions choose one that improves the subsequent top-layer state.

Benefit — May yield an OLL that is easier to recognise/execute, and improves the F2L→OLL transition.

When to learn — F2L is stable and you can recognise basic OLL features.

Risk — Do not significantly increase F2L pauses to chase a particular OLL; it is an umbrella concept, not a single algorithm set.

Prerequisite — Stable last slot, solution choice, OLL recognition.

VHLS

Definition — A restricted subset of ZBLS; when the last Pair is already paired or close to directly insertable, it solves the last Slot while completing top-layer edge orientation.

Benefit — On entering OLL, all top-layer edges are already oriented.

When to learn — First understand last-slot influence, and prefer to start from the restricted entry.

Risk — Adds last-slot recognition and algorithm load; it should not replace ordinary F2L that is not yet stable.

Prerequisite — Last-slot influence.

ZBLS

Definition — Under a more general last-slot state, solves the last F2L Pair while completing top-layer edge orientation.

Benefit — Merges last slot with LL edge orientation.

When to learn — VHLS and ordinary last-slot execution are stable, and the learner is willing to take on a larger recognition/algorithm load.

Risk — This is a large algorithm and recognition system; this page provides no Case count, algorithm table or training course.

Prerequisite — VHLS.

X-cross

Definition — Complete the first F2L Pair during the Cross stage — Cross + 1.

Benefit — On formally entering F2L, one Slot is already done, possibly improving the first-Pair transition.

When to learn — You can stably plan the Cross during inspection and track the first Pair.

Risk — Forcing an X-cross can make the Cross longer, overload inspection or destabilise execution; without a clear opportunity, an ordinary Cross is better.

Prerequisite — Cross planning + F2L Pair tracking; not part of the ZBLS dependency chain.

• Treating zero rotation as the goal: a sensible rotation beats a poor fingertrick, a blocked view or an unstable rotationless solution.
• Treating lookahead as raising TPS: first let the current solution execute automatically and lower your turn speed, then move your eyes; a fast hand does not mean seeing the next Pair.
• Treating a solved Slot as empty workspace: Keyhole/multislotting and the like must satisfy their context preconditions; an advanced solution is not a universal solution.
• Hunting the theoretically optimal Pair until analysis paralysis: when the prediction is uncertain, a clear, stable, preservable solution takes priority.
• Stacking advanced tricks before core F2L is automatic: advanced terms cannot replace Family recognition, random-Slot execution and boundary restoration.