A PLL skip is when you finish OLL and the whole cube is already solved, so you don't need any PLL algorithm. An OLL skip is the same thing one step earlier: you finish F2L and the yellow face is already all yellow. This post explains both kinds of skip, why they are rare, how to work out the odds yourself, and a few ways to get them more often.
This post is for anyone who can already solve with CFOP, even with 2-look OLL and 2-look PLL. Every algorithm here was tested on the EasyRubic cube simulator.
What is a skip?
CFOP has four steps: the cross, F2L, OLL and PLL. A "skip" means a step is already done on its own, without you doing anything.
- OLL skip: after F2L, the top face is already all yellow. You go straight to PLL.
- PLL skip: after OLL, every top-layer piece is already in the right place. At most one U turn and the cube is solved.
- LL skip (skipping the whole last layer): once F2L is done, the whole cube is done.
The picture below is every CFOP solver's dream: the last F2L pair is waiting in the top layer, so the top face still has three stickers that aren't yellow. Put the pair into its slot with R U R', and the cube is solved.
- You start with (input)
- After the formula
What are the odds of a skip?
According to Wikipedia's CFOP article, if you do nothing to "force" one, a PLL skip comes about once in 72 solves, an OLL skip once in 216, and a skip of the whole last layer about once in 15,552 solves (Wikipedia).
| Type of skip | Chance | What that means |
|---|---|---|
| PLL skip | 1/72 | about 1.4% of solves |
| OLL skip | 1/216 | about 0.46% |
| LL skip | 1/15,552 | very rare |
If you do 50 solves a day, on average you'll get a PLL skip about every day and a half, and an OLL skip only about once every four days.
Why 72 and 216?
You can count these two numbers yourself, no need to memorise them.
OLL: 216 cases. There are four edges on the top layer, and each one can be in two ways: yellow facing up or not. But the fourth edge is always decided by the other three, because the number of flipped edges is always even. So there are 2 x 2 x 2 = 8 edge patterns. The four corners can each be twisted three ways, and again the last one is decided by the other three: 3 x 3 x 3 = 27 patterns. That makes 8 x 27 = 216 patterns in total, and only one of them has the yellow face already done.
PLL: 72 cases. The four corners can be arranged in 24 ways, and so can the four edges. But only half of those combinations can happen on a real cube, which leaves 24 x 24 / 2 = 288. Four arrangements that differ only by a U turn count as one, because a U turn doesn't count as a PLL step. That leaves 288 / 4 = 72, and one of them is already solved.
LL skip: you need both at the same time, so it's 216 x 72 = 15,552.
Skips with 2-look
If you still solve the last layer with 2-look, each small step has its own chance of a skip, and the chances are much higher:
- Yellow cross step of OLL: 1/8 of the time, all four edges are already yellow (the 8 edge patterns from above).
- Yellow corners step of OLL: once you have the cross, 1/27 of the time all four corners are already yellow.
- Corner swap step of PLL: 1/6 of the time the corners are already right (headlights on all four sides).
- Edge swap step of PLL: 1/12 of the time the edges are already right.
You'll find the odds for each 2-look PLL case in the 2-look PLL post.
How to get more skips
Skips are usually luck, but fast solvers don't just wait for luck. They choose how to solve the last F2L pair, or pick up extra algorithms, so the top layer ends up in fewer possible cases. Fewer cases means a higher chance of a skip.
Edge control: orienting edges as you insert the last pair
SpeedSolving Wiki calls this technique partial edge control: when you insert the last F2L pair, you choose the insert that leaves every top-layer edge with yellow facing up (SpeedSolving Wiki). That page uses the very example below.
Starting from a solved cube (yellow on top, green in front), do F R' F' R to set up a last pair case. There are two ways to put the pair into its slot:
- U R U' R': the pair goes into the right slot, but two top-layer edges are flipped. You still need to do OLL.
- R' F R F' (the sledgehammer): the pair goes into the right slot, and all four edges are yellow. In this example, the cube is solved straight away.
- You start with (input)
- After the formula
When all four edges are yellow after F2L, there are only 27 corner patterns left instead of 216 OLL cases. The chance of an OLL skip goes up from 1/216 to 1/27, which is 8 times more. You also avoid the OLL cases with only a single yellow dot in the middle, which are usually hard and long.
How to practise: whenever you see a last pair you could insert with the sledgehammer, look at the yellow edges on top before you insert, then compare the two inserts on a real cube. You don't need to get them all at once; each case you get the hang of is one more hard OLL you get to avoid.
Forcing a PLL skip with a different OLL algorithm
Wikipedia describes how to "force" a skip: spot the OLL case, then choose a different OLL algorithm that also solves it but solves PLL at the same time (Wikipedia).
The easiest example to see is the Sune, R U R' U R U2 R'. The picture below shows a Sune case where the top-layer pieces are already in the right places, so once you finish the Sune the cube is solved: a PLL skip. If you look at the colours on the side faces before doing OLL, over time you'll start to notice moments like this.
See it in 3D, before and after
- You start with (input)
- After the formula
Fast solvers go further by knowing several algorithms for the same OLL case, each one suited to a different arrangement of the pieces. That's only worth doing once you know all 57 OLL algorithms and 21 PLL algorithms by heart.
Advanced algorithm sets
The three algorithm sets below all need the top-layer edges to be yellow after F2L, so they usually go together with edge control or the ZZ method:
- Winter Variation: according to SpeedSolving Wiki, this is 27 algorithms put forward by Lucas Winter in 2005. When the last pair is already paired up in the top layer, you insert the pair and twist the corners at the same time, so you get an OLL skip every time (SpeedSolving Wiki). After that, a PLL skip is still 1/72. See also the Winter Variation post.
- COLL: 42 cases counting symmetry, which solve the top-layer corners while keeping the yellow edges as they are (SpeedSolving Wiki). After COLL only the edges need swapping, with 12 possible arrangements, so a PLL skip is 1/12 instead of 1/72. See the COLL post.
- ZBLL: solves the whole last layer in one algorithm when the edges are already yellow, but it takes hundreds of algorithms. The What is ZBLL post has the details.
Should you chase skips?
If you're a beginner, no. Skips are too rare to change your average: one PLL skip in 72 solves only saves a few seconds on that one solve. What makes you faster every time is smoother F2L and spotting OLL and PLL cases more quickly.
A sensible order:
- Get the hang of all of 2-look OLL and 2-look PLL.
- Move on to full PLL, then full OLL.
- Practise edge control with a few common last pair cases, like the sledgehammer example above.
- Once you're fast and want to keep going, look at Winter Variation or COLL.
When you get a skip during a timed solve, enjoy it, but don't treat that solve as a measure of your level. An average of 5 or 12 solves is what tells you where you really are; the post how to work out Ao5 and Ao12 explains how. Practise and time yourself with the 3x3 timer.
With the EasyRubic app, you can scan your real cube with the camera and follow the solution step by step on a 3D cube. Get EasyRubic to practise the last F2L pair on your own cube.
Video
Want to see it in motion? The short video below from the Brian Sun channel on YouTube shows more of what this post covers. The video belongs to that channel and is embedded straight from YouTube.
Sources
Frequently asked questions
What is a PLL skip?
It's when you finish OLL and every top-layer piece is already in the right place, perhaps needing just one U turn. You don't have to do any PLL algorithm.
What are the odds of a PLL skip?
About 1/72 if you don't force it, according to Wikipedia. The number comes from the 72 ways the top layer can be arranged after OLL, not counting U turns.
Is an OLL skip rarer than a PLL skip?
Yes. An OLL skip is 1/216, three times rarer than a PLL skip. Skipping the whole last layer is about 1/15,552.
How can I get more OLL skips?
Practise edge control when you insert the last F2L pair, so all the top-layer edges are yellow. Then an OLL skip is 1/27 instead of 1/216. Winter Variation gives you an OLL skip every time, but you need 27 algorithms for it.