This'd only be worth it if the algs proved significantly faster than what people already do for those cases. I don't know of nay low movecount skewb methods, might be cool to see one though. I guess you might look into improving Sarah's advanced with advanced 2x2 techniques (face, solve the rest with EG 1/2, TCLL variants, etc.)

Yeah, I know that anything I make has almost zero chance of surpassing normal Sarah’s Advanced because it’s so good, so I’m just doing this for fun basically.

I’ll be so surprised if people actually use whatever I come up with.
 
Last edited:
I was wondering about genning algs for last 3 edges on 5x5 . How many algs would there be, and is it worth learning them?
 
I was wondering about genning algs for last 3 edges on 5x5 . How many algs would there be, and is it worth learning them?
Try generating some of the algs and I think you'll find the answer to both of those questions very quickly.
There should be 600+ cases not counting L2E (300+ without parity, 300+ with), although a lot of them are one-move setups into each other. The parity algs will be mostly 12+ moves long, possibly significantly longer if you're looking for speed-optimised algs.
 
A new (?) Heise method variant:
1. 2x2x3 block (done by expanding 2x2x2 or 1x2x3 block)
2. Edges Orientation (the same as in the Petrus method)
3. 2x2x1 block + Edges Permutation. Can be done intuitively or with TPELL algorithms (something like LPELL, but solving 3rd slot instead of 4th, 24 cases)
4. Last 5 corners orientation - 22 cases
5. Last 5 corners permutation - 16 algorithms.

Similar to Petrus, Heise, Y-Move, and FLFL. Average movecount should be around 50 HTM, less, than in FLFL method (with less, easier to recognize cases).
 
So, earlier this week I said I was developing a low movecount Skewb method.
Well, here's the next thing I've done- Half A Skewb. No this isn't a meme, its a real method that not many people know of and made by Cube Roll here! This is a unique method that interested me.

Basically, you solve Half of a skewb minus one corner, then solve the rest in one alg.

This has a average movecount of just 14.044. Crazy low. Beating optimized Sarah's Int. by four whole moves.
Calculation is 4.800 for Half-1, 0.666 for AUF, and 8.578 for Second Half algs.

This sheet made by me has like all the info you really need: https://docs.google.com/document/d/12fa5s96lzPunGiZEVrRBX3sk-zrWO6hAfVcPG5xx-8Y/edit?usp=sharing

I don't know if I can make up anything that's lower than 14, without getting super complicated or involving 100s of algs. Maybe I can.

Is this useful? I don't know. Some people might struggle with making the first Half-1 and the algs aren't made to be fingertricky.
What do you guys think?
 
Last edited:
I want to introduce to everyone about my new method. It was created within 30 minutes.
I named it DoubleV because of step 2 and 3 I described below.
Step 1: EO+DB (EODB)
The same as ZZ but you don't need to solve DF edge.
Move: 6-8 moves
Algs: intuitive
Step 2: Make a V-block. (V-block)
Solve FB in DL and insert corner-edge at BR.
Move: 10-15
Alg: intuitive
Step 3: Make V-shape on U. (V-shape)
a) Create 1x2x2 block on UBL.

Move: 5-9
Alg: intuitive.
b) Solve FR edge.
Move: 0-4
Alg: intuitive
c) UBR and DFL corners.
You can do this step with commutator or conjugate.
Move: 8-16
Alg: intuitive
Step 4: Solve the last 4 edges and 2 corners. (L4C2E)
Move: 6-15
Alg: about 72 cases ( i will add them later)
Some of these cases can be done with the commutator. (Edges permutation or corners orientation only)

Movecount: 52-63
Number of algs: 72 or less.

Pros:
-Ergonomics - <R, L, U, r, M> for S2 and S3a+b
-All are easy intuitive steps except for some cases in step 4.

Cons:

-Lookahead is so hard.
-Some cases are difficult to recognize quickly.
-Sometimes you have to do commutator two times to solve UBR + UFL corners. This's why it has up to 16 moves. (I will try to improve this)

What do you guys think about this?
Sorry if my English was bad.
 
I don't much care for this method. It has a lot of steps to get to the L4EC state with little lookahead. Also, L4EC seems like it has really terrible recognition. Even though you only have to look at 2 faces, twisted corners makes things quite difficult to recognize. However, if you want to continue developing this, you might consider this method instead:

1)FB on L
2)Block in DBR or DFR (DBR, DB, RB or DFR equivalent) w/ centers
3a)Square in UBL (you may need to rotate so empty slot is in FR)
3b)EO+FR
3c)UFL+UBR
4)L4E+L2C
 
This is a method idea that I think has potential. I've genned a few of the algs, and they're looking pretty good. here are the steps:
2x2x3 (13, according to one of the best Petrus solvers)
Corners (13, 2 parts, set up, alg, like in waterroux)
FR+BR (9, mode of all the algs I've genned (about 90%). Also, S slice fingertricks need to be on point)
L5E (10, a combination of ELL and L5EP. Probably quite good, as all algs MU).

Advantages:
-Lots of alg steps means good tps.
-Efficient (45 moves)
-only one rotation, and it's planned and you can use it to recog.

Disadvantages:
-more alg steps means less lookahead.
-lots of algs
-2x2x3 is awkward, as are S slice fingertricks.

B' D2 R2 B2 D' L2 U B2 L2 D R2 D B F' L' R' U F' R' U2 R2
x2//Inspection
U R' D' L D' F2 D//2x2x2, 7/7
y R D R D' R2//2x2x3, 5/12
y' U2 R2 U'//corner setp up, 3/15
R U R' U R U2 R'//corners, 7/22
U' R2 U2 S R' S' R' U2 R2//FR+BR, 9/31
y U M' U2 M' U2 M U M2 U M' U'//L5E, 11/42
 
Hi. I'm Dalton Thomas. I may or may not have discovered a method that improves upon the beginner method by using less algorithms. It uses the same algorithms but only uses 5 of them. The algorithms that is uses are
F U R Ui Ri Fi
U R Ui Ri Ui Fi U F
R U Ri U R 2U Ri
U R Ui Li U Ri Ui L
R Bi Ri Ui Bi U F Ui B U R B Ri Fi

If you think that this may have an opportunity at being a new advancement. I would be glad to hear it.
 
Using less algs is not an improvement. Ok, it may be less BUT algs you solution will be longer.

More algs => More improvement. And you can solve the entire cube using only one alg and setups
 
Last edited by a moderator:
Please write in standard notation it makes it so much easier

<RLFBUD> and <MES>
R' instead of Ri (R Prime, not R Inverted)
U2 instead of 2U (2U is sometimes notation for slice moves, usually big cubes)

I'd need to see what each alg actually does in context of a solve to see if it'd be worth it over beginner LBL
 
Hey guys,

I wanted to make a suggestion for 4x4x4. My last contribution to this community was probably about 10 years ago when I developed the BH method with Chris Hardwick for solving the Rubik's cube and particularly the 4x4x4 and 5x5x5 blindfolded.

I can solve the 3x3 with Roux in about 18 seconds when I practice heavily.

I'm testing some things on 4x4 now.

My proposed method is getting fine tuned now:

1. Solve L/R Centers
2. Solve Left 1x3x4 block.
3. Solve Right 1x3x4 block.
4. Solve Last 4 Centers
5. CMLL
6. LSE (Orient, Permute L/R wings, Permute M Slice)
7. Direct Solve remaining edges.
7a. Fix Parity if Needed

What most interests me about the method or approach is that I get to mix Roux and commutators.

Next, the LSE step can permute 5 wings on average. Leaving you with 3 commutators at the end. The parity fix is faster using (r U2) (r U2 F2) (r F2) (l' U2 l U2) as opposed to an OLL parity fix.

Take a look at the LSE of Roux. If you look define the wings as we do in BH, there are 12 wings that need permutation. 4 on the l slice. 4 on the r slice. 2 on the f slice and 2 more on the b slice.
URb, UBl, ULf, UFr
DFl, DBr, FDr, BDl
FUl, RUf, BUr, and LUb

By always solving and permuting 2 L/R wings and using another one as your buffer. There are 72 possible permutations. 56 of them involve 2 l/r slice wings. So you can use commutators like x' U'RU r2 U'RU r2.

Anyway, I'll post more soon.
 
Hey guys,

I wanted to make a suggestion for 4x4x4. My last contribution to this community was probably about 10 years ago when I developed the BH method with Chris Hardwick for solving the Rubik's cube and particularly the 4x4x4 and 5x5x5 blindfolded.

I can solve the 3x3 with Roux in about 18 seconds when I practice heavily.

I'm testing some things on 4x4 now.

My proposed method is getting fine tuned now:

1. Solve L/R Centers
2. Solve Left 1x3x4 block.
3. Solve Right 1x3x4 block.
4. Solve Last 4 Centers
5. CMLL
6. LSE (Orient, Permute L/R wings, Permute M Slice)
7. Direct Solve remaining edges.
7a. Fix Parity if Needed

What most interests me about the method or approach is that I get to mix Roux and commutators.

Next, the LSE step can permute 5 wings on average. Leaving you with 3 commutators at the end. The parity fix is faster using (r U2) (r U2 F2) (r F2) (l' U2 l U2) as opposed to an OLL parity fix.

Take a look at the LSE of Roux. If you look define the wings as we do in BH, there are 12 wings that need permutation. 4 on the l slice. 4 on the r slice. 2 on the f slice and 2 more on the b slice.
URb, UBl, ULf, UFr
DFl, DBr, FDr, BDl
FUl, RUf, BUr, and LUb

By always solving and permuting 2 L/R wings and using another one as your buffer. There are 72 possible permutations. 56 of them involve 2 l/r slice wings. So you can use commutators like x' U'RU r2 U'RU r2.

Anyway, I'll post more soon.
I'm not going to lie, this looks an awful lot like stadler. If you can find a mostly intuitive approach to LSE I'll be pretty interested though.
 
Ok, so this method is terrible, but I have an idea for corners first. Solve the corners with a 2x2 method such as Ortega or CLL, solve the centres with slice moves, then fill-in 3 first layer edges with M and U moves, then remember which edge is unsolved. Then, using keyhole-like techniques, fill in all the last layer edges except the one that matches the colour of the missing edge from the first layer. Finally, you would rotate to have the unsolved edges on top and do LSE like you would with Roux.
 
Here is a concept and idea that makes CFOP easier. I called it CEFOP(Cross - Edge Control-F2L-OLL and PLL). Basically it is to introduce edge control after cross. The idea is as follows.

Basically a cube has 12 edges and a scrambled cube has an even number of good and bad edges. If you use ZZ, you need to be proficient to inspect 12 edges and turn them good using F,F' for 4 bad edges in F face and using B or B' move for 4 bad edges in B face.
If 2 bad edges, you can use FUF', FU'F', F'UF,F'U'F' , FU2'F' to turn bad edges in equator from bad to good and similar for back face by replace the F turns with B turns.

But once you done that the whole solve is 2 gen with just R,L, U and D moves. But the EO line is the quite difficult for people who are used to making cross optimally by using all moves including F and B without restriction. What if a CFOP user wants to make a solve 2 gen? It is harder to be color neutral on ZZ and sometimes it feels restrictive

My suggestion for CFOP cubers who are optimal with cross solving, color neutral and extended cross solvers is to do edge control after cross.
This can turn your whole F2L into rotation less solving and get only OLL cross cases. The reason it is easier to focus on finding and fixing bad edges if the cross which is already solved are all good edges. You end up just focussing on 8 edges. You can have zero or even number of bad edges such as 2,4,6 or 8 bad edges.

You can do edge control for just the 4 middles layer edges or all 8 edges to make a F2L solve more seamless. The result is all the difficult F2L case with flipped edges are avoided and dot OLL can be avoided.

Basically after doing the cross, if you have bad edges in UF and FR position, a sledgehammer will change them to good. A reverse sledge hammer will turn UF and UR edges to good.

Changing bad edges to good
1) Use R'FRF' to change UF and RF bad edges to good edges and
LF' L'F for bad edges in UF and LF position
2) Use rUR' URU2r' to change UF and UR bad edges to good edges
or R' (R'FRF') R to change UF and UR bad edges to good
or F (U'RUR')F'
or double sledghammer ( R'FRF') x 2
3) Use rU2R'U'RU'r' to change UF and UL bad edges to good edges
or L (LF'L'F)L' to change UF and UL bad edges to good
4) Use F( RUR'U')F' to change UF and UB bad edges to good edge
5) Use RU2R' (Sledghammer) U2(sledgehammer) to change UF, UR, UL and UB to good edge
6) Use FBUF'B' for bad edges in UF and UL and FL and BR postions etc.

I like to focus on 1-5 above on fixing bad edges and how to be fast in fixing bad edges with quick look ahead and inspection.

After solving cross, you should focus on the 3 edge pieces (FU,FR and FL) facing you. You should also track any good edge with the top color on the U face(eg. Yellow) , If they are good edges of the top color in the equator make a quick decision to do a Y or Y' turn so you have most Yellow edges that are good in equator. Track the good Yellow edges on top as well. If both edges in equator is good, do a quick d2 turn.

The aim is to quickly have 2 edges in the back of equator to be good. You can quickly use 3 move insert(such R'UR') to put a yellow edge into equator if possible. Use 1-4 method to correct remaining bad edges. Once there are no more edges to fix, the rest of the solve is simply rotationless F2L and OLL. You can either focus on just fixing middle layer edges or both.

If we are fixing only middle layer bad edges, it is important to look at just the middle layer edges on the top layer. Imagine if all the top middle layer edges is on the U layer. Assuming yellow is the top color, you will usually have blue ,green red and orange. If blue and green is dominant color, either blue or green should be the front face to have less rotation. If orange and red is the dominant color, that should be the front face. You need at most one rotation and one sledge hammer to have a rotation less F2l subsequently.

In fact all F2L can be solved with at most one rotation and no F moves to fix bad edges. if we solved by those pairs having F and B colors first and bring bad edges to the U face and rotate 90 degrees about Y axis and solve all remaining pairs. All difficult F2L cases can be avoided. The trade off is solving selected pair rather than solving any pairs you see in usual F2L which may slow you down initially.

Some explanation of the step is in the video below


Middle layer edge control


Here is variant of edge control after cross that I is useful for OH solving since it use double layer moves and avoid F and B' wherever possible.


After solving cross, inspect middle layer layer edges front and back and remember where the bad edges are. Do a u or u' rotation, this is the important setup move to fix bad edges similar to keyhole method

Note that the bad edges in equator becomes a good edge after a u or u' move. So replace it with a bad edge on the U layer via 3 moves inserts (R U'R', RU2'R', RUR' FR Slots, R' R, R'U2'R, R'U' BR Slots and similarly for the left side slots. Note all 4 slots can be use to fix bad edges. After fixing the edges undo the setup move with the opposite u' or u move. the replaced bad edges will turn good. This method is especially good to fix bad edges at BR and BL slots which is notoriously hard to solve during F2L and makes the whole solve easy.


Fixing bad edges at back slots



Do take note of special situation,

1) All middle layer edges are good, use earlier methods to fix bad edges on U layers.
2) 4 bad middle layer edges in equator, take out 2 of them before fixing
3) 2 bad edges in diagonal corner. take out one bad edge before u move or use sledgehammer to change it.
Basically you can't have more bad edges in equator than U layer before using u moves to fix it. Otherwise use earlier methods such as sledgehammer to fix u and equator bad edges simultaneously


A very useful triggger

The highlight of this method is a trigger that converts 4 bad edges to good.
It is basically
RuR'u'R' converts 4 bad edges in BR, BL UR and UB position to good
R'u'RuR converts FR, FU, UR and UF edges from bad to good.
Lu'L'u'L' coverts FR, FL, UL and UF
L'uLu'L converts BR, BL, UL and UB
 
Last edited:
Or you could just use zz with eocross. Better ergonomics, more efficient, and better than what you described. You can sledge in CFOP to help with edge control, but anything else isn't very good, including what you've described. It's too much thinking mid solve.
 
Or you could just use zz with eocross. Better ergonomics, more efficient, and better than what you described. You can sledge in CFOP to help with edge control, but anything else isn't very good, including what you've described. It's too much thinking mid solve.

I did mentioned ZZ is hard for people who are used to doing efficient cross and 12 edges is harder to inspect and even harder to achieve full color neutrality. This method is a suggestion as a mid step for intermediate CFOP users. You can do some slow solve based on what is described above especially the last part.

"If we are fixing only middle layer bad edges, it is important to look at just the middle layer edges on the top layer. Imagine if all the top middle layer edges is on the U layer. Assuming yellow is the top color, you will usually have blue ,green red and orange. If blue and green is dominant color, either blue or green should be the front face to have less rotation. If orange and red is the dominant color, that should be the front face. You need at most one rotation and one sledge hammer to have a rotation less F2l subsequently."
 
You might as well do conjugated ollcp and L5E, which is kinda zipper, and faster, and it's the inverse of what you said btw. So F2L-1, corners, edges instead of edges, corners.
 
I did mentioned ZZ is hard for people who are used to doing efficient cross and 12 edges is harder to inspect and even harder to achieve full color neutrality. This method is a suggestion as a mid step for intermediate CFOP users. You can do some slow solve based on what is described above especially the last part.

"If we are fixing only middle layer bad edges, it is important to look at just the middle layer edges on the top layer. Imagine if all the top middle layer edges is on the U layer. Assuming yellow is the top color, you will usually have blue ,green red and orange. If blue and green is dominant color, either blue or green should be the front face to have less rotation. If orange and red is the dominant color, that should be the front face. You need at most one rotation and one sledge hammer to have a rotation less F2l subsequently."
Interesting, but it loses too much in efficiency and having to check EO in the middle of a solve. It seems like most of the time you'll spend three or more moves to fix only two bad edges, whereas the EO step in beginner's ZZ fixes up to four bad edges in one move. I did ten solves each with EO-then-cross (x2 neutral) and cross-then-EO (full CN) to compare the move count, and the former is a clear winner (9.9 moves versus 12.5 moves). I'm not particularly efficient at the cross, so the difference might be smaller for better CFOPers. However, determining EO during inspection takes just a bit of practice, and I actually think that planning EO (as in, EO by itself, not EOline) is a lot easier than planning cross.

I think the best "advanced" CFOP variant is to influence edge orientation throughout F2L by using alternative inserts (sledgehammer, F R U/U' R' F'). Doing EO by itself is definitely a waste of moves; doing EO while solving F2L is fine.
 
Back
Top