I invented my own rubik'a cube method

[Discomantis]

I wanted to make my own rubik's cube method, and I think I did it! I know this method is very bad, however I don't really care about speed for this method.

If this is a re-invention of a method. please let me know as I don't want to steal someone else's work.

I also have a video tutorial I made for it here:

Step 1: Solve edge orientation with yellow on top



Step two: Solve E layer



Step 3: Turn the top such that a white corner is on the top-right-front (If this is not possible you have skipped this step and step 4)



Step 4: Turn the bottom such that yellow corner is on the bottom-front-right spot



Step 5: Perform this algorithm: R2 U R2 U R2 U2 R2



Step 6: Repeat Step 3, 4, and 5 until you can no longer perform step 3. Then move on to step 7.



Step 7: Turn the top such that a white edge is on the top-front spot



Step 8: Turn the bottom such that a yellow edge is on the bottom-front spot



Step 9: Perform this algorithm: M' U2 M



Step 10: Perform step 7, 8, and 9 until step 7 is no longer possible. Then move on to step 11.



Step 11: Turn the top such that either 2 or 4 edges are solved on the top . If you solve 4 edges, skip to step 13. If you have already done step 13, skip to step 14.



If the two edges that are solved are not adjacent to each other, then do y rotations until they form a vertical line and perform this algorithm: R U R' U R U2 R'. Then perform the adjacent case. (The adjacent case is right below)



If the two edges that are solved are adjacent, do y rotations until one edge is in the back and the other is on the right. Then perform this algorithm: R U R' U R U2 R' U



Step 13: Do an x2 rotation. Then repeat step 11, and 12.



Step 14: Check how many corners are per-mutated correctly on top.



If none are per-mutated correctly, do this algorithm: R' F R' B2 R F' R' B2 R2. Then check how many corners are permuted correctly. If the answer is still 0, do the algorithm again. If not, do either the one correct case or the 2 correct case. (Cases shown below.)





If one is per-mutated correctly, do a y rotation until you have a corner per-mutated correctly on the top front left. Then do this algorithm: R' F R' B2 R F' R' B2 R2. If you still have one corner per-mutated correctly after that, then do the algorithm again. Then go to step 15. If you have already gone to step 15, go to step 16.



If two corners are permuted correctly and are a diagonal swap, do a y rotation until one of the correct corners is on the front-left-top and then do this algorithm:

R L M2 U' M2 U2 M2 U' M2 U2 L' R'

If two corners permuted coreectly and are an adjacent swap, do this algorithm: x' R' U' R' D R U R' D' R U R' D R U' R' D' R2

If you have four corners per-mutated correctly, skip to step 15. If you have already gone to step 15, go to step 16.



Step 15. Do an x2 rotation. Then repeat step 14.



Step 16: Locate two unsolved corners. Rotate the cube such that one of the unsolved corners is on the top layer and the other is on the bottom.



Then, turn the top layer such that the unsolved corner is on the top-front-right.



Then, turn the bottom layer such that the other unsolved corner is on the bottom-right-front



Then, do this algorithm: R F2 R' F' R F' R' L' F2 L F L' F L







Then, undo the moves you did to the top and bottom layer



There are 3 possibilities for what could happen next:



It solves one corner



It solves both corners



It solves no corners







If it solves one or two corners, then repeat step 16. If it solves no corners, repeat the algorithm once. Then repeat step 16.







Repeat step 16 until you solve the cube!


Edited because wrong copy and paste

 

[UnknownCorrespondent]

Sorry for the late reply - I only have a few hours access a day, and it took until now to get my account set up (12 hours for email responses to be received).

This is an interesting approach. Belt methods make for a fun challenge. I did mine while learning beginner's CFOP, so I used keyholing/F2L to separate the corners, [M2 U2]*2 and [R2 U2]*3 to separate the edges (yours is much simpler), then solved each layer with 4LLL plus whatever I knew to fix layer parity.

There are some errors in the parity algorithms you use in step 14. The first one is very clever, but is missing the last two moves (L' R'). The second one got garbled and doesn't do an E perm like you intended. which I assume is something like whatever the transposition of z x' (R U' R') D (R U R') D' (R U R') D (R U' R') D' x z’ to remove the cube rotations would be.

You could simplify the process by using the modified H perm for both, if you are willing to give up solving the top layer and fixing parity simultaneously. In fact, you can simplify everything by reversing the order your solve. Corners are harder to handle than edges, and the beginning of the solve is always easier than the end. Also, it is better to orient first, because permutation recognition then becomes trivial.

Putting all that together, my suggestion would be:

Steps 1-10 as written.

Step 11: Do Step 16.
Advance: Note that the alg twists URF CW and DRF CCW. Whenever possible, arrange the spots so that the top and bottom colors are both on F, which will solve both.
Alternate: Put unsolved corners at ULF (CCW, top color on F) and DRF (CW, bottom color on R) and do [(U R’) B2 (R U’) F2]*2. This is slightly shorter, and considerably easier to learn.

Step 12: Look at the side spots of the top layer corners.

If each pair of sides are the same color (headlights), then turn U to match the centers, turn the cube over and apply step 12 to the other side.

If each side has opposite colors, do the step 14 alg (A perm) once, which will give you the case directly below.

If there is one set of opposite colors, hold them on the top front and do the A perm, then adjust the solved corners to match the centers.

Step 13: Look at the side spots of the top layer edges. If one is solved, or if there are two swapped pairs, skip to the next paragraph. If 2 are solved, or if all four are in any pattern other than swapped pairs, do [M² U²]*2, which will give you one of the above patterns.

If you have two swapped pairs, do R² U (F B') R² (F' B) U R² (U perm), which will leave you with one edge solved.

If you have one edge solved, hold it on the left and do the U perm. You may have to do it twice.

When the top is solved, turn the cube over and solve the other side.

Advance: When you are holding the solved edge on the left, look at the front spot. If it is the opposite of the center, then the U perm will solve it. If it isn’t the opposite, do the U perm, except with U’s instead of Us.

You use most of the same algs, only adding a U perm and the trivial double pair swapper, while losing the H and E perms, but there should be fewer repetitions of each alg, especially if you pay attention to what each alg is doing.

 

[Discomantis]

Sorry for the late reply - I only have a few hours access a day, and it took until now to get my account set up (12 hours for email responses to be received).

This is an interesting approach. Belt methods make for a fun challenge. I did mine while learning beginner's CFOP, so I used keyholing/F2L to separate the corners, [M2 U2]*2 and [R2 U2]*3 to separate the edges (yours is much simpler), then solved each layer with 4LLL plus whatever I knew to fix layer parity.

There are some errors in the parity algorithms you use in step 14. The first one is very clever, but is missing the last two moves (L' R'). The second one got garbled and doesn't do an E perm like you intended. which I assume is something like whatever the transposition of z x' (R U' R') D (R U R') D' (R U R') D (R U' R') D' x z’ to remove the cube rotations would be.

You could simplify the process by using the modified H perm for both, if you are willing to give up solving the top layer and fixing parity simultaneously. In fact, you can simplify everything by reversing the order your solve. Corners are harder to handle than edges, and the beginning of the solve is always easier than the end. Also, it is better to orient first, because permutation recognition then becomes trivial.

Putting all that together, my suggestion would be:

Steps 1-10 as written.

Step 11: Do Step 16.
Advance: Note that the alg twists URF CW and DRF CCW. Whenever possible, arrange the spots so that the top and bottom colors are both on F, which will solve both.
Alternate: Put unsolved corners at ULF (CCW, top color on F) and DRF (CW, bottom color on R) and do [(U R’) B2 (R U’) F2]*2. This is slightly shorter, and considerably easier to learn.

Step 12: Look at the side spots of the top layer corners.

If each pair of sides are the same color (headlights), then turn U to match the centers, turn the cube over and apply step 12 to the other side.

If each side has opposite colors, do the step 14 alg (A perm) once, which will give you the case directly below.

If there is one set of opposite colors, hold them on the top front and do the A perm, then adjust the solved corners to match the centers.

Step 13: Look at the side spots of the top layer edges. If one is solved, or if there are two swapped pairs, skip to the next paragraph. If 2 are solved, or if all four are in any pattern other than swapped pairs, do [M² U²]*2, which will give you one of the above patterns.

If you have two swapped pairs, do R² U (F B') R² (F' B) U R² (U perm), which will leave you with one edge solved.

If you have one edge solved, hold it on the left and do the U perm. You may have to do it twice.

When the top is solved, turn the cube over and solve the other side.

Advance: When you are holding the solved edge on the left, look at the front spot. If it is the opposite of the center, then the U perm will solve it. If it isn’t the opposite, do the U perm, except with U’s instead of Us.

You use most of the same algs, only adding a U perm and the trivial double pair swapper, while losing the H and E perms, but there should be fewer repetitions of each alg, especially if you pay attention to what each alg is doing.

This seems like a significantly better approach than what I did. Doing corners first is probably what I should have thought of in the first place. Plus, the last step for twisting corners is very hard to do. I ended up having to use a two corner twist ZBLL in order to actually solve the cube. Parity is a huge issue for twisting corners, and since four twist algs are very hard to find and are very hard to use, doing that in a "top then bottom" fashion wouldn't work.


anyway, I fixed the mistakes. The step 14 parity alg for the adjacent swap case was meant to be like this: R2 (E perm) R2

 

[Deleted member 36445]

I can't see the video

 

[Discomantis]

I can't see the video

fixed

 

[UnknownCorrespondent]

I've been noodling around with my version of this, and I've found a few ways to reduce the move count without adding any more algs.

Corner Separation: If you need to swap 3 or 4, M² z [M² U²]*2 will swap all 4, leaving you with 1 or 0 to solve.
Edge separation: Not as useful, but you can use [M² U²]*2 to swap 4 at once if they form two opposite pairs. If all 4 need swapping, you can use two of those.
Corner Orientation (my version): You can do most of the work with sune/antisune. If you have parity, one sune or antisune on each face will reduce it to one twisted corner, then finish with one pure twister. Although they are rare, you can reduce the move count for worst case scenarios drastically. WCS would be 7 twisted one way, one the other, which will have parity. You can solve it with two sune/antisune and a twister instead of 5 twisters.