New method idea

[jtjogobonito]

I was thinking the other day about how we use the 3x3, and not the 2x2, for the last solves on every cube, except 2x2(excluding beginner methods). Then, I was timing the steps in my solve and found that my cross usually takes 3 seconds. After that, I started to figure out why my cross was so bad. This is when I though of this idea. When you create the cross you only put the edges on the D layer correctly orriented, without worrying about permutation. Then you solve the F2L and OLL normally, and finish it off with a D AND U permutation. Here is an example solve:

Scramble on D: D B F' L' B F' U B2 D2 F L' D2 U B' L' B' U2 B2 F' L' U2 F' U' B R
Cross: y L2 F' U' R' F
F2L 1: y R' U' R2 U R'
F2L 2: U R' U R d L' U L
F2L 3: U R U' R' d' R' U R
F2L 4: d' L U L' U L U2 L' U2 L U L'
OLL: y' r U2 R' U' R U' r'
Permutation(using the Cube Explorer optimal algorithm): U B' F2 F' D R2 B U' B L2 U B' U2

Please leave your comments about this method.



P.S. I could not find a way to upload my other method onto my site so PM me if you want those algorithms.

 

[martijn_cube]

last permutation alg doesn't work.

 

[rjohnson_8ball]

Interesting, but there can be a parity problem if you are not careful, like only 2 edges needing to be swapped on the first layer. Similarly, you could just randomly get all the first layer cubies oriented correctly (both edges and corners), then use PLL on that first layer later on... But again, there is that parity problem of 2 edges (or 2 corners) needing to be swapped.

The nice thing about doing the first 2 layers as they are commonly done is because there is so much freedom in movement, allowing short cuts. When you are down to only one layer left, movement is constrained.

 

[jtjogobonito]

Alright I fixed the last permutation alg. rjohnson I do not understand what you mean about the "freedom in the F2L" Doing it this way destroys no freedom in the F2L. Also, the corners shouldn't be incorrectly placed because the centers are there. Lastly, you don't permutate the first layer later on because A) that requires too many moves B) there can be parity. The parity is in the example solve that I have given you. The final permutation gets rid of the parity.

 

[Lt-UnReaL]

Permutation(using the Cube Explorer optimal algorithm): U B' F2 F' D R2 B U' B L2 U B' U2

U B' R2 F' D R2 B U' B L2 U B' U2

I've thought of this before, don't know how many algs or time it would save, though...

 

[Harris Chan]

Solving a cross normally would take less than 2 seconds (perhaps 1 to 1.5 on average) with enough practice (Of course there's cases when people try to preserve pairs, etc.). I don't know if just orienting would actually shorten the number of moves/times by much, at the cost of the parity at the end.

 

[Lotsofsloths]

Solving a cross normally would take less than 2 seconds (perhaps 1 to 1.5 on average) with enough practice (Of course there's cases when people try to preserve pairs, etc.). I don't know if just orienting would actually shorten the number of moves/times by much, at the cost of the parity at the end.

+ the immense amount of algs

 

[Jai]

( people try to preserve pairs, etc.).

Was that a reference to me?

Yeah, I don't think that just orienting would save much more moves. anyway, cross isn't really a major problem, and only orienting won't make the cross 1-2 moves on average, would it? And plus, with all the extra permutation algos you'd have to learn to "master" this method, you'd might as well be putting your efforts into learning something like ZB.

I'm not saying it's a bad method; it's interesting, it's an interesting take at a sort of Ortega for 3x3 edges, but takes too much work to learn, with the results not really being worth the effort.

 

[Harris Chan]

( people try to preserve pairs, etc.).

Was that a reference to me?

Just in general You're not the only one that do that you know

 

[Lt-UnReaL]

There's 21 cases for just the top layer, and 10 cases(I think?) for the bottom layer, so it would be 210 cases?

 

[Jai]

And I believe this is where Johannes comes in with long calculations

 

[Lucas Garron]

Nah.
PLL: 22 even, 22 odd
EPLL: 5 even, 5 odd

Makes 22*5*2-1 = 219 non-solved cases (AUF/ADF, but no mirroring)

 

[rjohnson_8ball]

jtjogobonito, Sorry, I did not have a cube handy and did not read closely. I thought you were talking about doing PLL on the U face and then the D face separately, but now I see you mean one permutation involving the cross and U layer. The algorithms might be awkward to remember or difficult to recognize. I am not sure if your method saves time.

There have been times when I try to solve the cube under low lighting, and get to the final layer when I notice I goofed the cross, like a yellow-red where a white-red should be. I use M and U twists to swap the pieces, then do the final layer.

 

[brunson]

Lastly, you don't permutate the first layer later on because A) that requires too many moves B) there can be parity.

Since everyone else is picking apart the method, I thought I'd jump on the bandwagon and point out that "permutate" actually means the exact opposite of what I think you intended. Sorry, just being a mathematician.

BTW, it seemed like a fun idea, it just needed to be thought through. I think it's always good to hear and explore new ideas.

Since everyone else is creating a method (edges first, corners first, 2x2x2-> 2x2x3->etc, 1x2x3+1x2x3, etc.) I'm claiming this method: Solve a 1x2x3 block, then orient all edges using M2 cycles, then extend to 2x2x3, permute remaining edges, permute corners on one unsolved face, permute remaining corners.

That's it, it's mine, I said it first. If anyone ever uses this method they must call it EBURPHECCE, which stands for "Eric Brunson Uber Roux/Petrus Hybrid Edge Corner Corner" method.

Algorithms to follow (someday, yeah right).