I learnt rubiks cube before the YouTube days from a blog. I do not find the algorithm anymore online but it is something like

  1. White cross
  2. White corners (except one)
  3. Second layer corners (except one)
  4. Yellow cross and the one left out corner in second layer
  5. Exactly 5 pieces are left and you solve them one by one by a standard 8 move algorithm. (You may have to orient the cube or make one 'fake' move to get the cube in right orientation for the algorithm to work)
  6. (Optional): More likely than not exactly 2 corner pieces are left, oriented incorrectly at their right places. A standard 16 move algorithm works to orient them correctly.

Now I am very comfortable solving by this method but step number 4 is kind of devised by me and it is slightly cumbersome and takes like 30 seconds for me. I never figured out what is the documented approach for step 4 and now I am unable to find it. So my question is

Can someone help me find a documented approach of this algorithm especially step number 4?

Slightly unrelated rant I have is that all the new algorithms online are certainly faster but they have lot many moves to memorise (one for every possible possible position). I think are more 'mechanical' but take away the fun of 'real-time' thinking and puzzle solving to bring the cube in the right orientation before applying those algorithms. (Limited ones to memorise but happens a lot in step 5 and 6).

  • $\begingroup$ " but they have lot many moves to memorise (one for every possible possible position)." Wouldn't all valid solving algorithms have this property? $\endgroup$
    – Chipster
    May 18, 2020 at 21:09
  • 1
    $\begingroup$ @Chipster No. Older algorithms like the ones I know had only 3 routines to memorize. You bring the cube to a position where these routines can be applied. This bringing the cube to a desired position has no routine to memorise. You just solve it out by thinking analytically. $\endgroup$
    – PagMax
    May 20, 2020 at 4:53

1 Answer 1


This seems to be the 8355 method by Reheart Sheu. It seems the original blog was Taiwanese but no longer exists, but the speedsolving wiki has a description.

Step 4 consists of solving 3 of the last layer edges intuitively (U face, yellow edges), and then the last two edges (one in the U layer, and the keyhole edge in the middle layer at FR) may need to be swapped or flipped.

That wiki only lists the swaps R U R' U R U R' (U2) and its mirror image F' U' F U' F' U' F (U2). The first sequence is used if the piece at FR has yellow on the front, and will insert that piece at the UF location. The second sequence is used if the piece at FR has the yellow on the right, and will insert the piece at the UR location.

If the last two edges are in their correct locations, but need to be flipped, then you swap them first with one of the sequences, and then again with the other.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.