8355 method: easy way to solve cube?

[felix]

I learned how to solve the cube using the LBL method, but found that 1) the middle layer algorithm is too long and 2) there are 4 complicated algorithms you need to perform multiple times to solve the last layer.

I found this 8355 method by Mr. Reheart Sheu and I think it is much better. There are only 3 easily memorizable algorithms. I would like to know is this method available in English?

Website
Video

8 bottom layer
Cross on bottom, 3 corners. Total of 8 pieces solved on the bottom. The unsolved corner is the keyhole, keep it at FRD.

3 edges
Use keyhole to solve 3 of the 4 middle edges.

5 edges
Use keyhole to solve the remaining 5 edges.

The last two edges may require you to use algorithm(s).
Rotate the middle layer until the unsolved edge is at FR.
Rotate the top layer until the unsolved edge is at TR.
If the two edges are in the incorrect location need to be swapped, apply algorithm 1: (hold F face with left hand) F'U' FU' F'U' F
If the two edges are in the correct location but need to be flipped, apply algorithm 2: (hold F face with left hand) F'U FU' (hold R face with right hand) RU' R
You may need to use both algorithms.

5 corners
Now the cube is solved except for 5 corners. Hold the cube so that there are 4 unsolved corners on the top face and 1 unsolved corner at FRD.
Look at where the corner currently at FRD is supposed to go. Rotate U until the place it is supposed to go is directly above it at FRU.
Perform algorithm 3: R'D'RD
until FRU is solved. The bottom layer of the cube will be messed up but that doesn't matter. Again look at FRD, turn U layer, and repeat alg. 3 until done!

It may be the case that the top layer is done but the bottom layer is messed up. If this happens, repeat alg.3 until all the cube is done except for 2 corners. Rotate the cube until the two unsolved corners are at FRU and FLU. Repeat alg.3 until corner at FRU is done. Perform U'. Again repeat alg.3 until corner at FRU is done. U and you're done!


I like this method because it feels more like solving the cube instead of just recalling some formulas. i tried teaching some people and they can generally get it after a few times. (Alg.1&2 take a bit longer.) There are 2 more algorithms you can learn to make it faster. (post #13)

 

[jcuber]

It's a method called the keyhole method and he uses a very long (use those 3 algs many times) LL method.

 

[EmersonHerrmann]

1) The alg for the middle layer edges is NOT long.

2) The LBL algorithms are not hard to learn at all if you have the decency to take the time to learn them.

The keyhole method at the beginning has already been invented, nothing new. The LL sucks because it's restricted and takes a long time.

Edit: Maybe I'm being harsh...

 

[rahulkadukar]

Well practice 3 Look LL and the middle layer edges can be done in under 8 seconds even by a slow guy like me

 

[Stefan]

is this method available in English?

Your presentation here is the only one I know, at least of this exact method. One reason might be that today's dickhead kids simply can't appreciate its beauty and attack it for no good reason when they see it (other examples are Bruce being attacked for novel PLL statistics and George for his opposite sides method).

Are you asking for yourself or for others? You seem to have understood it already.

 

[chevyLi]

I have to argue if you call it suck, This method is invented by a math teacher in Taiwan, who plays rubik's cube as long as it was invented.

As a teacher , his concerns to the method is a non-algorithms method for beginners or those who mock cubers of their algorithms to appreciate to solve a rubik's cube type puzzle without any algorithms but only some few concepts. So it's not for speedcubing, but in a better(maybe) way to promote the cubing world. Though it's only available in Chinese. I'll make an English version if I have time.

I'ts also very efficient on solving megaminx, you also need no algorithms to solve a megaminx with this method.

 

[ISuckAtCubing]

to the people who is trying to lear to solve the cube, the edges alg is long for them, they usually give up there and how to put to corners in after the cross

 

[Tetris Cube]

is this method available in English?

Your presentation here is the only one I know, at least of this exact method. One reason might be that today's dickhead kids simply can't appreciate its beauty and attack it for no good reason when they see it (other examples are Bruce being attacked for novel PLL statistics and George for his opposite sides method).

Are you asking for yourself or for others? You seem to have understood it already.

Wtf you can say dickhead on this forum but not the P word synonymous with urine?

Back on topic...

The method seems interesting for an almost algorithmless solution. It'd be a great beginner's solution that requires more thinking than memorization. I've never seen that LL approach, orienting and solving edges with the last slot edge solved, but not the corner.

 

[blgentry]

Very unusual method. First a quick correction:

Alg 3 is listed as R' B' R B . That's incorrect. The algorithm should be:

R' D' R D

I find the last stage to be counter intuitive. It's very easy to learn, but I don't "see it" happening on the cube. I just blindly apply the (short) algorithm over and over. So in that sense, I don't think this solution is more "natural" or teaches you more about the cube or something.

It's an interesting method. Getting those last two edges correct *without* using algs 1 and 2 is a challenge. You have to think ahead a bit.

Brian.

 

[blgentry]

After playing with this for a while, doing a number of solves, I think I see the appeal. It's a really neat method. I haven't mastered exactly how to do it, but you can always place the last 5 edges without using any algorithms. There's this funny three way swap that you have to do to finish it, but it works every time.

In the last step, where you solve the final 5 corners, you sometimes end up with the correct cubie in the FRD position. This can happen at the start of the step, or part way through, after positioning one or more corners. If that happens, all you have to do is run R' D' R D once, and then continue as normal.

I had a very hard time following the videos, as I only speak something like 15 words of Mandarin. The java apps work, but it's hard to gain insight from them. So if the original poster wants to elaborate more, I'd enjoy reading it.

Brian.

 

[Stefan]

There's this funny three way swap that you have to do to finish it

What's that?

 

[blgentry]

^ The "three way swap" I'm referring to is of the final three edges that are placed. At that stage, two edges, in the upper layer, are correctly placed and oriented. The other three edges are all in the wrong places. In one step, you then move an edge out from the second layer, placing it correctly in the upper layer, move the 2nd layer edge from the upper layer to the second, *and* swap the position of the incorrect U layer edge to it's proper position.

Example:
Put your LL face on top and set up with:
U' R2 D2 F2 D R2 B2 U' R2 D F D2 F' L2 R2 U' B' y

The edges at UL and UB should be correct, while the ones at UF UR and FR should be incorrect. The "three way swap" to solve all of them is:
U' R U R'

I set this one up to try to make it easier to see. But getting that set up during a solve requires some forethought. I'm working on being able to do it every time, on my first try. Sometimes it takes me 2 or even 3 tries to set it up (during a solve) so that the 3 way swap happens properly. Plus you sometimes have to orient edges by pulling them into the keyhole slot and then reinserting them.

Brian.

 

[felix]

1) The alg for the middle layer edges is NOT long.
2) The LBL algorithms are not hard to learn at all if you have the decency to take the time to learn them.
The keyhole method at the beginning has already been invented, nothing new. The LL sucks because it's restricted and takes a long time.

1)3 steps is shorter than 8.
2)I would argue the LL for the LBL method sucks more. You need to learn 4 algorithms and the orientations to hold the cube just to solve the last layer. As a beginner, those algorithms seems a bit intimidating.

Are you asking for yourself or for others?

More non-Chinese speaking people should know about this method.

I have to argue if you call it suck... I'll make an English version if I have time.

Thanks. I don't have a webcam so no video from me.

The algorithm should be: R' D' R D

Thanks. The "three way swap" is nice, I need to try it.


Improvements: after you have mastered the beginner method, you can add 2 more algorithms to speed it up.

Algorithm 1 to swap the last two edges can be divided into 1a and 1b.
If keyhole is at FR, F is orange, R is yellow, U is green.
Alg 1a: If the wrong piece on U is yellow on top and orange on side, twist U so it is at UR, directly over the yellow face. F'U' FU' F'U' F
Alg 1b: If the wrong piece on U is orange on top and yellow on side, twist U so it is at UF, directly over the orange face. RU R'U RU R'

Alg.3 can also be divided into 3a and 3b.
3a: R'D'RD (aka +1)
3b: D'R'DR (aka -1)
These two algorithms swaps and rotates the corners at FUR and FDR.

Assume U is green
If the corner is at the bottom and you want to move it to the top:
Use 3a if green is at R
Use 3b if green is at F
Use either 3a or 3b three times if it is at D (but I will usually avoid it and solve another piece first)

If the corner is at the top and you want to move it to the bottom:
Use 3a if green is at R
Use 3b if green is at F

Other pieces in the bottom will be messed up, it will be fixed after you do 3a/3b six times. Keep a running total in your head: start at zero, add 1 every time you do 3a and subtract 1 every time you do 3b. Your total at the end must be zero, +6 or -6, so when you have one corner left on the top layer, just ignore the orientation and do 3a/b until the total is zero or six.

 

[bundat]

I used to use this method in the past (I knew quite a few beginner methods before).

I just wanted to share that I did LL edges purely intuitively,
I didn't have to memorize any algorithms for it at all.

I just position 3 edges in their right places with simple intuitive moves:
i.e. if the edge at FR is oriented correctly (like ZZ's EO orientation)
AUF-(R U/U2/U' R'), just making sure that the initial R twist puts an edge in its right place, and R' pulls back down into FR an unsolved edge.

Then just mirror it to the left hand with (F' U/U'/U2 F) if the edge is not oriented. Or just pull out an unsolved edge out of the LL using those algs if the edge at FR is the correct edge for FR.

I usually end up solving all 4 edges this way, or having solved 3 and the last one left in FR.
Which I easily solve with my only "alg" (mirrored), which is intuitive as well.I just do what is similar to assembly's "rotate in the carry bit", the carry bit being the FR edge, i.e. shift all the edges by one position.

e.g. if oriented - put the unsolved LL edge on FU, then do R U R' U R U R'
It's not an alg you have to memorize.
It's pretty intuitive, R replaces one solved edge with your FR edge, then with (U R') you place the solved edge that you took out back next to it while taking out another edge, then you do (U R) to place the edge you took out next to it while taking out the last LL edge, which you place in the last slot with (U R').
Don't memorize it, just do the inverse and watch how it works.

Just mirror it to the left-hand, and you only have one "special alg" case that isn't exactly non-intuitive either. (Very finger-trick friendly too)

Sometimes I also end up doing the "3-way swap" and the other "double-flip" alg accidentally, purely intuitively without even attempting to run a specific alg.

...I realize that I really didn't add anything new to the posts, I just tried to share the logic I was using when using this (or a similar) method in the past, which was purely intuitive from my perspective.

 

[qqwref]

I've never heard about this method before, but it's really cool. It seems like a very clean and simple way to solve the cube, and I like how there's only one algorithm that you have to do lots of times (as opposed to the standard Western beginner method where you have to do every single LL algorithm more than once in most cases).

 

[Scigatt]

This is, essentially, my beginner method with a few differences(L D2 L'/F' D2 F is used instead of R' D' R D, for example) . (see http://jeays.net/rubiks.htm, solution 2)

 

[daniel0731ex]

i planned to translate the 8355 method tut long ago...but i wasn't able to because i was working on the unicube.tw purchasing instruction

anyway, i might be working on it soon, but don't expect me to finish it quickly


http://teach.ymhs.tyc.edu.tw/t1086/RC/l3-8355/8355.htm

 

[Stefan]

I really didn't add anything new to the posts

Wrong.

I just tried to share the logic I was using when using this (or a similar) method in the past, which was purely intuitive from my perspective.

And you did it well. I think about it the same way, btw (shifting all the edges by one position around the LL center).

 

[Cride5]

What a cool method! I used to teach beginners petrus because its more interesting, but in hindsight, its probably too difficult for a complete beginner. This is amazingly simple, and amasingly easy!!

Muchos graicas Felix

 

[felix]

^
You should thank Mr Sheu, I'm just the translator.

Thanks to Scigatt, I found this 1982 Handbook of Cubik Math by Alexander Frey and David Singmaster.

It describes a method that is similar to the 8, 3 and 5. The last 5 corners are put in place using a 3-cycle swap first and then rotated in pairs, it's kind of confusing. The keyhole is located at the back, for some reason. The book describes it as being "chosen from the many known methods" so it seems 75% of the 8355 method was discovered before 1982.