4x4 Roux?

[unirox13]

recently i learned roux and have started using it as my main 3x3 method. today i was fiddling around with my 4x4 solving it using the original method i had learned; solving the centers, then the edges, then proceeding to solve the cube using the LBL method of a 3x3. after a few solves i began to start using Roux to solve it after solving my centers and edges. it actually works surprisingly well (and is helping to continue my practice of roux). The parody's still come into play, and same algorithms solve them however they do need to be used at different times than with the LBL method.

does anyone else use this method to solve 4x4 or larger cubes? also, is there a way to use Roux to completely solve the 4x4. I've tried a times to solve the entire cube with a roux type method to no avail.

just figured i'd throw that out there and see what peoples thoughts were on it

 

[waffle=ijm]

You can use Redux...or stadler where you make 2 1x3x4 blocks and CMLL then Centers and Edges and L6E. Or use Meyer where you make opposite centers, first block and then centers and edges then 3x3 stage with the first block already done.

using redux and Roux is fine, I'm avging 56-57 with it

EDIT - most preferred method by fast Roux users is K4

 

[Cyrus C.]

Wasn't Goins working on a Roux-like 4x4x4 method before he left? Anyone know whether that was any good?

 

[waffle=ijm]

whatever he was thinking. it was probably just stadler

 

[The Puzzler]

I made a petrus method on 4x4.

 

[vcuber13]

You can use Redux...or stadler where you make 2 1x3x4 blocks and CMLL then Centers and Edges and L6E. Or use Meyer where you make opposite centers, first block and then centers and edges then 3x3 stage with the first block already done.

most preferred method by fast Roux users is K4

meyer sounds like yau but with roux instead of fredrich.
I dont think k4 is prefered by rouxers, i use k4 and cant sub 30 with "fake roux"
and what about kenneths big cube method? (i think thats what is called)

 

[The Puzzler]

I use a strange 4x4 method where I get centers, solve cross. Make F2l like pair of dege and corner, insert it, slice over, insert matching dege, repeat. Solve LL by making paired deges while orienting/permuting. Is this a good method and what is it called?

 

[masterofthebass]

meyer sounds like yau but with roux instead of fredrich.
I dont think k4 is prefered by rouxers, i use k4 and cant sub 30 with "fake roux"
and what about kenneths big cube method? (i think thats what is called)

more like yau is meyer with fridrich instead of roux. Richard was doing that method way back in 2007 (maybe even 2006)

 

[jms_gears1]

whatever he was thinking. it was probably just stadler

Ya, just stadler. He was all liek LOOK AM HAVE NEW METHOD. I was all like, ive been using that since forever ago... Its stadler.

 

[waffle=ijm]

I dont think k4 is prefered by rouxers, i use k4 and cant sub 30 with "fake roux"

So you're implying that you can only be good at K4 if you are a good Roux user?

Right, and I can't be good at redux because I'm not a good CFOP user

 

[blade740]

more like yau is meyer with fridrich instead of roux. Richard was doing that method way back in 2007 (maybe even 2006)

THE TRUTH ABOUT DAN COHEN IS OUT.

 

[irontwig]

There's also Kenneth's method: http://www.speedsolving.com/wiki/index.php/Kenneth's_Big_Cubes_Method
(F2B, CMLL, F3L, ELL)

 

[Kenneth]

Yes, my method is pretty close to Roux if you look at the steps I do but I think only CMLL is exactly the same as the original method. I use several ways for F2B and sometimes I even use Akimoto starting from columns.

The most obscure part of my method is MF3L (the part after F2B and CLL) but it is really easy and all intuitive, I start with D and B (or F) centres followed by the pairing of the DB dedge that is then placed (use F for keyhole = r r' or l' l), then I do triplets of one edge and two centres for the F side (example: r U' r' U r U' r' U2 r U2 r' .. inverse for setup).

The last part is ELL 1-3 where ELL 3 also takes care of the parity if there is one.

Edit: I made an example solve and placed it at this page

 

[Kenneth]

I made a tiny scrambler for Roux 4x4 (it preserves F2B and LL corners)

Some examples:

Spoiler

1. U r U2 l2 U l U' r l2 U2 r' U' l U' M2 U2 M U' r2 U2 l U' r l2 U2
2. M2 U' r2 l' U2 l2 U' M2 U' r2 U l U2 r l U l' U2 r' l' U r2 l U2 r l2 U' r' l'
3. r2 l U' M2 U2 r l2 U M U' r l2 U r l U r2 l U2 r l U2 r' l' U M' U' r' l2 U2 M2
4. r2 U' r2 l' U2 r l2 U' r2 U' r' l2 U2 M2 U r' U2 M' U' M U2 r' l' U' l2 U' M
5. U2 r l U' M U2 r2 l' U M' U' r2 U r' U M2 U' r' l' U2 r2 U M U' M' U
6. U' l2 U r2 l U M' U2 r' l2 U r2 U' r l2 U2 l2 U r2 U' M2 U2 r2 l' U' l2 U
7. U2 r U r2 l U' r' l' U2 r' U' r2 l' U M U' M U2 M2 U2 l2 U' r2 l' U2 M U'
8. U r U2 M U' M' U' r' U r' l' U M2 U' M U l' U' l2 U M' U' r' U'
9. r2 U' r2 l' U l U2 r' l' U r U2 r' U' M' U l2 U' l' U r2 U2 r' l2 U2 r l2
10. M' U M2 U' r2 l' U' r' l2 U' l U2 r2 l U' l' U r' l2 U r U l U' M2 U r'
11. r2 l' U2 l' U M' U l U l2 U r' l' U2 r2 l' U2 r U2 r2 U l U2 l2 U' M
12. U2 r2 l' U r' l2 U2 l U2 r2 U2 M2 U2 M2 U2 r U r' l' U2 M' U2 M' U' M U'

It is not really user friendly as is but I will remake it and put it at the web for download later.

Edit, it is written in QBASIC, here is the source:

Spoiler

RANDOMIZE TIMER
DIM mov$(18)
mov$(0) = "U "
mov$(1) = "U' "
mov$(2) = "U2 "
mov$(3) = "r "
mov$(4) = "r' "
mov$(5) = "r2 "
mov$(6) = "r l "
mov$(7) = "M "
mov$(8) = "r2 l "
mov$(9) = "M' "
mov$(10) = "r' l' "
mov$(11) = "r2 l' "
mov$(12) = "r l2 "
mov$(13) = "r' l2 "
mov$(14) = "M2 "
mov$(15) = "l "
mov$(16) = "l' "
mov$(17) = "l2 "
CLS : PRINT "Kenneth's 4x4x4 Roux scrambler v1.02 December 2010": PRINT
ax% = INT(RND * 2)
FOR n% = 0 TO 22
  ax% = (ax% + 1) MOD 2
  IF ax% THEN
    mov% = INT(RND * 3)
  ELSE
    mov% = INT(RND * 15) + 3
  END IF
  PRINT mov$(mov%);
NEXT n%

First line seeds the random generator using the timer as source (like srand in C), then follows a indexed string variable that holds the diffrent moves that will be printed and after that the strings are initialised. The rest is pretty self explaining, ax% is the axis (% = integer variable) and mov% is the index for the table.

Edit again: Now I'm using the scrambler to make some examples for step 5 in my method (the sole reason for me making the scrambler, I needed it for this =), one done, more to come.

 

[Kenneth]

Here you can download my DOS executable: https://sites.google.com/site/kubbuth/4x4roux/4xRoux.zip

Usage is simple, click the prg icon and it will open a DOS window, that first asks for the number of scrambles and then the length of the scrambles. Just type and enter... and then nothing happens!! (there are no input error checks so it is probably easy to make it bug). Then open the text file that is in the same folder as the prg and you have your scrambles =)

Note that the resulting length is variable, any combination of r and l counts as one move (axis metric).

 

[vcuber13]

So you're implying that you can only be good at K4 if you are a good Roux user?

Right, and I can't be good at redux because I'm not a good CFOP user

thats the opposite of what i implied

 

[Kenneth]

This is fun:

Scramble : r l2 U r2 l' U2 r2 l' U l2 U r2 l' U2 r' l' U' l2 U M U' r' l2 U2 r' l2 U2 r U2 r2 U2 M

Block up centres (1x2) : U l' U r2 U' l' U' l2 U' l' U' ... last two moves I put a solid centre in U so I can AUF
UR: l2 F R' F' M' F R F'
UL: l2 F' L F M' F' L' F

(orient the F's to U)

Roux redux! =) solve the rest as the middle layers of a 3x3x4!

 

[peedu]

Hi,

I reduced 4x4 into 3x3 yesterday and tried to solve the 3x3 part in Roux style. Then I noticed that I'm doing the M' (rl') and M2 (r2l2)moves with middle and/or ring finger of both hands symmetrically and M (r'l) turns mostly with two index fingers (holding the R and L layer between a thumb and a middle finger) or sometimes a combination of index finger and thumb of different hands.

I'm not an everyday Roux user, but sometimes I just play with it for fun.

That fingertrick with both index fingers to turn M would look more funny if I had a mini QJ. Qrrently I only have QJ and MF8/Dayan.


Peedu