# [Help Thread]Roux Discussion and Help

#### DNF_Cuber

##### Member
It is called CMLL:

D

#### Deleted member 55877

##### Guest
I’m old so go easy. I need a link to a tutorial (I’m not sure of the name of the step) for the part after building the second block. Right now I do a weird alg to get the corners in the right place then the R U R’ U R U2 R’ over and over to get all yellows on top. I think there must be a set of algs that combine those steps, right? And I would prefer ease over speed. I’m at 48 seconds so I don’t need the absolute fastest way, just an improvement.
thanks!
The step you are referring to is called CMLL (where you solve all 4 yellow corners). It seems that you repeat the sune algorithm to orient the corners, which is fine for beginners. If you want to learn the algs to orient the corners in one step I'd recommend watching:
But at your level I think you should focus more on L6E and blockbuilding

#### oneshot

##### Member
The step you are referring to is called CMLL (where you solve all 4 yellow corners). It seems that you repeat the sune algorithm to orient the corners, which is fine for beginners. If you want to learn the algs to orient the corners in one step I'd recommend watching:
But at your level I think you should focus more on L6E and blockbuilding
Yeah so it’s basically what I do, I guess I’m just extremely slow with both recognition and execution of CMLL. I’ll have to learn to finger trick the algs eventually. Thanks

#### Athefre

##### Member
I often think about how great of a discovery this method is. It is simple and symmetrical. It is parallel Waterman in appearance and solves the look-ahead problem of corners first methods. The first block solves all of the pieces where your left hand holds the cube and would normally be blocking your view of that area. After the first block, almost everything is in view with a great moveset. This freedom means that second block doesn't take many moves. Then CMLL only requires the solver to learn 42 algorithms to be competitive at the top level.

Not only did Gilles find this simple method, he found an amazing way to finish. Anyone else might have decided that it was enough to have FB+SB+CMLL and just used any of the various LSE methods that already existed for corners first methods. Initially the Roux method was using those older LSE methods. Then what is used now was found. I think the Roux LSE method is an even greater discovery than the 3x3 method itself. The iterative process of orienting the edges was a new concept that turned out to be the best way to do it. Even further, it was found that the UL+UR edges can be solved during this EO process and the M-Slice edges can be influenced during the final turns of solving UL+UR. Each substep melts together to form a single step.

It is a great thing to see that this method has grown to have such a large community of solvers.

#### Athefre

##### Member
ACMLL

Here is a new technique and group of algs for Roux. It is the application of A3 where F2B is built using what's called passive blockbuilding. You build and attach pairs or pieces without worrying about the blocks being perfect 1x2x3s. The blocks will then be corrected during CMLL. This reduces the move-count of F2B and allows for more planning during inspection. There is still a lot of work to be done. For now what has been generated are the block types that seem most likely to have good CMLL algs. I have tested other block types and many of those also seem promising. Another helpful thing would be analysis for the move-count difference for each passively built block arrangement compared to a normal block.

A3 can be applied to other methods. For those methods, the right algorithm step would have to be identified. A3 works so well for Roux because there are many unsolved pieces after F2B. This results in good algs for CMLL. For CFOP, there would be more to memorize if OLL is the step used to correct the blocks. However it is possible that it would have better algs compared to PLL. For ZZ, OCLL is likely the best step. There would be a very small number of algs and the algs would most likely be very good.

#### tkoalas

##### Member
Hi, everyone! So I am learning the Roux method, and I got into a position I have no idea how to get out of. There isn't a Roux algorithm for the position I am in, and I am really confused and frustrated. I have been staring at it, swapping corners, trying the Sune algorithm over and over trying to get it to work (R U R' U R U2 R'), all to no avail. I can rescramble and try again, but there shouldn't be any stalemate positions. What am I missing? Do you guys see something I missed?

I just got this cube, and so I am frustrated as this is my first solve with it. I was so excited, but now I'm stuck.

The upper layer is oriented as the front layer in the images. I solved the first two blocks (blue and green, two layers; white is on bottom). Can anyone help?

Last edited:

#### Pyjam

Welcome to the forum.
One corner is twisted. Turn the yellow/orange/blue corner clockwise (put blue on trop).
Then F R' F' R U2 R U2' R'

#### tkoalas

##### Member
Welcome to the forum.
One corner is twisted. Turn the yellow/orange/blue corner clockwise (put blue on trop).
Then F R' F' R U2 R U2' R'
Thanks! So, I figured out what I did.... the cube is a lot looser than I am used to, and so I think I accidentally flipped one of the corners. I wasn't expecting that to be the cause, but it makes sense now. *facepalm*

#### Nir1213

##### Member
I often think about how great of a discovery this method is. It is simple and symmetrical. It is parallel Waterman in appearance and solves the look-ahead problem of corners first methods. The first block solves all of the pieces where your left hand holds the cube and would normally be blocking your view of that area. After the first block, almost everything is in view with a great moveset. This freedom means that second block doesn't take many moves. Then CMLL only requires the solver to learn 42 algorithms to be competitive at the top level.

Not only did Gilles find this simple method, he found an amazing way to finish. Anyone else might have decided that it was enough to have FB+SB+CMLL and just used any of the various LSE methods that already existed for corners first methods. Initially the Roux method was using those older LSE methods. Then what is used now was found. I think the Roux LSE method is an even greater discovery than the 3x3 method itself. The iterative process of orienting the edges was a new concept that turned out to be the best way to do it. Even further, it was found that the UL+UR edges can be solved during this EO process and the M-Slice edges can be influenced during the final turns of solving UL+UR. Each substep melts together to form a single step.

It is a great thing to see that this method has grown to have such a large community of solvers.
I hope more cubers use Roux method, it doesnt have as much advice and users as CFOP.

#### Rouxster

##### Member
Just found out an alg for pi left bar cmll. ( I am 100% sure that it has been invented before) : r U' r' U2 R' F R F' U2 r u r'
I still think that the standerd alg ( R' U' R' F R F' R U' R' U2 R) is better but this one preserves EO.

#### Pyjam

I do R' F' U' F U' R U R' U R
F' with the right index.

#### Platform_

##### Member
I want learn LSE oriented edges.

#### abunickabhi

##### Member
Sure, what problem have you faced till now? LSE is quite intuitive. To make EO in LSE, you have to make an arrow.

Otherwise check out the video resources by Kian for more clearer explanation, R U2 F' U M' U M U2 F U2 R'.

#### OreKehStrah

##### Member
I want learn LSE oriented edges.
Hey, just wanted to remind you not create a thread every time. Instead of creating another repeat thread you should have just searched up one of the many Roux threads already made!

#### Platform_

##### Member
Sure, what problem have you faced till now? LSE is quite intuitive. To make EO in LSE, you have to make an arrow.

Otherwise check out the video resources by Kian for more clearer explanation, R U2 F' U M' U M U2 F U2 R'.
my problem is EO

#### abunickabhi

##### Member
my problem is EO
Then you have to understand how the arrow is made, and how using [M,U] you can get the arrow, and then solve EO using M U M' or M U' M' and so on (assuming the arrow is formed at the back)

#### Athefre

##### Member
This is my 15 year anniversary of using Roux! I'm one of the first five Roux users. Back then, after Gilles Roux first presented his method, the first people to ever seriously use the Roux method as their main method were me, Thom Barlow, and a few others. At that time our only guide was Gilles Roux's website. It was, and still is, an amazing site. But there are so many more resources and such a large community today.

Like most, I started with LBL. The use of algorithms breaking everything that I worked so hard to build in F2L just didn't make sense to me. So I soon decided to use corners first instead. After that, I started really researching the faster methods. Back then there weren't nearly as many methods as there are now. We had CFOP, Petrus, Salvia, Waterman, Human Thistlethwaite, Roux, and Heise. ZZ didn't exist yet. I pretty much immediately threw away the idea of using CFOP because I disliked LBL and I also didn't want to learn 57 OLL algs. Salvia was interesting, but didn't really feel fast. Human Thistlethwaite was really fun and eventually partially inspired my MI2 method, but it also wasn't really a speedsolving method. I used Petrus for a while and came very close to choosing that, but the EO step felt like a waste of time. Heise was just too difficult at that point. So it came down to Waterman and Roux. Both were similar to the CF method I was using and enjoying. I loved the Waterman method, but chose to give that one up because of the high alg count. So Roux it was. It just felt right.

Choosing the Roux method is the reason that I have developed so many method-related things. Everything started with me solving the recognition problem for NMCMLL (in coordination with Gilles Roux). After that I spent a lot of time in the pseudo world, developing NMLL, discovering transformation, and developing A2 and A3. My latest development, ACMLL, is something I feel will really push the Roux method further.

There were so many doubts about the potential of the Roux method for the first several years. It was often called too intuitive, meaning it couldn't compete with the easy F2L pairing and heavy algorithm use that is in CFOP. Now look what people are able to achieve. Slice moves were criticized as being too slow. Then Roux users proved that incorrect by developing and using great fingertricks. Nowadays Roux is considered by many to be one of the top two methods. It has been amazing to see the Roux community grow from the original five of us to it now having many YouTube guides, WRs, a large Discord server, a Facebook group, and inspiring many other methods. And everything points to the method continuing to grow in popularity.

#### abunickabhi

##### Member
This is my 15 year anniversary of using Roux! I'm one of the first five Roux users. Back then, after Gilles Roux first presented his method, the first people to ever seriously use the Roux method as their main method were me, Thom Barlow, and a few others. At that time our only guide was Gilles Roux's website. It was, and still is, an amazing site. But there are so many more resources and such a large community today.

Like most, I started with LBL. The use of algorithms breaking everything that I worked so hard to build in F2L just didn't make sense to me. So I soon decided to use corners first instead. After that, I started really researching the faster methods. Back then there weren't nearly as many methods as there are now. We had CFOP, Petrus, Salvia, Waterman, Human Thistlethwaite, Roux, and Heise. ZZ didn't exist yet. I pretty much immediately threw away the idea of using CFOP because I disliked LBL and I also didn't want to learn 57 OLL algs. Salvia was interesting, but didn't really feel fast. Human Thistlethwaite was really fun and eventually partially inspired my MI2 method, but it also wasn't really a speedsolving method. I used Petrus for a while and came very close to choosing that, but the EO step felt like a waste of time. Heise was just too difficult at that point. So it came down to Waterman and Roux. Both were similar to the CF method I was using and enjoying. I loved the Waterman method, but chose to give that one up because of the high alg count. So Roux it was. It just felt right.

Choosing the Roux method is the reason that I have developed so many method-related things. Everything started with me solving the recognition problem for NMCMLL (in coordination with Gilles Roux). After that I spent a lot of time in the pseudo world, developing NMLL, discovering transformation, and developing A2 and A3. My latest development, ACMLL, is something I feel will really push the Roux method further.

There were so many doubts about the potential of the Roux method for the first several years. It was often called too intuitive, meaning it couldn't compete with the easy F2L pairing and heavy algorithm use that is in CFOP. Now look what people are able to achieve. Slice moves were criticized as being too slow. Then Roux users proved that incorrect by developing and using great fingertricks. Nowadays Roux is considered by many to be one of the top two methods. It has been amazing to see the Roux community grow from the original five of us to it now having many YouTube guides, WRs, a large Discord server, a Facebook group, and inspiring many other methods. And everything points to the method continuing to grow in popularity.
Wow a nice backstory, and congrats on completing 15 years with the Roux method. I only started using Roux 5 years back, when I saw Kaijun's video where is solving quite fluidly using Roux.

Its took to see how you have summarised all the methods you have learnt till now, and also pushed the frontiers a bit with your own theories.

#### jdh3000

##### Member
I just completed my first real Roux solve. Yes it was slow, yes it was sloppy, but I didn't use anything from cfop to finish it.

Up til now cfop has been my method, but after discovering oRoux, the block building interested me. I would do the right and left block, but fizzle out and give up trying to go further, resorting to finishing the last part by fixing the 2 bottom edges, then cfoping the last layer.

I thought it was the only way I was going to be able to do it because each time I would try to learn cmll algorithm, my mind would jump to oll/pll.

But just forced myself to do it to fix the four corners. I used some tutorials and honestly had to find the absolute beginner types.

I worked on the 2 look cmll, and the algorithms are not really unlike some off the algs I know from full oll/pll, yet I have trouble recognizing the cases and solving them.

I worked on them until I pretty well had them down and recognizable.

Then came lse, and it seemed a bit foriegn to me. I've done cfof so many years that the only m slices I do are in a few algorithms. Outside of those, any intuitive m slices are new territory for me.

I managed to get the edge orientation and set up an arror case, then finish the L & R sides, and finally the very last step.

I have no desire to quit cfop, but I wanted to know Roux. I know this is the most basic form of it, but once I get used to it, I can build on with more advanced stuff.

Sometimes when I've done a lot of solves with cfop, I get a little worn out. I think Roux can give me a little break because it's fun, and seem to require a different way of thinking. I can spend a day with it, or just a few solves here and there... then again I may like it so much as to devote a lot of time to it.

I've asked many questions here and I appreciate all the help I've been given.

#### GodCubing

##### Member

I wonder what step I should work on...
Any tips for second block efficiency (and lookahead tbh)