# [Help Thread]Roux Discussion and Help

#### Etotheipi

##### Member
Um...no comment?
? I think the reason th ar method is bad is the AUF might not already be done, half the time you need to do a U2

#### Attachments

• IMG_20200408_130157.jpg
608.9 KB · Views: 7
• IMG_20200408_130202.jpg
754.6 KB · Views: 6

##### Member
You've not solved centres correctly. Do an M or M', reorient edges and continue.

In future, please use the Roux help and discussion thread as opposed to creating a new thread.

#### Jasiko

##### Member
You've not solved centres correctly. Do an M or M', reorient edges and continue.

In future, please use the Roux help and discussion thread as opposed to creating a new thread.
Thank you and sorry, I'm new

##### Member
Thank you and sorry, I'm new
No problem! It's just that keeping similar posts in common and established threads keeps the forum neater and easier to search/read

#### Etotheipi

##### Member
Any tips on L cmll recog? I learned all the algs at one point, but the recog stumped me and I forgot them.

#### AbsoRuud

##### Member
Any tips on L cmll recog? I learned all the algs at one point, but the recog stumped me and I forgot them.
It's something I struggle with as well. I am just trying to follow Kian Mansour's documents as best as I can.

#### Micah Morrison

##### Member
I'm not a rouxer, but as someone who knows COLL I tried to learn all the algs from the same angle if possible and then notice the different color patterns.

#### Athefre

##### Member
How to Use Non-Matching Blocks in a Roux Speedsolve

Below is my method for using non-matching blocks. Roux users still don't use this move saving technique. Not because it saves just a few moves and really lucky cases are rare. It is for three reasons - corner recognition, EO recognition, and aligning the left and right layer at the end of the solve. I will show that all three of those aren't actually issues and there are simple and fast ways to accomplish this.

First Block:

This will be the same as every other solve - solve your first block. If you are color neutral in some way, even better. Find the easiest first block within the inspection time.

Second Block:

1x2x2: Now you can take advantage of the lucky cases that you see. Even better, you don't have to be constrained to forming a block that matches the one on the left. It gives a sense of freedom to simply pick whatever you see and just solve. You may have even had time during inspection to plan at least part of the 1x2x2 if you saw an easy case.

Last Pair: Now that a 1x2x2 is formed, you have two options for the next pair. In a solve where you are required to match the two 1x2x3 blocks, you wouldn't get to take advantage of the other pair option if you noticed that it is easier.

EO Preparation: This is the important part for LSE EO. During the second block, track the UR edge and note its location. You will have even more time to track the UR edge during CMLL. At the end, when aligning the second block, use r or r' to align the centers so that they match the block on the left. Because you now know where the UR edge is, after CMLL you can proceed with EO with no issues. You can also choose to track the UL edge instead and orient centers to match the right side block. This center orientation can also be done during the performance of the CMLL algorithm.

CMLL:

I have developed two corner recognition methods for non-matching corners.

Method 1 (Fastest): The best way to recognize corners is using ACRM, a recognition method that is based on the L/R colors. First, you find the orientation of the colors that match the left and right side of your cube. Then you check just two pre-determined sticker positions. This is just as fast as the recognition Roux users currently use for matching blocks. This recognition method works equally well for matching and non-matching blocks. It also works for Conjugated CxLL (CCLL). Below is a link to the recognition method and algorithms.

Athefre's Corner Recognition Method (ACRM)

Method 2 (Slower): An alternate solution is to continue to use U color recognition. You find the orientation of the colors that should be on the U layer, then check pre-determined spots for a pattern.

To locate the U layer colors, it helps to pay attention to the blocks that you have built. If you have an orange block on the left with white on the D layer, two of the U layer corners will need to have yellow on the U layer and they both will have orange colors. The block on the right will work the same way, the other two U layer corners will contain two sets of matching colors based on the block. It may help to start predicting orientation while finishing the second block. Now that you see your orientation, you can check the pattern. This will work mostly the same as usual. There are a few rules to remember.

1. L/R colors in the pattern will be recognized as normal. Orange and Red will always be opposite for example.
2. If your blocks are an R/R' away, any F/B colors in the pattern will match the status of the two 1x2x3 blocks. If the left block has blue on the front and the right block has white on the front, then those two colors will be considered matching colors. That also means green and yellow will be matching colors. Then blue and yellow will be opposite and green and white will be opposite.
3. If your blocks are an R2 away, those F/B colors in the pattern will be the opposite from usual. Two blues will be considered opposite and a blue and green will be considered matching.

Once you've chosen one of the above recognition methods, perform the algorithm. During this algorithm, you can continue to track the UR (or UL) edge as described in the second block step.

LSE:

Now that you know where the UR (or UL) edge is, EO will be exactly the same as normal. There will be no recognition problems. For 4c, there are some additional tricks that will be useful for aligning L and R at the end. Below are some examples of the many possibilities.

r U2 M' U2 r
r U2 M U2
r' U2 M' U2
r u2 M' u2 (versus r' U2 M U2 r2 or M U2 M U2 r2 R')

Last edited:

#### Exotic Butters

##### Member
Hello! How I average around 15 with roux, and one of my biggest struggles is first block. If i can find a pre-made 1x1x2 pair, I can usually figure out a good and efficient solution for the entire block, which often gets me good times. On the contrary, however, whenever I have a scramble which has no 1x1x2 blocks, I usually cant cannot figure out a good first block solution (solution ends up being 11+ moves), and as a result, I end up getting terrible times (17s+).

Any help would be greatly appreciated

#### Rainger

##### Member
Hello! How I average around 15 with roux, and one of my biggest struggles is first block. If i can find a pre-made 1x1x2 pair, I can usually figure out a good and efficient solution for the entire block, which often gets me good times. On the contrary, however, whenever I have a scramble which has no 1x1x2 blocks, I usually cant cannot figure out a good first block solution (solution ends up being 11+ moves), and as a result, I end up getting terrible times (17s+).

Any help would be greatly appreciated
I’m having the same issue but double your time haha

#### Rainger

##### Member
How to Use Non-Matching Blocks in a Roux Speedsolve

Below is my method for using non-matching blocks. Roux users still don't use this move saving technique. Not because it saves just a few moves and really lucky cases are rare. It is for three reasons - corner recognition, EO recognition, and aligning the left and right layer at the end of the solve. I will show that all three of those aren't actually issues and there are simple and fast ways to accomplish this.

First Block:

This will be the same as every other solve - solve your first block. If you are color neutral in some way, even better. Find the easiest first block within the inspection time.

Second Block:

1x2x2: Now you can take advantage of the lucky cases that you see. Even better, you don't have to be constrained to forming a block that matches the one on the left. It gives a sense of freedom to simply pick whatever you see and just solve. You may have even had time during inspection to plan at least part of the 1x2x2 if you saw an easy case.

Last Pair: Now that a 1x2x2 is formed, you have two options for the next pair. In a solve where you are required to match the two 1x2x3 blocks, you wouldn't get to take advantage of the other pair option if you noticed that it is easier.

EO Preparation: This is the important part for LSE EO. During the second block, track the UR edge and note its location. You will have even more time to track the UR edge during CMLL. At the end, when aligning the second block, use r or r' to align the centers so that they match the block on the left. Because you now know where the UR edge is, after CMLL you can proceed with EO with no issues. You can also choose to track the UL edge instead and orient centers to match the right side block. This center orientation can also be done during the performance of the CMLL algorithm.

CMLL:

I have developed two corner recognition methods for non-matching corners.

Method 1 (Fastest): The best way to recognize corners is using a method I developed based on L/R colors. First, you find the orientation of the colors that match the left and right side of your cube. Then you look at pre-determined spots for matching colors. This is just as fast as the recognition Roux users currently use for matching blocks. This recognition method works equally well for matching and non-matching blocks. Below is a link to this recognition method.

NMCLL

Method 2 (Slower): An alternate solution is to continue to use U color recognition. You find the orientation of the colors that should be on the U layer, then check pre-determined spots for a pattern.

To locate the U layer colors, it helps to pay attention to the blocks that you have built. If you have an orange block on the left with white on the D layer, two of the U layer corners will need to have yellow on the U layer and they both will have orange colors. The block on the right will work the same way, the other two U layer corners will contain two sets of matching colors based on the block. It may help to start predicting orientation while finishing the second block. Now that you see your orientation, you can check the pattern. This will work mostly the same as usual. There are a few rules to remember.

1. L/R colors in the pattern will be recognized as normal. Orange and Red will always be opposite for example.
2. If your blocks are an R/R' away, any F/B colors in the pattern will match the status of the two 1x2x3 blocks. If the left block has blue on the front and the right block has white on the front, then those two colors will be considered matching colors. That also means green and yellow will be matching colors. Then blue and yellow will be opposite and green and white will be opposite.
3. If your blocks are an R2 away, those F/B colors in the pattern will be the opposite from usual. Two blues will be considered opposite and a blue and green will be considered matching.

Once you've chosen one of the above recognition methods, perform the algorithm. During this algorithm, you can continue to track the UR (or UL) edge as described in the second block step.

LSE:

Now that you know where the UR (or UL) edge is, EO will be exactly the same as normal. There will be no recognition problems. For 4c, there are some additional tricks that will be useful for aligning L and R at the end. Below are some examples of the many possibilities.

r U2 M' U2 r
r U2 M U2
r' U2 M' U2
r u2 M' u2 (versus r' U2 M U2 r2 or M U2 M U2 r2 R')
I would really like to see a video of all this. It makes a lot of sense but will reach less rouxers sitting here in the forums. Awesome work.

#### Athefre

##### Member
I would really like to see a video of all this. It makes a lot of sense but will reach less rouxers sitting here in the forums. Awesome work.

Thank you. I will look into making a video.

#### Etotheipi

##### Member
I timed some splits:
FB: 3.67
SB: 5.55
CMLL: 3.38
LSE: 5.66

Each is an Ao12 of just doing that step, which ones need most work?

#### mukerflap

##### Member
I timed some splits:
FB: 3.67
SB: 5.55
CMLL: 3.38
LSE: 5.66

Each is an Ao12 of just doing that step, which ones need most work?
LSE and SB, cmll will improve naturally, SB and LSE can use some work. Also work on planning fb in inspection

#### Rainger

##### Member
Hey everyone,

After CMLL and the arrow you solve the left and right layers leaving only the middle layer unsolved. Right when you do the M2 to solve the left colors and right colors you are left with..

..my question is.. how do you know which way to turn here (U or U’) to be most efficient? What is your method here? What do you look for? I usually get it right by looking at the left top edge but occasionally I get it wrong and it slows my times down.
Thanks,

#### WarriorCatCuber

##### Member
Hey everyone,

After CMLL and the arrow you solve the left and right layers leaving only the middle layer unsolved. Right when you do the M2 to solve the left colors and right colors you are left with..

View attachment 11970

..my question is.. how do you know which way to turn here (U or U’) to be most efficient? What is your method here? What do you look for? I usually get it right by looking at the left top edge but occasionally I get it wrong and it slows my times down.
Thanks,
Are LU and LR stickers opposite colours? If so, which direction will make them match the centers?

#### Rainger

##### Member
Are LU and LR stickers opposite colours? If so, which direction will make them match the centers?
I’m not asking for just THIS case, I should have been more clear, am asking for what is going through your head and what should I be thinking and watching for. Or are you telling me to watch to see if they are opposite colors earlier in my solve so I’ll know that at this step? Just I get stuck here for 2 seconds occasionally because I don’t know what exactly to look for.