[Help Thread] Roux Discussion and Help

[PenguinsDontFly]

well, if you get used to it it won't be awkward. like using your less dominant hand for OH is more awkward until you practice it.

switching to zz for OH is worth it but its easier to get fast with roux OH (assuming you're faster at roux 2h than zz) instead of wasting time getting used to zz coughadamcough. similarily, why waste time adjusting to doing first block on right when its just as good/maybe better to do it on left?

 

[GuRoux]

switching to zz for OH is worth it but its easier to get fast with roux OH (assuming you're faster at roux 2h than zz) instead of wasting time getting used to zz coughadamcough. similarily, why waste time adjusting to doing first block on right when its just as good/maybe better to do it on left?

true, i agree, that's why i picked right hand OH. in the end, i don't think it really matter much, first block on left or which hand you use.

 

[JamesDanko]

true, i agree, that's why i picked right hand OH. in the end, i don't think it really matter much, first block on left or which hand you use.

I never thought about that. That put's a new perspective on why some Rouxers choose right hand.

Also, nice 4444

 

[ZZTrooper]

I'm not sure if this is the right place to ask this, and it's probably been asked before, but what other algorithms can be learned in addition to CMLL? I use ZZ but I occasionally mess with Roux, and I have a noobish friend who is transitioning into Roux from beginners method. I'm learning COLL for ZZ which you could use for the corners in Roux. If I know all of those algorithms, will I basically know "full" Roux? After learning CMLL, is it basically just working on lookahead, and efficient blockbuilding? Or are there more advanced alg sets for Roux users?

 

[sqAree]

What about an alg set that combines EO and CMLL?

 

[dominugget]

I'm not sure if this is the right place to ask this, and it's probably been asked before, but what other algorithms can be learned in addition to CMLL? I use ZZ but I occasionally mess with Roux, and I have a noobish friend who is transitioning into Roux from beginners method. I'm learning COLL for ZZ which you could use for the corners in Roux. If I know all of those algorithms, will I basically know "full" Roux? After learning CMLL, is it basically just working on lookahead, and efficient blockbuilding? Or are there more advanced alg sets for Roux users?

You will know "full" roux after learning CMLL. You could memorize block cases but that'd be so time consuming.


Sent from my iPhone using Tapatalk

 

[henrysavich]

I'm not sure if this is the right place to ask this, and it's probably been asked before, but what other algorithms can be learned in addition to CMLL? I use ZZ but I occasionally mess with Roux, and I have a noobish friend who is transitioning into Roux from beginners method. I'm learning COLL for ZZ which you could use for the corners in Roux. If I know all of those algorithms, will I basically know "full" Roux? After learning CMLL, is it basically just working on lookahead, and efficient blockbuilding? Or are there more advanced alg sets for Roux users?

CMLL is "full" roux, but I believe (non-roux solver here so correct me if I'm wrong) OLLCP can be used during CMLL to influence/skip the EO step of LSE. However it's not as great as it sounds because fast roux solvers almost always influence UL/UR during EO, so by skipping EO you make UL/UR slower (And you also made the CMLL step slower by using OLLCP). There really aren't any roux alg subsets with as good a return on your time spent learning as some of the supplemental CFOP subsets.

 

[dominugget]

CMLL is "full" roux, but I believe (non-roux solver here so correct me if I'm wrong) OLLCP can be used during CMLL to influence/skip the EO step of LSE. However it's not as great as it sounds because fast roux solvers almost always influence UL/UR during EO, so by skipping EO you make UL/UR slower (And you also made the CMLL step slower by using OLLCP). There really aren't any roux alg subsets with as good a return on your time spent learning as some of the supplemental CFOP subsets.

This is very true, a eo orientation supplemental would interrupt the flow of the solve, which is the most important reason of why I chose roux.


Sent from my iPhone using Tapatalk

 

[shadowslice e]

Well, I did propose an extension a while back which will always force a 4-flip but TBH, you don't really need it unless you are very, very fast and even then it will take less than half a second off.

 

[PenguinsDontFly]

CMLL+EO would be very nice, but its like 1176 algs or something insane, and many of them would be ****. Knowing 3-4 algs for each CMLL is good enough. Then you can have 1 alg to preserve EO, one to flip 6, and 2 others to flip other edges to force arrow/avoid something bad.

edit: I think its 1302 algs...maybe?
edit 2: **** I forgot mirrors cuz some H Pi U and T cases are symmetrical...approximation 1000-1300

 

[shadowslice e]

CMLL+EO would be very nice, but its like 1176 algs or something insane, and many of them would be ****. Knowing 3-4 algs for each CMLL is good enough. Then you can have 1 alg to preserve EO, one to flip 6, and 2 others to flip other edges to force arrow/avoid something bad.

edit: I think its 1302 algs...maybe?
edit 2: **** I forgot mirrors cuz some H Pi U and T cases are symmetrical...approximation 1000-1300

I reduced it to 85 by forcing 4 flips inteasd of eo skips cause I think that the 4-flip is nicer than the skip

 

[PenguinsDontFly]

I reduced it to 85 by forcing 4 flips inteasd of eo skips cause I think that the 4-flip is nicer than the skip

so by learning 85 algs you can get arrow every time? or by 4-flip do u mean anny 4 flip? cuz some of those are ugly.

 

[shadowslice e]

so by learning 85 algs you can get arrow every time? or by 4-flip do u mean anny 4 flip? cuz some of those are ugly.

I should have put the sub-text- if you keep 1 and only 1 misoriented edge on D, you can look at the pieces on the top. If you have 1 misoriented edge, you do CMLL+4 flip on U, if you have 3, just do COLL.

 

[TDM]

I should have put the sub-text- if you keep 1 and only 1 misoriented edge on D, you can look at the pieces on the top. If you have 1 misoriented edge, you do CMLL+4 flip on U, if you have 3, just do COLL.

We were discussing this on Reddit quite recently. You're still going to have some very bad algs, and you're going to need setups for quite a few cases which will slow you down enough for this method to actually be slower than standard Roux.

Also,

Well, I did propose

So many of the things "you" have proposed have been thought of many times before.

 

[shadowslice e]

Also,So many of the things "you" have proposed have been thought of many times before.

Yeah, in the original post detailing the method (it was a couple of months ago or so in this very thread), I did say I guessed it has been thought up before but I was just throwing it out there to see if anyone was interested along with any cons.

Which is pretty much why all the stuff I do now is pretty abstract rather than the "low hanging fruit" so to speak.

 

[dominugget]

Hey guys I feel like my blocks are kind of linear, and that I'm doing F2L with M-Slices instead of blocks. Any advice on how to reduce move count? (I'm at 9-14 STM.)


Sent from my iPhone using Tapatalk

 

[Praetorian]

What about an alg set that combines EO and CMLL?

dot OLLCPs in CFOP

 

[2180161]

Does anyone solve a 2-flip case (one in U one in D, setup: M' CCW U-perm M) like this: M U M' U2 M' U M' ?

 

[GuRoux]

Does anyone solve a 2-flip case (one in U one in D, setup: M' CCW U-perm M) like this: M U M' U2 M' U M' ?

that's what i would do for that case.

 

[2180161]

How many cases are there for EO+ULUR? Would it be worth it?