One-Answer 3x3 Roux Question Thread

[AreRouxAmused]

How would you people teach Roux to a beginner? So it has a minimum amount of algs.

 

[shadowslice e]

How would you people teach Roux to a beginner? So it has a minimum amount of algs.

I learned it as 2 look CMLL and basic instruction on how EO works which I just experimented with a lot

 

[Oatch]

How would you people teach Roux to a beginner? So it has a minimum amount of algs.

I'd recommend checking out this video tutorial series. It's aimed at a complete beginner and teaches a simplified version of the Roux method, using only 2 algorithms for CMLL (sune and niklas) if I remember correctly. The remainder of the solve would be intuitive.

 

[AreRouxAmused]

Thank you I am trying to teach my cousin's easily so they can get a handle on it.

 

[Umm Roux?]

How would you people teach Roux to a beginner? So it has a minimum amount of algs.

For cmll, you could teach them to spam sune and then adjacent swap. For eo, m' u m' is a large part so that sequence and a few other 3-movers are not too bad, recognition is a concept that will take a day or two. For ul ur, if the pieces are not both on the top or bottom, do some variant of m u2 m and then join. For the final cases, only the 3 cycle is hard. How i remembered it is that you need to find the edges that pair, pair them and then push the pieces out centers-first. Adjust m-slice as necessary. This was basically my thought process when I first learned roux.

 

[Shiv3r]

bump. For beginners, instead of teaching them like T or A perm for adjacent swap, I taught my sister like this(with LBL but still works):
-spam Sune to orient the corners(teach rules as well, and it is important to give the algs names its easier to remember them for beginners)
to solve corners, teach them Niklas. I tell them that Niklas swaps the 2 corners on the Left side, but it also twists the corners, so you have to do a U2 and then a Sune to fix it(tell them that sune does not disturb the position of the corners)

The amount of moves is only 5 more than A-perm and about the same as T-perm, and it is easier to remember because the move sequences are shorter(7 moves for both niklas and Sune, and by tracking F2L pairs it becomes easy to remember).
Also, what are your thoughts on slow turning solves as a way to improve? should I do them more often than speedsolves?

 

[Teoidus]

Also, what are your thoughts on slow turning solves as a way to improve? should I do them more often than speedsolves?

Roux solves average 48 moves

Do an algorithm that you know really well at 4 tps and you'll realize it's quite slow

48 moves / 4 tps = ~12 second solves

 

[Shiv3r]

Roux solves average 48 moves

Do an algorithm that you know really well at 4 tps and you'll realize it's quite slow

48 moves / 4 tps = ~12 second solves

also, What was your average when you got your first sub-10 solve? I still have not gotten a single one... and I'm sub-16..

 

[Teoidus]

I think I was sub-15. It was a ~30 stm solve I think

Just solve enough and you'll get there eventually. singles don't really mean much imo

 

[Shiv3r]

I think I was sub-15. It was a ~30 stm solve I think

Just solve enough and you'll get there eventually. singles don't really mean much imo

wow. lemme see, you had under 3 TPS on average if I did the math correctly(iirtmc?)

 

[Teoidus]

wow. lemme see, you had under 3 TPS on average if I did the math correctly(iirtmc?)

Probably, yeah

For example here's a sub-15 avg at rly low tps:

 

[Sion]

I want to learn one look CMLL, but I feel intimidated with the amount of algs. Any tips on How I should learn it?

 

[TDM]

I want to learn one look CMLL, but I feel intimidated with the amount of algs. Any tips on How I should learn it?

It's fewer algs than OLL!

I would learn one set at a time. That way you can clearly see how you're progressing each time you finish a set. Learn one alg a day: it's not much and you'll get the whole thing done in a little over a month. (you might actually find it helpful to learn the algs in pairs for some cases, which would get it done even quicker)

 

[shadowslice e]

I want to learn one look CMLL, but I feel intimidated with the amount of algs. Any tips on How I should learn it?

Well, it's not really that many. If you know s-look then you only need to learn another 33 algs (also, I would learn it in sets probably H, Pi, T, L, S, AS).

 

[TDM]

Well, it's not really that many. If you know s-look then you only need to learn another 33 algs (also, I would learn it in sets probably H, Pi, T, L, S, AS).

Why not L earlier? It's probably the fastest set.

 

[shadowslice e]

Why not L earlier? It's probably the fastest set.

It's just the order I learned it. I would only really say H first then S/AS last but otherwise up to the cuber.

 

[deadcat]

I want to learn one look CMLL, but I feel intimidated with the amount of algs. Any tips on How I should learn it?

Alternatively, you could start by learning which cases your 2-look algs solve, so you can predict when you get a skip / adjacent swap / diag swap. That will start training your 1-look recog. Then you can add one by one algs for the diag swaps, so you can stop doing Y perms all the time
In any case, take it slowly, and make sure you practice all the algs you already know before adding each new one.

 

[Shiv3r]

This

Alternatively, you could start by learning which cases your 2-look algs solve, so you can predict when you get a skip / adjacent swap / diag swap. That will start training your 1-look recog. Then you can add one by one algs for the diag swaps, so you can stop doing Y perms all the time
In any case, take it slowly, and make sure you practice all the algs you already know before adding each new one.

I did the exact same thing, just learning which algs my 2look algs solved. But that was only after I learned a set at a time.

 

[Teoidus]

tbh I learned S/AS first because they looked like they had the weirdest recognition, so by the time i had learnt the other sets i had gotten used to it.

Plus all the S/AS CMLLs are basically R U2 R' U2 R' F R F' mirrored and inverted all over the place + some niklases

 

[Shiv3r]

tbh I learned S/AS first because they looked like they had the weirdest recognition, so by the time i had learnt the other sets i had gotten used to it.

Plus all the S/AS CMLLs are basically R U2 R' U2 R' F R F' mirrored and inverted all over the place + some niklases

yeah, sunes are the easiest to learn algs but hardest to recog I just learned CMLL by going just sequentially down the table(I still like the original table style), but instead of just learning all the algs from other tables I genned a lot of my own. Also, I disagree with Kian that You should learn the nicest and best algs first. I say that the most important thing to learn is recognition, and while learning good recognition you should learn the easier-to-learn algs. but As soon as you got full CMLL recog down, Start optimizing your CMLL set that you use in solves, which I call my CMLL "battle set".

Here is a good example
case G4(pi, front swap). When I first learned this case, I learned the easier to learn algorithm:
(R U R' U R U2 R') U (L U' R' L' U' R).
This case is easy to remember and learn because it is a Sune, AUF and then a Niklas. It also happens to be the COLL for the case, so I still use it from time to time.

However, as I got better and started replacing the slower algorithms with quicker algs, and so I learned this alg:
R U2 R' U' R U R' U2(R' F R F')
Which is faster, but Is a little harder to learn at first, and learning new algs now and then is easier than mass-learning a bunch of fingertricky but hard-to-learn algs.

Just my 2 cents on the idea.
-Shiv3r