• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Roux 4b to 4c Transition

tukra

Member
Joined
Dec 12, 2014
Messages
18
I've been interested in optimizing Roux LSE 4c cycles for a while and finally sat down and figured something out. I see many before me have come up with solutions. I started to look at them but I find it's best to make my own solution first then I can understand others better in comparison. Please criticize the following approach and tell me how it compares to the commonly used methods. I'm sure most/all of what's useful here is just another way of looking at existing ideas but still curious what you think of this viewpoint.

Orientation: D-White L-Blue

At the end of 4b, M2 to move Blue/Yellow and Green/Yellow edges to the top.
Before any U move look at the LU and RU Red/Orange stickers.

If LU/RU are different (one Red, one Orange), then turn U so FU matches F center.
If Green is on the left then you are already out of the cycle.
If Blue is on the left: M2 U2.​

If LU/RU are the same (both Red or both Orange), then turn U so Blue is on the left.
If FU matches U center: M U2
If FU does not match U center: M' U2​

This seems like an efficient reduction of possible cases and it seems possible to get some lookahead, but I'm also not fast enough to know what enough lookahead is so maybe not.

EDIT: Simplified final recognition further.
 
Last edited:

Teoidus

Member
Joined
Feb 11, 2016
Messages
573
Location
Char
I think BU prediction is attempting to do a similar thing to what you're doing. I wrote up a post on that recently here; have a look and let me know what you think.

(Also, is reviving a thread this old acceptable on this forum? I'm not too familiar with the rules.)
 

shadowslice e

Member
Joined
Jun 16, 2015
Messages
2,923
Location
192.168. 0.1
YouTube
Visit Channel
(Also, is reviving a thread this old acceptable on this forum? I'm not too familiar with the rules.)

Generally it's not encouraged unless you have something new to say although it usually better just to start a new thread. You won't get banned or anything for just one thread but some people may complain if it's over a year old.

You won't get banned or anything unless you bump multiple threads for no real reason at all. Usually anything from the first two pages is fine and though in general there will already be a general topic which you can post in.
 

TDM

Member
Joined
Mar 7, 2013
Messages
7,006
Location
Oxfordshire, UK
WCA
2013MEND03
YouTube
Visit Channel
I've been interested in optimizing Roux LSE 4c cycles for a while and finally sat down and figured something out. I see many before me have come up with solutions. I started to look at them but I find it's best to make my own solution first then I can understand others better in comparison. Please criticize the following approach and tell me how it compares to the commonly used methods. I'm sure most/all of what's useful here is just another way of looking at existing ideas but still curious what you think of this viewpoint.
I actually really like this. Much easier to learn than BU prediction. I think I'll learn it. :) I've been having problems recognising cases with misoriented centres, but I like how this works for that too.
 

tukra

Member
Joined
Dec 12, 2014
Messages
18
I think BU prediction is attempting to do a similar thing to what you're doing. I wrote up a post on that recently here; have a look and let me know what you think.

It was actually seeing your post on Reddit that got me to think about this again and finally cement the thinking I've done on it in the past. I'm sure BU is complete and efficient but when I see instructions like this it just seems overly complicated. What I like about my method is that it's simple, clear and efficient. For my skill level and speed it seems ideal. I'd be really interested to hear how it might not be as good as BU though but I can't tell on my own.

The other thing is that when I looked at BU things in the past they start with U3 sometimes which seems easy to avoid using my method. That always dissuaded me from going further with learning it but maybe there is some useful tradeoff in there?

is reviving a thread this old acceptable on this forum?

When I go to post a new thread I get this message. "Before creating this thread, make sure you have searched the forum to ensure the topic does not already exist." I almost ignored it but the fact that you have to click through it before you can post convinced me it meant what it said.
 

Teoidus

Member
Joined
Feb 11, 2016
Messages
573
Location
Char
I'd be really interested to hear how it might not be as good as BU though but I can't tell on my own.

I think the main advantage of BU prediction is that you get all the information you need from very little "looks": you can see BU during ULUR, and then you get all the other info you need by glancing at F after ULUR is solved. The annoying part of BU prediction is that it takes a lot of thinking at first to figure out what you're supposed to do, which is it's major downside; the idea though is that that processing eventually isn't needed as you start to recognize the cases (there are only 6 really) automatically and they become muscle memory, much like any other step. The hard part is figuring out what stuff you actually want to look at in order to identify a case, and it's not one of those things where you can do it over and over and just sort of figure it out (at least it wasn't for me; I kept looking for blocks and didn't realize I had to look specifically for pieces with BU color). That's what I try to outline in the linked post--a system describing exactly what you should be looking for.


The other thing is that when I looked at BU things in the past they start with U3 sometimes which seems easy to avoid using my method. That always dissuaded me from going further with learning it but maybe there is some useful tradeoff in there?

This sort of bugs me about BU as well, but I think you can avoid it if you pay attention to F while resolving ULUR--you just have to make sure you don't see a solid bar with BU color on F, and you can proceed normally (otherwise just AUF backwards in order to cancel out the U3 and you're OK). I've never been able to put this into practice, though, since I'm currently only able to track/predict BU.
 

GuRoux

Member
Joined
May 6, 2013
Messages
1,712
Location
San Diego, California
WCA
2014TANG03
YouTube
Visit Channel
This sort of bugs me about BU as well, but I think you can avoid it if you pay attention to F while resolving ULUR--you just have to make sure you don't see a solid bar with BU color on F, and you can proceed normally (otherwise just AUF backwards in order to cancel out the U3 and you're OK). I've never been able to put this into practice, though, since I'm currently only able to track/predict BU.

yes, I still do U3 because i never see the cancellation in time, somehow, alexander lau can pretty much always see it in time. could you guys post an example of these different methods with a scramble and solution so it is easier to follow? I think instead of BU, you can predict any of the hidden pieces, and that would work too. meaning you could also do DF, DB, BD, or BU.
 

Teoidus

Member
Joined
Feb 11, 2016
Messages
573
Location
Char
My method isn't really anything new--it's just my way of recognizing cases with BU prediction. I look for the block that involves BU color and recognize like that. There are 6 distinct cases.

Alg.cubing.net demo of all 6:

M'U2M'U2 // BU vert 2x1 down -> U2
(M'U2M'U2)'

MU2MU2M2 // BU 1x1 down -> M2
(MU2MU2M2)'

M2U2MU2M // BU 1x1 up -> M'
(M2U2MU2M)'

M2U2M'U2M // BU 2x1 up -> M'
(M2U2M'U2M)'

U2MU2M' // BU flat 2x1 up -> M
(U2MU2M')'

U2M'U2M' // BU 2x2 up -> M
(U2M'U2M')'


Out of curiosity, how are you deciding how to solve the rest using BU? In your video you say that you essentially solve the 3x1 line with BU color in it. Do you just approach this intuitively? (If so, it's possible you just learned the 6 cases I found by solving them so many times)
 

GuRoux

Member
Joined
May 6, 2013
Messages
1,712
Location
San Diego, California
WCA
2014TANG03
YouTube
Visit Channel
My method isn't really anything new--it's just my way of recognizing cases with BU prediction. I look for the block that involves BU color and recognize like that. There are 6 distinct cases.

Alg.cubing.net demo of all 6:

M'U2M'U2 // BU vert 2x1 down -> U2
(M'U2M'U2)'

MU2MU2M2 // BU 1x1 down -> M2
(MU2MU2M2)'

M2U2MU2M // BU 1x1 up -> M'
(M2U2MU2M)'

M2U2M'U2M // BU 2x1 up -> M'
(M2U2M'U2M)'

U2MU2M' // BU flat 2x1 up -> M
(U2MU2M')'

U2M'U2M' // BU 2x2 up -> M
(U2M'U2M')'


Out of curiosity, how are you deciding how to solve the rest using BU? In your video you say that you essentially solve the 3x1 line with BU color in it. Do you just approach this intuitively? (If so, it's possible you just learned the 6 cases I found by solving them so many times)

note there is always going to be a solved block, kind of like a 1x2x2:
for example: scramble: M' U2 M' U2
the solve "block" would be Front center, DF edge, and Down Center

the solution never breaks this solved "block."
so you just have to figure out how to connect the BU color into a line without breaking the block, should be pretty easy to figure out after playing around with it.

In this case, BU matches the F center color and the FD color, so you just connect those with U2. then the rest should be obvious.
 

Teoidus

Member
Joined
Feb 11, 2016
Messages
573
Location
Char
ALright. I basically used that reasoning as basis when developing the 6 cases I had--so basically I just wrote out explicitly what you ended up figuring out. If only I had the intuition to do it implicitly...
 
Joined
Oct 1, 2010
Messages
361
Orientation: D-White L-Blue

At the end of 4b, M2 to move Blue/Yellow and Green/Yellow edges to the top.
Before any U move look at the LU and RU Red/Orange stickers.

If LU/RU are different (one Red, one Orange), then turn U so FU matches F center.
If Green is on the left then you are already out of the cycle.
If Blue is on the left: M2 U2.​

If LU/RU are the same (both Red or both Orange), then turn U so Blue is on the left.
If FU matches U center: M U2
If FU does not match U center: M' U2​

If LU/RU are different: align M-edges and do the obvious.

If LU/RU are the same: align L/R-edges and do the obvious.

Writing it this way is much less precise, but it emphasizes how easy (and helpfull) it is.
(especially helpfull, if you missed the lookup step in Athefre's method, because of an unintended 4b skip )
 

TDM

Member
Joined
Mar 7, 2013
Messages
7,006
Location
Oxfordshire, UK
WCA
2013MEND03
YouTube
Visit Channel
To reduce the number of moves, it is useful to remember the D2:

MU2MU2M2 // BU 1x1 down -> M2
M'D2MD2

M2U2MU2M // BU 1x1 up -> M'
D2MD2M'

M2U2M'U2M // BU 2x1 up -> M'
D2M'D2M'
D2 is very slow in solves, it requires a regrip and the longer solution is usually faster. You can use Uw moves: e.g. for the first one you can do M' Uw2 M' Uw2, but it's still difficult to do quickly. I think almost everyone uses the 5 move algs.
 
Top