The New Method / Substep / Concept Idea Thread

[V Achyuthan]

Not proposing this thing but just asking everyone if this is a variant of hoya cause I used it in a solve recently (for some reason)
Basically before pairing the last cross edge with hoya why don't we just finish the centres and solve the Last cross edge like Yau.

 

[Eli Apperson]

Not proposing this thing but just asking everyone if this is a variant of hoya cause I used it in a solve recently (for some reason)
Basically before pairing the last cross edge with hoya why don't we just finish the centres and solve the Last cross edge like Yau.

bc it interrupts the flow. Having to completely change how you solve edges just for one edge doesn't really make sense.

 

[TheEpicCuber]

bc it interrupts the flow. Having to completely change how you solve edges just for one edge doesn't really make sense.

I agree. If you solve the last cross edge, then you have to do F L' to be able to preserve the cross edges and continue solving the centers.
Therefore, doing just those moves to solve the centers breaks the fluidity of the solve and you are gonna have to fix the set-up later.

 

[Eli Apperson]

I agree. If you solve the last cross edge, then you have to do F L' to be able to preserve the cross edges and continue solving the centers.

And for less than a 4-5 move difference it isn't really worth it.

 

[Skewbed]

Hexagonal Francisco Variant:

A Random Example Solve (over 90% 2-gen <RUru>, 64 HTM):

alg.cubing.net visualizer link

Scramble: U F2 D F2 L2 D2 R2 D' L2 F2 D' R2 L B2 U' R' F' U R U R

R D L2 // Square
R F2 // Square + Extra Corner Piece
U' R U R' u' R' // Hexagon
R U R' r U r' // 1st Edge
u R U R' r U r' // 2nd Edge
u U2 r U r' // 3rd Edge
u U2 r U2 r' // Pseudo-LSLL
y' U' R' U R U' R' U R D2 // LS
r U R' U' r' F R F' // OLL
U2 R U' R U R U R U' R' U' R2 U' // PLL

I'd imagine this method could do a lot better on a luckier scramble, and with someone who has better blockbuilding than me. This solve was pretty unlucky with having so many twisted E-slice edges, but is a good demonstration of how close to 2-gen this method can be.

 

[PapaSmurf]

Hexagonal Fransico is actually quite good. I feel like there should be a better way to finish though, but L5C->L6E is a lot of algs. Maybe some sort of conjugation approach could work.

 

[SciKE]

Y4c
A two-sided Roux 4c recognition system expanding on GodCubing's EZ4c system for 3 cycles.

Spoiler: Introduction

To start off, if you do not know already, there are currently 2 main 4c recognition systems everyone uses. Each recognize what 4c case you can get before solving 4b, aka the UL and UR edges by looking at certain pieces and where they would end up.

BU is the easiest. Before inserting ULUR, you look at what edge would be in BU after inserting and keep that edge in mind while you insert and AUF. Based on what color that edge is and what color your F center is, you can deduce what 4c case you have fairly easily.
DFDB is the better bigger brother to BU. You look at the colors of the UF and UB edges before inserting ULUR and keep those edges in mind. Based on that as well as the BU edge, you can deduce your 4c cases even further.

GodCubing made a recognition system that allows you to eliminate the need to look anywhere besides the U and F faces to deduce what cycle cases you get. The problem is this only really works for deducing 3cycles. GodCubing's technique is using the UF sticker and comparing it to whatever sticker our AUF is aimed at. So if we would have to AUF U to solve the corners after inserting ULUR, we would use UF and the UL stickers. Vice versa with U' and UR stickers. As well as that he compares the FU sticker to the F center color, similar to the BU and DFDB systems. Using these two, we can determine what type of cycle we get and if we can cancel out moves by AUFing the opposite direction without needing to look at anything else.

I wanted to try and keep that 2 side recognition and simplicity of opposite or same colors. Instead of remembering what edge/edges will be in BU/DFDB you can figure it out just by tying some cases to the 4c cases. After tinkering around with this, not really trying to remember what cases are what, I'm fairly used to and confident using this already. I feel like this could be learned as easily as DFDB with some simple rules to follow. As many have said before, it doesn't really matter what 4c system you use, as long as it works and you are consistent with it. I just wanted to go more in depth with an idea somebody else proposed.

Basically you are comparing the UF sticker to the UL and UR sticker, and then the FU sticker to the F center. If you draw the arrows these comparisons make, it makes a Y shape, hence Y4c. This is broken up into 3 "recog pairs", each recog pair can have one of two states. The stickers in the pair are opposite colors, or the stickers are the same color. Using the 3 pairs in 6 different combinations (with a smaller one off rule that is trivial), you can break down all of the possible 4c cases into these 6 unique cases.

The first recog pair is the UF sticker plus whatever edge your AUF would put into UB. So if you would do a U after inserting the ULUR edges, the target sticker would the be UL edge, vice versa with U' AUF and UR edge.
The second recog pair is the opposite edge of the target edge from the first recog pair. You can think of it as the edge your AUF would put to UF. So if you AUF U, it would be the UR edge, vice verse with AUF U' and UL edge.
The third recog pair is the FU sticker and the F center.

AUF normally unless specifically saying otherwise.
**Dots cases can either AUF normally or opposite way, it depends on how you solve dots cases. If you solve them using ME gen, AUF normally; if you use MU gen, AUF the other way

UF + target	UF + opp target	FU + center	4c case	Example
opp	opp	opp	ULUR line, dots** UFUB line, 4c skip, AUF opposite	U' M U M2 U M2 U M U' M2 U2 M2 U2 M2 U' M2
opp	opp	same	ULUR line, columns UFUB line, 4c skip	U' M U M2 U2 M2 U' M' M2 U M2 U M2 U M2 U' M2 U'
opp	same	(opp)	M or M' U2 cycle	U2 M U2 M' U2 M2 U' M2
opp	same	(same)	M or M' U2 cycle	M U2 M' U2 M' U2 M U M2
same	(opp)	opp	M2 U2 cycle	M' U2 M U M2 U M2 U2 M2 U'
same	(opp)	same	4c skip, AUF opposite	M' U' M2 U M2 U M' U M2 U'

Here is a link to the google sheet version (along with my mega sheet !!!)

Spoiler: Optimizations

Common Patterns/Observations/Optimizations

The easy cases (s/o)
These cases you do not need to recog the opposite target. Once you see that UF and the target sticker are the same, you can skip to recoging the FU+c pair. If the UF+target pair is the same color, the opposite target pair will always be the opposite color.

The intermediate cases (o/s)
These cases do not need to recog the FU+c pair. Once you recog that the target sticker is opposite and the opposite target is the same, you know you have a "raise the dot" 3cycle either way. The difference between the FU+c pair does not recog anything for these cases.

The "sometimes maybe good, sometimes maybe s**t" cases (o/o)
These cases dont have optimizations relating to the recog pairs. But they do have a small addition that allows you to distinguish them further. For both cases if you have a line connecting the UF and UB edges and the U center, it will be a 4c skip.

Making the recog even easier on your brain
So instead of remembering the cases like "opp, same, same" or something like that, you can just remember which of the pairs are the same color and recog off of that. This doesnt lose anything and still works 100% of the time. So the s/o/o case, which is a M2 U2 cycle, could just be recognized that only the UF+target pair is the same.
Eventually the patterns of each case will become brainless, possibly as easy as BU recognition.

Spoiler: Pros and Cons

Pros
2 sided, no need to head tilt
Simple recog, whether the colors match or not is trivial and brainless to recog
Could be or is as brainless as DFDB or BU after some practice
Works equally well with misoriented centers

Cons
The 3 recog pairs is a bit more than the recog for BU or DFDB, plus the one line case
Its very conditional, some cases recog some things while others dont, could be awkward to get used to
Being a somewhat different approach to 4c prediction, this can be awkward to switch too

I am planning to make a video for this method just in case I am bad at explaining things over text. If this is confusing please ask I am happy to answer.

 

[Quốc Hưng]

Hello, I have a new4x4 speedsolving method for LEOR solver.

Overall Step :
1.F2C like Meyer
2.First Block (Like Meyer)
3. Pairing andOrient DF DB edge then put them at DR,BR
4.L4C
5.Finish DF DB edge
6.EOPairing+OLL parity.
7.Right Block->COLL+EPLL(with pll parity)
How do you feel about this idea?

I think it very good because two step is done after reduction (FB,EOstrip)

 

[Quốc Hưng]

LEO4 METHOD​

Hi, I just had an idea a few days ago, about a 4x4 rubikscube method for LEOR solver.

Steps :
1.L/R Center
2.FirstBlock (Like Meyer)
3.F4C
4.Solve DF, DB dedge.
5.EO Pairing the remaining edges without messing up the block (Do this like Z4, but without messing up the 3x3x4 Block)
6.FLIP Final Badedge
7.Solve like LEOR or Petrus (RightBlock->LL+Parity PLL)

Pros
* Good ergonomics during edge pairing as no F/B moves are needed
* Final 3x3 phase is just RU blockbuilding followed by LL, no pause for inspection of EOSTRIP.
*Two Step is done after Reduction, resulting lower movecount for 3x3StAGE.
*Less cube rotation
*Better ways to deal with parity. The number of algs used in a solve with double parity is one less than with standard Yau with double parity, and PLL parity recognition is easier due to corners being solved.
Cons
* More moves during edge pairing.
* Good ability to see EO during lookahead is required, but this should be pretty natural for LEOR users already.
* In some cases, flipped ledges may accidentally be created but this should be avoidable with good lookahead.

Do you think this a good method ?

Inventer : Hồ Nguyễn Quốc Hưng​

 

[V Achyuthan]

LEO4 METHOD​

Hi, I just had an idea a few days ago, about a 4x4 rubikscube method for LEOR solver.

Steps :
1.L/R Center
2.FirstBlock (Like Meyer)
3.F4C
4.Solve DF, DB dedge.
5.EO Pairing the remaining edges without messing up the block (Do this like Z4, but without messing up the 3x3x4 Block)
6.FLIP Final Badedge
7.Solve like LEOR or Petrus (RightBlock->LL+Parity PLL)

Pros
* Good ergonomics during edge pairing as no F/B moves are needed
* Final 3x3 phase is just RU blockbuilding followed by LL, no pause for inspection of EOSTRIP.
*Two Step is done after Reduction, resulting lower movecount for 3x3StAGE.
*Less cube rotation
*Better ways to deal with parity. The number of algs used in a solve with double parity is one less than with standard Yau with double parity, and PLL parity recognition is easier due to corners being solved.
Cons
* More moves during edge pairing.
* Good ability to see EO during lookahead is required, but this should be pretty natural for LEOR users already.
* In some cases, flipped ledges may accidentally be created but this should be avoidable with good lookahead.

Do you think this a good method ?

Inventer : Hồ Nguyễn Quốc Hưng​

Here are the thing why I think this method is "not good"
1. During edge pairing you will always require F moves (not B moves though).
2. A "lot" more moves during edge pairing since it is EO Edge pairing.
3. EO edge pairing isn't really a good step. Partial EOedge pairing might be good.
4. This one is my personal opinion, this is just another way to get to a solved 223 state after edge pairing on big cubes. There are many other methods like this. For example OBLBL, 4PB, TSSH.
This might be a good method but these are just my opinions
Also a request to anyone. Can anyone do a comparison of the big cube methods which when arriving at 3x3 stage have 223 solved (preferably on dl) just like how @Athefre did for 223 based methods on 3x3?

 

[Reirto-RRNF]

LEO4 METHOD​

Hi, I just had an idea a few days ago, about a 4x4 rubikscube method for LEOR solver.

Steps :
1.L/R Center
2.FirstBlock (Like Meyer)
3.F4C
4.Solve DF, DB dedge.
5.EO Pairing the remaining edges without messing up the block (Do this like Z4, but without messing up the 3x3x4 Block)
6.FLIP Final Badedge
7.Solve like LEOR or Petrus (RightBlock->LL+Parity PLL)

Pros
* Good ergonomics during edge pairing as no F/B moves are needed
* Final 3x3 phase is just RU blockbuilding followed by LL, no pause for inspection of EOSTRIP.
*Two Step is done after Reduction, resulting lower movecount for 3x3StAGE.
*Less cube rotation
*Better ways to deal with parity. The number of algs used in a solve with double parity is one less than with standard Yau with double parity, and PLL parity recognition is easier due to corners being solved.
Cons
* More moves during edge pairing.
* Good ability to see EO during lookahead is required, but this should be pretty natural for LEOR users already.
* In some cases, flipped ledges may accidentally be created but this should be avoidable with good lookahead.

Do you think this a good method ?

Inventer : Hồ Nguyễn Quốc Hưng​

It more look like it created for Petrus user than LEOR user because the 3x3x4 block is solved first and then the EOPairing

 

[jyhjyi]

JJ_Method (for NxNxN)
No “long” parity algorithm and flipping algorithm
no more complicated last 4 edges pairing

step1: solve 6 centers
step2: pair 8 edges containing white and yellow
step3: solve white and yellow faces
step4: solve the lower-semi layers with ”cross layer commutator”
step5: solve the upper-semi layers with “self layer commutator “
step6: adjustment the centers(if necessary)

Examples
#1

#2

#3

#4

 

[PapaSmurf]

Could you write a text tutorial? To be quite honest, I don't want to watch 4 10 minute plus videos with no talking to figure out what's actually happening. It could be cool though.

 

[TGWCubing]

2x2 Alg Set that does OLL while permuting the bottom layer

May sound bad at first and "just like EG" but really it's not. About 21 algs (not counting solved OLL and if you don't want any non-PLL PBL then try to avoid OLL skips when not having a fully permuted layer.

FAQ (from other places where I talked about this):
Q: Why don't you learn EG?
A: Too many algs for a begginer-brain

Q: Why don't you learn Ortega?
A: Diag-Adj / Adj-Diag is pain

Q: Isn't that like EG?
A: Ortega is too but no-one's complaining

Q: How is Ortega like EG?
A: Well you solve the bottom layer and permute both layers (like a 2-look EG)

 

[PapaSmurf]

Ortega algs are pretty good. The OLLs are better than these, and the PBLs are better than PLLs overall. If you're stuck for algs, just use Cyotheking's. They're all pretty good.

 

[hydynn]

Isn't that just LBL with a bit more freedom in making the first side and having to learn 21 algs? If you're able and willing to learn 21 algs, just learn CLL since its just 40 algs.

 

[TGWCubing]

Isn't that just LBL with a bit more freedom in making the first side and having to learn 21 algs? If you're able and willing to learn 21 algs, just learn CLL since its just 40 algs.

well I was thinking of an alternate way which was CLL then permute D layer

 

[OreKehStrah]

This is not worth it for sure. Just use normal LBL if you don’t like Ortega and learning algs. I’m able to get sub4 avgs using white bottom only lol

 

[GenTheThief]

May sound bad at first and "just like EG" but really it's not.

It really is.

FAQ (from other places where I talked about this):
Q: Why don't you learn EG?
A: Too many algs for a begginer-brain

LBL has like 9 algs. Sub3 potential ez
CLL has 40 algs and is 2 look. Sub2 potential if you can turn fast. And what's the point of 2x2 if you're not turning fast.
This is 20 algs and still 3 look. It also leaves you with adj and diag, which are both long algs.

Q: Why don't you learn Ortega?
A: Diag-Adj / Adj-Diag is pain

I will give this to you. I hate those algs. adj-adj is cracked though.
However, I will say that it's really just a

Spoiler

Skill Issue

Q: Isn't that like EG?
A: Ortega is too but no-one's complaining

The difference is that Ortega predates EG by like 25 years. And even when it was rediscovered, that was still 6 years before EG was developed.

Q: How is Ortega like EG?
A: Well you solve the bottom layer and permute both layers (like a 2-look EG)

You also need to orient the top face, so EG is 2 look (if you don't predict anything). Ortega is is 3 look (if you don't predict anything).


I would:
1. Just stick with LBL if you're really adverse to learning algs.
2. Learn CLL if you want to be fast without learning 100+ algs.
3. But I get it, it's fun to come up with methods and get clout for being a method inventor. I'll try not to clown on someone if they recognize the downsides of their method. But if you're gonna try to tout this as a legitimate beginner/intermediate/advanced method, you're gonna have to show why it's better than other methods.

Even so, 2x2 mostly comes down to having good inspection/planning, a good pickup, fast recognition, and good turning. Your method or how many algs you know isn't super super important. Lucas Etter set the 1.51 2x2 WR average without knowing full EG because he was also cracked at planning and had stupid good turning.

 

[fabdabs1234]

Really not viable, 3 look and you get diag/adj swap 5/6 solves = average 20+ moves
Ortega is much better as there are fewer and faster algs. Im not sure what you have against adj/diag diag/adj, but the algs are very fast (faster than pure diag).