[Help Thread] FMC Discussion

[okayama]

Scramble : F' U2 F2 R' U F' R2 F' L F L2 U2 R' U' L (15f*) ... setup XX-cross + one wrongly solved pair, the rest is random.

Spoiler

How about

• All but 3 c/e pairs: U' R U' B F U' B' U F' U' R' U' , or
• All but 3 corners: B U2 B2 R B U' R2 U2 R U ?

Insertions are left to readers.

 

[mrCage]

Cubenovice: I just haven't felt like doing FMC lately, it feels more like chore atm.

Btw pair commutators aren't that difficult really, first you have to notice that pairs can be interchangeable by a face turn or by a double layer turn. I guess face turn is more common in practice. So you must have two interchangeable pairs and then you just insert the third pair like with corner commutators. In irontwig's solution the two interchangeable pairs are on the F-face. He inserts the third pair with L b2 L'. Notice how nothing else changes on the F-face except one pair is replaced by another. You could solve the case like F [D' R' D, F] F', where the insertion D' R' D is a bit different, but that's one move longer because you first setup to an 8-move case with F (one move is cancelled in the end). One example of a pair commutator where the pairs are interchangeable by a double layer turn could be [R U' R', d]. I haven't really experimented much with pair commutators, so there might be some tricky cases, but usually I have no problem figuring out how to solve them.

Pair commutators are actually quite trivial. They can be done on corners, edges or ce-pairs for that matter. One can achieve 3-cycle or double swap (or others with more experience...)

Per

 

[Brest]

Greetings FMC pros.

I have read about half of this thread and, while I'm not close to understanding exactly how the more advanced techniques work, I am beginning to grasp the concepts involved. The 1st time I formed a pseudo block and then figured out the pre-move to make it an actual block, I began to understand why that technique works. Before this I just knew that it worked because I had read that it did.

I can block build, but not during a speed solve and not close to using optimal moves. Unless a block is very simple to build, I find it hard to "see" how to bring the pieces together. This could be because I am primarily a CFOP solver, so I am used to breaking the parts down to cross +pairs. I've looked at Petrus and Roux, and understand the former far more than the latter.

Are there any specific ways to practice block building?
Do you have any advise on how to break the CFOP habit of pairing?

-*-*-*-*-*-

Here is a recent attempt I made:

B2 F' U L' D2 R2 L' B2 U B F L F2 R' L B' D' B2 L' U L' B2 L' U2 B2

U B' U F L F' R' F' D F' R
L' F L F'
D' B D B' D L' D' L2 D L'
D B' L B L'
D L D L' D L D2 L'

I don't yet understand how to complete the F2L to end up with a mostly finished LL. This solve was as close as any I've had, and it took me a long time to find that ending, and even longer to find that start. I haven't yet got to the stage of trying to insert a cycle into the solve to generate cancellations.

-*-*-*-*-*-

Here is a 5 move XX-cross & 8 move 3rd+4th pair:

F2 U' B2 D' L2 D B2 U R2 U2 L2 B L' D2 F D2 R' B' F' R B

F' R B L' R2
D' R D L D L' D' R'

I couldn't find a great way to finish the F2L, but it would be great to see what can be done with this scramble.

-*-*-*-*-*-

Even though I am struggling to comprehend the advanced techniques you guys use, I do find the FM challenge interesting. Most of the solves I've read in this thread are amazing and I look forward to finish reading the 2nd half.

I will appreciate any advice and teaching given. Thanks for your time.

 

[irontwig]

You could have continued that five move double x-cross with F' U' R' B R B' U D2 F D2 which leaves three corners and edges after 15 moves. It's basically just keyhole with some luckiness.

 

[Kenneth]

Greetings FMC pros. A.S.O....

Just practice block building, skill comes from experience.

Nice endings is a combination of trying several ways from the first blocks, using moves you now will improve the situation and pure luck.

Edit, nice ends... also from knowing the optimal alg that solves the case, the more algs you know the more cases are nice =)

 

[okayama]

-*-*-*-*-*-

Here is a 5 move XX-cross & 8 move 3rd+4th pair:

F2 U' B2 D' L2 D B2 U R2 U2 L2 B L' D2 F D2 R' B' F' R B

F' R B L' R2
D' R D L D L' D' R'

I couldn't find a great way to finish the F2L, but it would be great to see what can be done with this scramble.

-*-*-*-*-*-

Solution: L2 F' R' F L2 F' R B' R' B R B R B R B R D2 R' B' L' D2 (22 HTM)

Pre-scramble: L' D

2x2x2 block: F'
More 2x2x1 block: B' R' B
Expand to 3x2x1: R B
F2L minus 1 slot: R B R
More 2x2x1 block: B R D2 R' B'
All but 3 corners: L' D L
Premoves correction: L' D

Insert at the beginning: L2 F' R' F L2 F' R F

 

[Cubenovice]

Hi Brest, good to see you here!
Don't forget to check the other link too

In addition to the other proposals:
You could aslo sneak in some extra moves into the X cross to get better EO
F' R B L' F' R' D2 F R'
D' B' D' B L D' L'
now you have various options:
L' F L F' leaves 2 EO that you can check for OLL: Fat Antisune leaves G-perm...
D L D' L' gives Sune but bad edge permutation
D2 L D2 L' gives two adjecent edges. From here you can try 4 different sunes that will permute edges
Antisune L' D' L D' L' D2 L plus D leaves three corners after 25 moves
solve corners:
Direct solve: B' D R D' L2 D R' D' L2 B = 35 HTM

insertions:
label the remaining three cycle with small stickers 1, 2 and 3
solve, scramble ans walk move by move through the solution to check for basic 8 move commutators
at end of scramble -
F' here's one: U' D2 U R U' D2 U R' cancels 2 moves R' with the R that is coming up
R -
B -
L' another one, F' D F U2 F' D' F U2
F' another one, L U' L' D' L U L' D
R' another one, B' U2 B D' B' U2 B D cancels 1 move with the upcoming D2
etc etc

If I take the one after the initial F' I have already improved my solution by 4 moves!
Instead of 10 moves for set up - comm -undo setup I have placed the corners in 6 moves.

Offcourse I do not always write down the ful commutators, If I see a 3 cycle I scan the start and end move.
Only if they start / end with the pre- / post-comm moves I will write them down.
Exception: the very first 8 mover I do write down just in case I do not find cancellations further down the solve.




R B L' F' R'

 

[Brest]

Many replies. Sweet!

You could have continued that five move double x-cross with F' U' R' B R B' U D2 F D2 which leaves three corners and edges after 15 moves. It's basically just keyhole with some luckiness.

Nice! I've come to expect intriguing solves / sequences / insertions from you, and you didn't let me down. I like how you say it's "basically just keyhole". Basic, but in a super-advanced way! :tu

How would you solve those last two 3-cycles? Try to find insertions with cancellations?

Just practice block building, skill comes from experience.

Nice endings is a combination of trying several ways from the first blocks, using moves you now will improve the situation and pure luck.

Edit, nice ends... also from knowing the optimal alg that solves the case, the more algs you know the more cases are nice =)

Well, I know the mantra of *Practice Practice Practice* however I was hoping there might be specific ways to help with block building. You are all far more advanced than I am, so when you were learning did you find ways which helped your understanding of block building?

When searching for nice endings, do you find that you can predict what certain sequences will produce? Or is it more a case of "try as much as you can" until you find something good? Hit it with a hammer until it fits!

Solution: L2 F' R' F L2 F' R B' R' B R B R B R B R D2 R' B' L' D2 (22 HTM)

Pre-scramble: L' D

2x2x2 block: F'
More 2x2x1 block: B' R' B
Expand to 3x2x1: R B
F2L minus 1 slot: R B R
More 2x2x1 block: B R D2 R' B'
All but 3 corners: L' D L
Premoves correction: L' D

Insert at the beginning: L2 F' R' F L2 F' R F

OMG! That is epic. I mean this as a compliment, but are you some kind of robot?!?
I've just read the part of this thread where you first joined the forums and I have really enjoyed studying your solves. Although I don't completely understand everything you do, it sure does help to see your solves.

Hi Brest, good to see you here!
Don't forget to check the other link too

In addition to the other proposals:
You could aslo sneak in some extra moves into the X cross to get better EO
F' R B L' F' R' D2 F R'
D' B' D' B L D' L'
now you have various options:
L' F L F' leaves 2 EO that you can check for OLL: Fat Antisune leaves G-perm...
D L D' L' gives Sune but bad edge permutation
D2 L D2 L' gives two adjecent edges. From here you can try 4 different sunes that will permute edges
Antisune L' D' L D' L' D2 L plus D leaves three corners after 25 moves
solve corners:
Direct solve: B' D R D' L2 D R' D' L2 B = 35 HTM

insertions:
label the remaining three cycle with small stickers 1, 2 and 3
solve, scramble ans walk move by move through the solution to check for basic 8 move commutators
at end of scramble -
F' here's one: U' D2 U R U' D2 U R' cancels 2 moves R' with the R that is coming up
R -
B -
L' another one, F' D F U2 F' D' F U2
F' another one, L U' L' D' L U L' D
R' another one, B' U2 B D' B' U2 B D cancels 1 move with the upcoming D2
etc etc

If I take the one after the initial F' I have already improved my solution by 4 moves!
Instead of 10 moves for set up - comm -undo setup I have placed the corners in 6 moves.

Offcourse I do not always write down the ful commutators, If I see a 3 cycle I scan the start and end move.
Only if they start / end with the pre- / post-comm moves I will write them down.
Exception: the very first 8 mover I do write down just in case I do not find cancellations further down the solve.




R B L' F' R'

Dude, thank you! I was going to ask for a brief "how to" on inserting cycles, and you've preempted my question. When using stickers, how do you apply them to fix orientation as well as permutation?

Although I know what they are, I'm not proficient at commutators. I can see how they work, and I understand the way they are constructed, what I don't understand is how you know what sequence to use for a given case. Does this come down to knowing how each commutator effects the cube, like any algorithm? Or is it more a case of complete understanding will give you the knowledge to apply a comm for any case, without knowing it as a specific algorithm?

 

[Cubenovice]

When using stickers, how do you apply them to fix orientation as well as permutation?

In my example I had three corners left:
orange-green-yellow
green-red-yellow
blue-orange-yellow
the above order is random, also started the "name" with a random color, do not ready anything into it

Take a random sticker on one of the corners:
I picked orange on the orange-green-yellow cubie, I put a small tag on it with no 1
I know where this sticker should be on a solved cube.
On the sticker that is now occupying that spot I place a tag no 2 (this is the yellow sticker on the blue-orange-yellow cubie)
I look where no 2 is supposed to be: that sticker gets tag no 3 (this is the yellow sticker on the green-red-yellow cubie)

Note: with tag I mean I put a real sticker with 1, 2, 3 onto the cube's stickers.
This way you can easily keep track of the pieces when you walk trhough your solution.

Some important notes: always check if your cycle is closed
Check where the last tag needs to go, you should find no1 occupying this spot!
If this is not the case there are several options:
- you stickered wrong
- there is one or more corners twisted in place
- there are corners swapped
- the last two mean that you will have an additional cycle to solve

Does this come down to knowing how each commutator effects the cube, like any algorithm? Or is it more a case of complete understanding will give you the knowledge to apply a comm for any case, without knowing it as a specific algorithm?

Both are the same ;-)
If you understand comms you know how they affect the cube and thus can apply them to any case.
A great resource on comms is the BH tutorial thread in the BLD forum.
The basic 8 mover is typically enough. I cannot identify all the comm types mentioned in the tutorial but I can solve all of them with the 8 mover and set ups.

And the good thing about working with insertions is that it is almost impossible not to have an 8 mover in there.
But the fun really starts when moves are cancelled.
My record is 5 moves cancel in an 8 mover.
Some people have even had "free" insertions, cancelling the full 8 moves!

Commutator:
you do something,
you do something else,
you do the inverse of something,
you do the inverse of something else
and then you have done something...

 

[Kenneth]

Well, I know the mantra of *Practice Practice Practice* however I was hoping there might be specific ways to help with block building. You are all far more advanced than I am, so when you were learning did you find ways which helped your understanding of block building?

When searching for nice endings, do you find that you can predict what certain sequences will produce? Or is it more a case of "try as much as you can" until you find something good? Hit it with a hammer until it fits!

Many FMC newbies reads Heise to learn about blocks : http://www.ryanheise.com/cube/heise_method.html

To me making a block is most often a multi step job, first I do a 1x2 block of a corner and a edge, then the same but a edge and a centre that fits to the first block for a 2x2 and another edge-centre if it is a 2x2x2. When the blocks are done I put them together but not just like that, I try diffrent ways (if possible) to improve the situation for the next block (usally a 2x2 to exapand to 2x2x3).

I 'cheat' a bit, only seldomly I start from anything but yellow. So recognition is easy because I work in my standard colour. If I do like green, then it is much harder to spot the cases. The drawback is that it is not always the best side to start from. If there are no blocks in the scramble it does not make much difference thought...

If you have a good skeleton but a crappy ending, then you can step solve the blocks and look for a 'empty face'. An empty face is a side that does not contain any of the pices that belongs to the blocks, so you can turn it freely. That way you can have three ends that are diffrent from the first with just one extra move. If this is not possible the next option is to make such a side, but more moves then.

----
There is a lot of hammering

But of course, you force a lot on EO (fliped edges in LL usally sucks), a bit on EP but not that much on the corners. Three orientations makes corners much harder to control than edges, and also, solving the corners last is easier than solving the edges last (insertions).

As a CFOP user you know what R' F R F' does to LL edges compared to U R U' R' when solving the last pair. This is the most basic ZBF2L but there are nearly 500 cases, and the solutions are often intuitive.

ZBF2L; forming - placing the last pair and orient LL-edges, all in one go.

 

[Cubenovice]

From the weekly 10:

3x3x3 Fewest Moves
1. B R B' D F' R2 B' L' B2 L B L' U2 L B U2

On Inverse scramble
B R B' D' F' # R2 makes a 2x2x3 minus 1corner 6
L' B' U2 B2 L B' U2 makes F2L + EO minus 2 corners 13
B' L' B L' @ B' L2 B L' Antisune permutes two edges and solves a corner, ALF leaves 5 corners 21

At # insert F R F' L F R' F' L' to cancel 3 moves, note that I have to reverse the order of R2 L' in the original solution to cancel the 3rd move L' L' into L2
At @ insert L B L' F2 L B' L' F2 to cancel 3 moves

B R B' D' F' F R F' L F R' F' L' R2 L' B' U2 B2 L B' U2 B' L' B L' L B L' F2 L B' L' F2 B' L2 B L'

gives B R B' D' R F' L F R' F' L2 R2 B' U2 B2 L B' U2 B' L' B2 L' F2 L B' L' F2 B' L2 B L' as 31 HTM solution for the inverse scramble

Final solution:
L B' L2 B F2 L B L' F2 L B2 L B U2 B L' B2 U2 B R2 L2 F R F' L' F R' D B R' B'


So, Okyama mentioned the similarity of my start with the scramble.

I didn't even notice but he's definately got a point.

The initial six moves (apart from the 4th) are indeed identical to the scramble but how can one NOT see the three 2x1 blocks that make the 2x2x2?

Most of the time I actually do not even look at the cube during scrambling so I can honestly say that I did not see the block near the end of the scramble.

What do the experts in this thread think about this?
If this would give me a DNF or DQ in an official event I'll ask Mats to change the reults.

In a competition, if you noticed a similar situation, would you deliberately take a different start?

 

[marco.garsed]

insertionts rulez!

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



Solve:

F R F' U L'
D U2 L
D2 F D' U2 L U R U' L'
D2 M' D2 L'
x' R2 D2 R2
B R D L' D' L
D'

Explanation:

F R F' U L'
D U2 L
D2 F D' U' [R] ([R'] [U'] L U R U' L' ) [U']
(D2 M' D2 L' [R] x')
R D2 R'
[R'] B R D L' D' L
D'


32 HTM


I think that I could find something better changing my start but I love how the insertions give me a 8 moves LL!

 

[irontwig]

I'm afraid that the problem with solutions being somewhat similar to scrambles can only be solved with long scrambles. I think 30 random moves that Fredlund uses are resonable in both length and randomness (i.e. very close to random state scrambles). His 10 minute scrambles seem to be quite hand picked though.

 

[guusrs]

From the weekly 10:

3x3x3 Fewest Moves
1. B R B' D F' R2 B' L' B2 L B L' U2 L B U2

On Inverse scramble
B R B' D' F' # R2 makes a 2x2x3 minus 1corner 6
L' B' U2 B2 L B' U2 makes F2L + EO minus 2 corners 13
B' L' B L' @ B' L2 B L' Antisune permutes two edges and solves a corner, ALF leaves 5 corners 21

At # insert F R F' L F R' F' L' to cancel 3 moves, note that I have to reverse the order of R2 L' in the original solution to cancel the 3rd move L' L' into L2
At @ insert L B L' F2 L B' L' F2 to cancel 3 moves

B R B' D' F' F R F' L F R' F' L' R2 L' B' U2 B2 L B' U2 B' L' B L' L B L' F2 L B' L' F2 B' L2 B L'

gives B R B' D' R F' L F R' F' L2 R2 B' U2 B2 L B' U2 B' L' B2 L' F2 L B' L' F2 B' L2 B L' as 31 HTM solution for the inverse scramble

Final solution:
L B' L2 B F2 L B L' F2 L B2 L B U2 B L' B2 U2 B R2 L2 F R F' L' F R' D B R' B'


So, Okyama mentioned the similarity of my start with the scramble.

I didn't even notice but he's definately got a point.

The initial six moves (apart from the 4th) are indeed identical to the scramble but how can one NOT see the three 2x1 blocks that make the 2x2x2?

Most of the time I actually do not even look at the cube during scrambling so I can honestly say that I did not see the block near the end of the scramble.

What do the experts in this thread think about this?
If this would give me a DNF or DQ in an official event I'll ask Mats to change the reults.

In a competition, if you noticed a similar situation, would you deliberately take a different start?

Hi,

As long as you can explain how you came up with those moves its allowed to have some moves similarity with the scramble.

If a jury asks you about it, you will answer:

B R B' D' F' # R2 makes a 2x2x3 minus 1corner 6

Which make it acceptable you did not make use of the scramble and your solve would be legal.

Gus

 

[irontwig]

Well this solution is "explained" but it's just the same inverse scramble with two corner cycles:

L2 B R B' [Square]
L2 D F' R2 B' [X-cross]
L' B2 L B L' U2 L B [F2L]
L' B L' F2 L B' L' F2 L2 U2 [LL]

Note that I'm not saying that Cubenovice was cheating I'm just pointing out problems with the current regulations.

 

[marco.garsed]

I think it's also a problem of short scramble... inverse scramble, with 18-19 moves scramble, is often the optimal solution and in FMC we try to get as close as possible to this one..

 

[okayama]

@Cubenovice: I didn't intend to disqualify your solution, but just noticed the similarity.
I also saw B R B' for making a 2x2x1 block and actually used in my tries.

I'm afraid that the problem with solutions being somewhat similar to scrambles can only be solved with long scrambles. I think 30 random moves that Fredlund uses are resonable in both length and randomness (i.e. very close to random state scrambles). His 10 minute scrambles seem to be quite hand picked though.

Well this solution is "explained" but it's just the same inverse scramble with two corner cycles:

L2 B R B' [Square]
L2 D F' R2 B' [X-cross]
L' B2 L B L' U2 L B [F2L]
L' B L' F2 L B' L' F2 L2 U2 [LL]

Note that I'm not saying that Cubenovice was cheating I'm just pointing out problems with the current regulations.

I'm of the same opinion as Jernqvist.
Even if you make a solution based on the scramble, sometimes you can (try to) explain the solution.
See also this post. I made the solution based on the scramble, but I explained somehow.
There is no doubt explanation is important to prove your innocence, but explanation doesn't automatically mean the innocence.

 

[Cubenovice]

Thanks for your reply!
I thought I remembered a previous example of a scramble based solution but could not find it.

I agree that slightly longer scrambles could be a means of reducing the possible impact of scramble based solutions.
Afterall, a sub 30 solution in a competition wil get you a top 33 world ranking.

This seems very (too...) feasible by scramble based solutions now that the WCA is using optimal scrambles.

Could the scrambling software the WCA uses be programmed as such that the scrambled state is actually random but the scramble to get there is relatively long.

 

[Lucas Garron]

Could the scrambling software the WCA uses be programmed as such that the scrambled state is actually random but the scramble to get there is relatively long.

Interesting idea, but how about suggesting this somewhere where it can be taken seriously instead of being overlooked?

 

[Kenneth]

[wiki]LPELL[/wiki]

A method I made mostly for FMC. I will fix the mirror cases later (done now) and some day I may do all (2x48) cases where the edges are unoriented.