[Help Thread] FMC Discussion

[irontwig]

If you're learning for FMC you might just learn them according to length. There are only 10 cases <9htm all of which are imo valuable for FMC. After that stuff like R U2 R D L' F2 L D' R2 can come in handy.

 

[marco.garsed]

what's NISS?

 

[irontwig]

It's when you turn a promising start that you can't find any continuation into to pre-moves for the inverse scramble:

S (Scramble) X (Nice start)
=>
X' S' F (Finish from inverse scramble)
=>
S' F X' (Solution for the inverse scramble)
=>
X F' S (Solution to regular scramble with pre-moves)
=>
S X F' (Corrected solution to regular scramble)

 

[Cubenovice]

A discription by Guus can be found further up in this thread:
http://www.speedsolving.com/forum/showthread.php?13599-The-FMC-thread/page19&highlight=FEWEST+MOVES

To understand how to properly apply NISS you should really walk through some NISS solves.
Start with analysing solves with a single switch, multiple switches are quite hard to follow.

Apart from this thread you may want to check out the weekly competitions on this forum and fmc.mustcube.net to find detailled FMC solutions.

 

[marco.garsed]

wow! seems cool! too many times happened that i could find a great start but a damn bad LL... I'll try it..

 

[irontwig]

It won't help with bad LLs; as the inverse LL case will be just as bad. You would have to invert your solution to (say) F2L-1 rather than the entire F2L.

 

[Cubenovice]

I just wanted to share this little Sune matrix routine that I seem to have develloped over time.

If EO gives me two adjecent solved edges I always apply the following Sune-matrix to check for a good LL:

S = Sune R U R’ U R U2 R’
S' = Inverse of Sune
AS = AntiSune R’ U’ R U’ R’ U2 R
AS' = Inverse of Antisune

So with two adjecent edges solved I check all four Sunes mentioned since they will all solve EP but have a different effect on corners.
The lines in the center indicate how to position the two adjecent edges at the start of the alg.

Anyone else doing something similar?

 

[marco.garsed]

what a lucky solve!

B' D2 F' R U F' U2 F2 U' R (not mine )
D' F' D F L D L'
R' D' R D' R' D2 R D

25 HTM

 

[guusrs]

It won't help with bad LLs; as the inverse LL case will be just as bad. You would have to invert your solution to (say) F2L-1 rather than the entire F2L.

I think if you aim on a sub 30 solution, the maximum of useful pre-scramble moves is ±15 moves (0.5 * 30) . If using more than 50% pre-moves regular and inverse solution converge too much, so they will have the same difficulty and using NISS is a waste of time. In about 30 NISS solutions I did so far my maximum number of pre-moves was 12!.

Gus

 

[Kenneth]

Alg useful for FMC, it even got room for a insert : F R E D L U . D

At the dot, insert N-PLL

 

[Kenneth]

Trondheim FMC : F U2 R2 D2 F R2 B' U2 F L2 D L' U R' B L' F D2 L' F' R

Lucky scramble?

I have not tried it yet but white looks pretty easy for start...

Tommy 'the lucky bastard' Holm had 33 using more or less speedsolving, Sune + PLL-skip for LL. (U L U B D2 B' F2 R D R F' R' (F2) U' R U2 R' U2 L' U' L2 U L2 U' L2 F' L' U' F U' F' U2 F U2)

hmm, (F2) must be z2

Coinman (the other Tommy, my bro) had this start : U L D' R' U D' F2 R' B R2 . F D2 B' but ended in DNF.

I placed the dot there, you can insert any D at that point.

 

[okayama]

Trondheim FMC : F U2 R2 D2 F R2 B' U2 F L2 D L' U R' B L' F D2 L' F' R

Lucky scramble?

Hmm, looks nice, but tricky...

Some of my tries:
[1]
2x2x2 block: U L U L2 F L2
2x2x3 block: R' D' R' F2
F2L minus 1 slot: R2 D' R'
Almost all: D' F D2 F2 R F R' D

[2]
1st 2x2x1 block: F2 U2 B
2nd 2x2x1 block: D
Make 2x2x3 block: L B' L' R2

[3]
Pre-scramble: U2 R2 D2

2x2x2 block: F2 D
2x2x3 block: U L U'
Orient edges: U' L2 U2 B' U'
Make tripod: B' L2 B L' B2 L B
All but 3 corners: L B' L' B
Correction: U2 * R2 D2

Insert at *: U L' U' R2 U L U' R2

Solution: F2 D U L U2 L2 U2 B' U' B' L2 B L' B2 L B L B' L' B U' L' U' R2 U L U' D2 (28 HTM)

[4]
Pre-scramble: L2 R2 D2

2x2x2 block: F2 D
2x2x3 block: U2 B L2
Orient edges: B L' R B L R2 U' R
All but 3 corners: U' B U2 B' U' B U B'

[5]
Pre-scramble: L R2 D2

2x2x2 block: F2 D
F2L minus 1 slot: B U2 B L B L2
More 2x2x1 block: L U' R' F' U' F
All but 3 corners: U L' U R U' B U B'

[6]
(For inverse scramble)
Pre-scramble: R B' R' L2 B' U2 D' F2

All but 3c3e: D' R' D R' L2 U (or D2 L2 R U2 R U' also leaves 3c3e)
Correction: R B' R' L' + L' B' U2 D' * F2

Insert at *: D B2 D' F D B2 D' F'
Insert at +: F E F2 E' F

Solution: F' D B2 D' F' D B2 U2 B L F' E F2 E' F' L R B R' U' L2 R D' R D (27 HTM)

Might getting Sub 25 be not easy?

 

[Kenneth]

Yes, sometimes there are too many good blocks to start from, trying to preserve all makes many moves.

I did a cube explorer search on Holm's double xx-cross, he had 16 from there and the optimal solution is 14. Not bad =)

Forgot to say, Morten had the winning solution on the scramble (31), but I have not got it.

Edit, I just asked him, he will post it...

Mortens FMC solution:

scramble: F U2 R2 D2 F R2 B' U2 F L2 D L' U R' B L' F D2 L' F' R
x-cross: R' B R B2 L F' L' R U' R' Uw'
F2L: F' U' B' U' B U2 F2 U' F' U2 F' U' F
LL: R' F' U' F U R U2

Well, the only LL that beats this is a compleate skip

 

[guusrs]

I tried and got 28 moves, not "that easy" scramble!

 

[Kenneth]

LS coms

I don't know if this idea has been up before but.. While doing the last slot, you got the pair readymade in the U-layer and it is time to put it down. Then use a 3-cycle 1x1x2 block commutator to solve the pair and affect the last layer. Two of the blocks are fixed, the place where the pair is and where it shall go, but the last one is free, so you can alter the com depending on the current case.

A nice situation is all EO and two of the edges sitting in U are relativly solved, then cycle the pair and the other two edges and you will have LS and all edges solved in 7-9 moves (or better if you can cancle some). Example L' Dw2 L U L' Dw2 L

Use setups like Bw to orient 2 edges while cycling. (edges are the easiest to control using this style but of course also corners works).

In one attempt of about 25 where I tried this I got a compleate LL-skip from solving LS (I forced edges so it was a 1:324 corner skip)

 

[irontwig]

Yeah sure sometimes one can use a pair cycle to leave three corners, and sometimes you get really lucky:
http://www.speedsolving.com/forum/s...lve-quot-Game!&p=535454&viewfull=1#post535454

 

[Stini]

Pair commutators are pretty common in Heise, but I almost never remember to try those in FMC though. I tend to use pair commutators only when they solve the whole cube.

Edit: Hmm, it just came to me that you could use pair commutators to solve some non-obvious cases with a setup move, for example U' [L F L', D'] U. I wonder why I haven't thought of that before, it's pretty obvious actually. Recognition isn't that bad because the setup move must make a pair and you must also have another pair and a pseudo pair. I'm somewhat used to recognize pseudo pairs without NISS anyway so it's not a problem.

 

[Cubenovice]

I just walked through irontwigs 24 mover linked to above, I wouldn't have know what to do...
It's about time I look into pair commutators, so it's back to Ryan's page!

Stini,
what's keeping you from participating in the weekly comps?

 

[Stini]

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.

 

[Kenneth]

Simple speaking, put a 1x2 in a F2L slot that is not the one you are working on (and do it in such a way you don't move the goal slot, if it is FR don't do R or F).

Example: start with a L and then move in the pair to swap to this position using d', remove its content from F2L using L' and put another 1x2 down using U' L. Then move the 'virtual slot' back to it's place using d and restore setups (in this case it is just a L' )

L d' L' U' L d L' ... the same using single face moves : L U' F' U' F U L'

But you can also swap diagonal using d2, like this : L d2 L' U' L d2 L' ... this is the fast way because moving to single faces you get L U2 R' U R U2 L'

--------
Also other types of comutators works, but mabye not as intuitive as the simple method above...

R U Lw U' R' U Lw' [U'] ... no explain, but think Niklas =)

--------
Pretty one : Bw L' d2 L U L' d2 L U' Bw' ...

--------
Edit: Aproach for linear FMC, block build F2L up to LS but ignore if one of the pairs are in the wrong slot. Then, it this happen, pair up the pair that should be where that one is (improve LL first) and follow it up with a 3-cycle to compleate F2L.

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.

Solution (one attempt, got 29 moves, don't peek until you tried =)

Spoiler

Pair Up : U R U2 R' (4)
Improve LL (EO) : F U F' (3, 7)
Pairs : d' R U' R' d R U R' (8, 15)

To reach ZBLL

LL :
Sune minus one move : B U B' U B U2 B' (6, 21)
Niklas : B2 R B L' B' R' B L (8, 29)