[Help Thread] FMC Discussion

[Stini]

Scramble weekly comp 32: B2 L' R F' U F' L' R2 U B2 R2 B' R2 U' L2 U2 L2

L' R' U' R' B F2 EOLine 6/6
R D2 U2 L U2 R' D2 L' D2 L' 1st 1x2x3 block 10/16
D' R D R D' R 2nd 1x2x3 block 6/22
R' B R F' R' B' R F OLL 8/30
F' D2 F' D' L' F' L2 D' L' D L' F L D' F PLL 15/45

4 moves cancelled = 41 moves

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

Et voila, my first attempt at Fewest Moves.

The actual solution was found within the hour but I had quite a hard time in rewriting the algs for cube rotations and double layer turns.
Especially the F perm was messing me up all the time.

Feel free to comment on my (lack of) blockbuilding skills.
After making the EOLine there was this red/white center-edge pair waiting to be used but I could just not find a way to use it???

Over the next few days I am going to work through some of the the solves for this scramble in the weekly comp thread to see how the experts do this.
I have already read Arnaud's post on FMC and I see I have still quite a lot to learn.
Think I will start learning commutators so I can work with insertions without having to learn too many algs.

Not bad for a first attempt! Personally I think that (pure) ZZ is not really ideal for fewest moves though. Sure it might occasionally bring good results, but correct EO doesn't really guarantee a short solution. Also I find that the line is just on my way, restricting my solve too much. Often it takes less moves to build a block if you allow yourself to flip edges, and also you can often combine EO and block building. If you really want to do the EO-line, perhaps you could try to combine some more block building with it? That way you wouldn't "waste" moves for solving just EO.
I usually start with a 2x2x3, which takes less than 10 moves on average. I might then orient some edges if I can't see a good continuation and if the EO-step contributes to the block building (for example creates pairs). Otherwise I delay the EO even further, just not to restrict myself to certain moves too early.

Other than that, you seem quite good at block building! Only 6 moves for the second block is really nice! The first block isn't bad at all either, but perhaps you can see that the line is restricting you a bit? If you just had a better LL, that would have been a great solution. Generally you want to avoid OLL+PLL and either try something different during F2L to get a nicer LL or you could try some short algorithm (usually T-OLL or Sune) to solve the edges and leave 3 unpermuted corners, which you can then solve with a commutator insertion. In this particular solve you have two opposite edges solved, which isn't that nice (it takes 10 moves just to solve the edges), but on the other hand your corners are permuted so you might be able to find some ending with just 2-gen algs like Sune and Bruno just to avoid the F-perm. One example I came up with is: (finish 2nd block) D' R D R D R D R' D R / (Sune) R' D2 R D R' D R / (Sune) F' D' F D' F' D2 F D. Not terribly efficient, but that's just an example of the idea.

Keep practicing and learning more tricks, I'm sure you'll get under 35 moves in no time!

 

[hawkmp4]

ZB's solution works for me.
How do you guys practice for FMC? More specifically, if you're going to practice a full solve, do you do it under competition conditions?

 

[eastamazonantidote]

ZB's solution works for me.

He updated it. I think it should count because he didn't spend the full hour on it, but that's his call.

How do you guys practice for FMC? More specifically, if you're going to practice a full solve, do you do it under competition conditions?

I've never done competition nor am I very experienced in FMC, but I think I can chime in here. I practice many method openings: Roux, Petrus, ZZ, and Edges First just for the fun of it. I started seeing the cube differently and my speed solves became more advanced (multislotting, X-Crossing, etc.), though my times didn't drop that much.

I started with just pure Petrus blockbuilding, netting me two 42 move solutions to start my career. After your F2L is done, though, you have to look at the last layer. My first 2 tries I went linearly into the last slot, then evaluated my options, forcing a PLL skip both times (once without EO and once with). After that I started looking into premoves, though still baffled at how people knew what to do. It was then that I discovered NISS. Only then did I start practicing (because now it was fun!).

What I do is take the weekly scramble, do my solve, then observe the solutions of better and more experienced FMCers. I learn methods from them (leaving 3 corners, better blockbuilding, etc.). I probably spend 4 hours or so on the weekly comps. 1 hour on my solution, then 1 or 2 looking at other solutions. Then I spend another 1 or 2 trying to solve the scramble a different way, forcing myself to use something I'd just learned (usually not very successful but I like the challenge). Recently I've put a lot of work into learning about commutators and insertions, as it would've dropped my solution this week to 24 (Chris pointed this out to me).

 

[Tim Major]

ZB's solution works for me.

He updated it. I think it should count because he didn't spend the full hour on it, but that's his call.

In a comp if you spend 25 minutes on a solution, only for the judge to tell you it doesn't work, you can't then fix it and resubmit it just because you didn't use up the full time. Stini informed me on irc that my solution didn't work, so for that reason it isn't fair to count it (unlucky that I had the correct solution written down on paper :/)
I recounted my solution, and was surprised that it was 32 moves, when I was sure it was 34 moves lol. I'd left 2 moves out D:
And it's not my call, it's Mats' call And I'm sure Mats wouldn't count it unfortunately (it's meant to be run, like an official comp I suppose. Official rules anyway)

 

[ianini]

Okay... so I'm quite possible the n00best person in FMC but I just did one solve and I found this solution in maybe 5 minutes.

Scramble: R2 L2 D2 L D' U' B' F' L' U B2 D B' L F2 D' U2 L' U2 L2 F B2 U' L' U'
Solution: (38 moves)
Cross: x2 R' D F L
F2L: U' R' U R2 U R'
F2L #2: y' U2 L' U' L
F2L #3: R U R' L U L'
F2L #4: U R U' R' U R U' R'
OLL: R U R' U R' F R F' U2 R' F R F'
PLL: Skip

Any tips are extremely welcome as for I am very interested in FMC.

I also had a few questions:

1.) What is the difference between the lower bound and the upper bound?
2.) Why do people use pre-moves before the scramble?

 

[irontwig]

2.) Why do people use pre-moves before the scramble?

Say you find this block:
http://cube.crider.co.uk/visualcube.php?size=150&fmt=png&fc=llllyylyyoolrrlllllgglbblll
If you do "U2" before the scramble it will look like a normal 2x2x2. It's just a way to get more options for blocks.

 

[Tim Major]

2.) Why do people use pre-moves before the scramble?

Say you find this block:
http://cube.crider.co.uk/visualcube.php?size=150&fmt=png&fc=llllyylyyoolrrlllllgglbblll
If you do "U2" before the scramble it will look like a normal 2x2x2. It's just a way to get more options for blocks.

How do you implement it in a solve? If you do U2 scramble, then solution, how do you make the solution work for the original scramble?

 

[irontwig]

2.) Why do people use pre-moves before the scramble?

Say you find this block:
http://cube.crider.co.uk/visualcube.php?size=150&fmt=png&fc=llllyylyyoolrrlllllgglbblll
If you do "U2" before the scramble it will look like a normal 2x2x2. It's just a way to get more options for blocks.

How do you implement it in a solve? If you do U2 scramble, then solution, how do you make the solution work for the original scramble?

You just undo the premoves in the end of the solution.

 

[trying-to-speedcube...]

2.) Why do people use pre-moves before the scramble?

Say you find this block:
http://cube.crider.co.uk/visualcube.php?size=150&fmt=png&fc=llllyylyyoolrrlllllgglbblll
If you do "U2" before the scramble it will look like a normal 2x2x2. It's just a way to get more options for blocks.

How do you implement it in a solve? If you do U2 scramble, then solution, how do you make the solution work for the original scramble?

You just undo the premoves in the end of the solution.

You *do* the premoves, you don't undo them.

 

[Cubenovice]

Hi Teemu,

thank you for your feedback.
I chose the ZZ approach because it should give a lower move count vs straight forward CFOP (my main method).

I understand that it is not a typical FMC method but it has tought me already (a little bit) about block building. So much to learn...

 

[irontwig]

Couldn't be bothered to wait for a new scramble from Mats or Per, so I generated one with Cube Explorer and got a PB, yay!

Scramble: U F2 R2 F R2 D B U B2 F R F D F' U' L' D
Solution: F R L F' B2 L' B U L D' U' R' U2 B2 U R B R' B' R2 B R B' R' D' R' U' R D R' U (31)

Outline:
F R L F' B2 L' B U L D' U2 [2x2x3] (11)
U R' U2 B2 U [Triple X-cross] (15)
R B R' B' R2 B R B' R2 [Everything but last three corners] (24)
R D' R' U' R D R' U [Last three corners, simple commutator] (31)

Pretty bad start, but good continuation. Found in about 30 minutes.

 

[Tim Major]

Scramble: B' D2 L2 U' R' U2 L D B R' D' U2 L' D2 R2 F' L B' L F D R' D2 F U'
Solution so far: z2 F' R' D' x U R U' F U' B U x' y' R U2 R' D F' U' F y2 D F' U F D y

Now this is what I was trying after that.
R U' R2 F D' F D

And I'm left with 2 edges and 4 corners. How would I go about solving this? Insertions? If so, how?
I don't have much time before I need to go, but I was originally having D' R' B2 as premoves and keyholing to make f2l (not effieciently, just messing around with different methods of doing FMC)

So how would "a pro" finish this solution?
z2 F' R' D' x U R U' F U' B U x' y' R U2 R' D F' U' F y2 D F' U F D y (and possible R U' R2 F D' F D)
Thanks.

 

[okayama]

I couldn't find any 8-move corner 3-cycle insertion in the following frame.

Scramble: L B2 F D2 F' R2 D' U B2 L U' R D F' U L R2 B2 D' B' D' L' R' F' L'
Solution: F2 R F' L2 F R' F' L2 F' D2 L' F D2 L' R' F' R U' R D B2 L' D' L B R D R' D' B' D B D (33 HTM)

1st c/e pair: D2 L'
2x2x1 block (a): F D2
2x2x1 block (b): L'
2x2x3 block: R' F' R U' R
F2L minus 1 slot: D B'
More c/e pair: B' L' D' L B
Orient edges: R D R'
All but 3 corners: D' B' D B D

Insert at the beginning: F2 (R F' L2 F R' F' L2 F) F2

Could anyone find another good insertion?

 

[Cubenovice]

Weekly 34 scramble D' L2 D' R B' R' F2 R' B2 F2 D2 R2 B L' U B

I was trying to solve this with a ZZ approach and upon my fourth pair I came across something I found very interesting.
With a simple U perm I could get the cube solved apart from four cubies twisted in their correct spot.

But could not put it to good use... Is there really something there or does it just look nice and that's it?

z2 B2 D2 R F' D2 L' D' EOLine 7
L' U' L R' U R 1st pair 13
U2 L U2 R U R' 2nd pair 19
U L U' L' U2 L' 3rd pair 25
Nasty remaining corner edge pair...
All ways of inserting it correctly where too long and resulted in long OLL's

So I just "hammered" it in with U' F' L F L' 30
Now a CW U perm (11) permutes the edges and I am left with 2 twisted corners and 2 flipped edges in their correct positions.

(The edge cycle of the U perm can be inserted in various positions in the solve to check for cancellations)

I find the resulting state looks really interesting but is there a way to fix this with a commutator?

 

[DavidWoner]

Now a CW U perm (11)

11? 9 please F2 U R' L F2 R L' U F2

 

[CharlesOBlack]

Weekly 34 scramble D' L2 D' R B' R' F2 R' B2 F2 D2 R2 B L' U B

I was trying to solve this with a ZZ approach and upon my fourth pair I came across something I found very interesting.
With a simple U perm I could get the cube solved apart from four cubies twisted in their correct spot.

But could not put it to good use... Is there really something there or does it just look nice and that's it?

z2 B2 D2 R F' D2 L' D' EOLine 7
L' U' L R' U R 1st pair 13
U2 L U2 R U R' 2nd pair 19
U L U' L' U2 L' 3rd pair 25
Nasty remaining corner edge pair...
All ways of inserting it correctly where too long and resulted in long OLL's

So I just "hammered" it in with U' F' L F L' 30
Now a CW U perm (11) permutes the edges and I am left with 2 twisted corners and 2 flipped edges in their correct positions.

(The edge cycle of the U perm can be inserted in various positions in the solve to check for cancellations)

I find the resulting state looks really interesting but is there a way to fix this with a commutator?

There /is/ a way to fix it with a commutator, but it's too long for your solve. I'd probably find a way to insert that 3rd pair a different way so the 4th pair won't be so hard. [multislotting?]

 

[Cubenovice]

Initially I had a different 3rd pair but that gave me another unfavorable 4th pair

9 moves for a 3 edge cyle? try U2 M U2 M'

On a serious note: I do not know too many algs so just counted turns of my speedsolve U perm

 

[Forte]

Now a CW U perm (11)

11? 9 please F2 U R' L F2 R L' U F2

9 moves for a 3 edge cyle? try U2 M U2 M'

U2 M U2 M' with a setup is F2 U R' L F2 R L' U F2 >_>

 

[DavidWoner]

Also starting with ZZ and then messing up EO on your last pair is a pointless exercise. Having edges oriented is really only important for the end of your solve, so doing it at the beginning and then destroying it right when you need it is an awful idea. And inserting pure flips/twists generally turns out poorly.

 

[Cubenovice]

Fair enough
I was thinking about an insertion where the pieces where on the M slice.