[Help Thread] FMC Discussion

[Kryptonite]

Insertions aren't saving anything, you add moves with it (-;

Haha, true. I suppose it would be more accurate to say that not using insertions will typically have an opportunity cost

it saves about 1 move for each 7 moves in your frame

Oh cool, how did you get this number btw?

 

[guusrs]

Oh cool, how did you get this number btw?

Experience on about 300 insertions I did in the past 5 years.
But we could apply math on that I suppose

 

[Cube Equation]

I'm quite sure that this situation has occurred many times before, but I desperately need help...

I've finally decided to attempt FMC. After conducting some research, I've found that the general approach for a solution seems to be blockbuilding, and then finishing with corner commutators. I have little problems with the latter, as I have recently explored the BH method from BLD. However, I am encountering difficulties when blockbuilding in methods such as Petrus. I find that I would always blockbuild using inefficient CFOP techniques. Therefore, my question is: how can I decrease the average move counts of my blockbuilding? I suspect that the answer may simply be practice. But I would prefer a more specific answer. How exactly should I practise? I feel that simply practising without a specific goal is merely reinforcing my bad habits from CFOP.

Thanks.

 

[Kryptonite]

Experience on about 300 insertions I did in the past 5 years.
But we could apply math on that I suppose

That's good enough for me!

I am curious though, how would one go about the math for this? The cube state during a solve isn't a truly random, so it seems like it would be difficult to determine the "true" value behind insertions.

I suppose one could do an empirical study though by analyzing solves from the FMC challenge. Though, processing the data properly would be no small feat.

EDIT: Also, our guesses are surprisingly similar. I said 3.5 moves, and you said 1 per 7 in the frame. So that means 3.5 moves would be saved in a 24.5 skeleton, giving a final move count of 29.8, and I've personally always considered 30 moves as "the number to beat."

 

[kinch2002]

I can't think of a better formula off-hand right now, but I certainly don't average over 3 moves per insertion, and my skeletons for L3C do average at least 21. I think I average just over 2 moves cancelling. That's just my feeling - I have no exact figures to back that up.

 

[Kryptonite]

Well, insertions provide a hidden benefit, because they (practically) guarantee an 8f* commutator, whereas applying the 3-cycle at the end, as Guus pointed out, averages 8.8 moves. So the average value of an insertion is .8 + average # of cancellations.

 

[guusrs]

But statistics can lie too: Tomoaki Okayama got a 22 move solve for Per's FMC #372 (http://fmc.mustcube.net/results.php?round=372&format=classic&metric=htm)

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

Solve: U2 L' U' L2 U L2 F U B' U' B L' F L' F' U' L2 R U2 B' F D'

(For inverse scramble)
Pre-scramble: U2

2x2x1 block: D F'
2x2x2 block: B U2 R'
2x2x3 block: L2 U
Finish F2L: F L F' L F'

Then

All but 3 corners: B' U B U' * L2 U' L2 U L
Correction: U2

Insert at *: U B' U' F U B U' F' (8 moves cancel!)

Congratz Tomoaki, but you could have done that insertion straight into your linear solve.

 

[marco.garsed]

But statistics can lie too: Tomoaki Okayama got a 22 move solve for Per's FMC #372 (http://fmc.mustcube.net/results.php?round=372&format=classic&metric=htm)

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

Solve: U2 L' U' L2 U L2 F U B' U' B L' F L' F' U' L2 R U2 B' F D'

(For inverse scramble)
Pre-scramble: U2

2x2x1 block: D F'
2x2x2 block: B U2 R'
2x2x3 block: L2 U
Finish F2L: F L F' L F'

Then

All but 3 corners: B' U B U' * L2 U' L2 U L
Correction: U2

Insert at *: U B' U' F U B U' F' (8 moves cancel!)

Congratz Tomoaki, but you could have done that insertion straight into your linear solve.

WTF!? congrats for F2L in 12 moves!

 

[irontwig]

No offence, but that seems to be limited LL knowledge rather than a fantastic insertion though

 

[marco.garsed]

But statistics can lie too: Tomoaki Okayama got a 22 move solve for Per's FMC #372 (http://fmc.mustcube.net/results.php?round=372&format=classic&metric=htm)

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

Solve: U2 L' U' L2 U L2 F U B' U' B L' F L' F' U' L2 R U2 B' F D'

(For inverse scramble)
Pre-scramble: U2

2x2x1 block: D F'
2x2x2 block: B U2 R'
2x2x3 block: L2 U
Finish F2L: F L F' L F'

Then

All but 3 corners: B' U B U' * L2 U' L2 U L
Correction: U2

Insert at *: U B' U' F U B U' F' (8 moves cancel!)

Congratz Tomoaki, but you could have done that insertion straight into your linear solve.

WTF!? congrats for F2L in 12 moves!

 

[cuBerBruce]

WTF!? congrats for F2L in 12 moves!

Huh? I count 13 moves.

 

[marco.garsed]

you're right! 13 with premove..

 

[okayama]

Congratz Tomoaki, but you could have done that insertion straight into your linear solve.

No offence, but that seems to be limited LL knowledge rather than a fantastic insertion though

Yeah, you're right. If you know all of the 10-move LL, you can solve that LL in one step:
F B' U B U' F' L2 U' L2 U
but I currently know only COLL + quite-easy LL... :fp
By the next competition I will learn more LL.

EDIT: In my experience, two moves (or so) cancel on an average by a corner 3-cycle insertion.

 

[guusrs]

you're right! 13 with premove..

13+1 makes 14 moves for me

 

[marco.garsed]

ok sorry I'm not able to count in these days XD...

 

[Cubenovice]

Spols Open fail-fest

Scramble 1: B' D2 F R2 B F L B U' F U2 F R F L' U R' B R
Scramble 2: L' D2 R2 F U B' L D L F2 R B2 D' B2 D2 B D2 R2

1st
Premoves F D L'
L' F' U2 L2 - 2x2x2
R . F' D2 - 2x2x3
R2 F R2 F2 - F2L-1
could not find good continuation


After comp, same start
Premoves F D L'
L' F' U2 L2 - 2x2x2
R . F' D2 - 2x2x3
R' F R' F R2 F2 - F2L-1 (just inserting F into R2 of original solution)
R2 F D' F' D R' D R' D' R - leaves 3 corners
F D L' - undo premoves
At . insert R B R' F' R B' R' F to cancel 3 moves

L' F' U2 L2 R2 B R' F' R B' R' D2 R' F R' F R2 F2 R2 F D' F' D R' D R' D' R F D L' = 31 HTM

2nd
I cannot believe I missed the obvious 2x2x2 block...
I did see that L' made a pair, how could I not see that F' made the square????
At a given moment I decided to just put in a ZZ solve just to get "a result"
6 move EO-line into bad continuation and a notation error to top it off.


On a more positive note:
I've been helping manhobbyfreak a little with FMC techniques and he came very close to a 31 on the 1st scramble.
19 HTM for L5C
1st insertion with 2 moves cancel
2nd insertion with 2 moves cancel but unfortunately the com was faulty. (time pressure)
We've got a future winner here!

Tristan's solution
L' B2 F' L - 2x2x2
R . U2 R2 U' R - 2x2x3
U' R U R' F2 U' F U' F' U2 - leaves 5 corners
at . insert R' B2 R F2 R' B2 R F2 to cancel 2 moves
don't have the wrong comm anymore...
We checked what happend, it included an R2 as interchange move but there was only one corner in the R face

I also ran a check on the L5C and found a 4 move comm after the very first move, leaving 3C in 23 HTM.
We'll both do a full check and compare.

 

[kinch2002]

20 move FMC

[22:50] <+Kirjava> scramble: R F U2 R2 U L2 D2 L2 F' L' F' B' L2 B2 D F' U2 R F' D2 F2 L' B' D F'
F2L-1: L2 B' F' R F D2
Continue: D L F L' F2 D F D2 L (2 twisted corners left)
So the skeleton is: L2 B' F' R F * . D' L F L' F2 D F D2 L (14)
Insert F' R' F L F' R F L' at * to cancel 6 moves (standard conjugate cancellation)
Insert L' U' L D' L' U L D at . to cancel 4 moves

Solution: L2 B' L F' R F L2 U' L D' L' U L2 F L' F2 D F D2 L (20 moves)

 

[Cubenovice]

I like how you got this 2-twist shortly after discussing conjugate cancellations on the chat

 

[Zane_C]

20 move FMC

[22:50] <+Kirjava> scramble: R F U2 R2 U L2 D2 L2 F' L' F' B' L2 B2 D F' U2 R F' D2 F2 L' B' D F'
F2L-1: L2 B' F' R F D2
Continue: D L F L' F2 D F D2 L (2 twisted corners left)
So the skeleton is: L2 B' F' R F * . D' L F L' F2 D F D2 L (14)
Insert F' R' F L F' R F L' at * to cancel 6 moves (standard conjugate cancellation)
Insert L' U' L D' L' U L D at . to cancel 4 moves

Solution: L2 B' L F' R F L2 U' L D' L' U L2 F L' F2 D F D2 L (20 moves)

That solution is incredible, the F2L-1 is simply ''. You've shown me that 2 twisted corners aren't as bad as I thought.

 

[Attila]

20 move FMC

[22:50] <+Kirjava> scramble: R F U2 R2 U L2 D2 L2 F' L' F' B' L2 B2 D F' U2 R F' D2 F2 L' B' D F'
F2L-1: L2 B' F' R F D2
Continue: D L F L' F2 D F D2 L (2 twisted corners left)
So the skeleton is: L2 B' F' R F * . D' L F L' F2 D F D2 L (14)
Insert F' R' F L F' R F L' at * to cancel 6 moves (standard conjugate cancellation)
Insert L' U' L D' L' U L D at . to cancel 4 moves

Solution: L2 B' L F' R F L2 U' L D' L' U L2 F L' F2 D F D2 L (20 moves)

My solution also 20 moves
F2U2BF’L’U’FU’R’F2L2B2L’FB’U2FB’RF2
F2U2BF’L’U’FU’R all corners and 5 edges,
R2F2L2B2L’FB’U2FB’RF2 2 edges swap , L5E.