[Help Thread] FMC Discussion

[THE_V01DL0RD]

Okay! I took a quick look at your solution and here are some tips:
(Disclaimer: My solution shown below might not be (close to) optimal either since I just took a quick look!)

1. You should get better at doing DRs. 18 moves is a lot for DR. For your EO I quickly found 2 different 14 move DRs:
DR1(on white-yellow axis):

Your EO
(D' L) //4c4e
(B2) //creates a second 1x2 "bar" on DLB corner
(D R2) //creates a 1x3 bar on FD position
(B2) //creates a second 1x3 bar on UR position AND sets up the to correct e-slice edges to be on the same side
(U2 D' L) // solves the rest.

DR2(on red-orange axis):

Here you already have a 4c4e case after doing your EO
(L' U2) //creates a 1x3 bar on UF position
(R2 B2) //creates a 1x2 bar on URB position
(L' U2) //creates a second 1x3 bar on UB position
(L' R2 D) //solves the rest

And additionally there was even a 16 move DR using your 9 move 4c2e solution:
DR3:
Your 4c2e
(D' L2 D U) //setup to trigger
(L D2 L') //trigger

Take a close look at those examples above and make sure you learn how to more efficiently setup the triggers. Also only try to generally use 2 moves max to get from EO to 4c2e or 4c4e cases. If you cant find any 2 move or less setups you propably should try another EO.

2. Speaking of EO. Make sure you find optimal EOs AND make sure you check multiple different EOs:

I dont know if you saw it and checked it but there are some 4 move EOs on inverse on different axis.
(R D2 B2 D) and (R U2 F2 D) are both 4 move EO solutions.
Also it is VERY important to check many EOs. Most of the EOs dont have a good continuation so you should check as many of them as you can.

3. HTRs:
You are correct that your DR solution is actually 3 qt away from solved. The reason you got fake HTRs is because you used 2 or 4 or any even number of quarter turns to get to your (fake) HTR. To get real HTR you need use odd number of QTs since you are an odd number away from solved. You should google "hyper-parity". There is a document from which you can learn to quickly recognize how far (in QTs) you are from HTR and how you can move closer to HTR. That is vital in finding HTRs. How ever that is quite advanced so I would recommend you getting more experienced and comfortable with finding efficient DRs first and foremost.

Here is however a HTR I found for your start quickly:
First we notice from the hyper-parity graph that we are 3qt away from solved (bars+bars case, we could "solve it in 2 moves but that would give us a fake HTR since we have parity). We also immediately notice that just doing a U will move us to solved+one face case which is only 2qt from solved so we do that.

(U') //puts us 2qt away from solved.

Now we look at the hyper parity graph again and see that next we need to get to slashes+oneface case if we want to finish in 2 quarter turns.

(L2 U') // puts us 1qt away from solved

Note that you also want to have exactly 2 M or S-slice edges on the correct slice since we want to solve the edges in 1 quarter turn next along with corners. R2 instead of L2 would work for corners here but not for edges. And that will always be the case here: turning the opposite layer will have similar impact to corners but different impact to edges, which is very useful to understand.

Now we just want to create two 1x3 bars AND have the correct edges be in the "same side" of the cube once we are done. Just like when solving 4c4e DR case. Except we can only use half-turns now. So lets imagine we are trying to just do DR on blue-green axis. We already have two 1x2 bars and just doing R2 would make two 1x3 bars. But the problem with this is that it would leave the 2 edges (white-blue and green-yellow) on different sides of the cube. So we must first do:

(B2 L2 B2) //moves the edges to different sides of the cube

Now that the (whiteblue and greenyellow) edges are on different sides of the cube we can do

(R2) //creates two 1x3 bars while moving the 2 edges to the same side of the cube.

Rest is easy. Just finish like it was DR.

(D2 F2 D) HTR


Now dont worry if the HTR stuff seemed too complicated. It is hard at first but getting better at DRs helps. And learning the hyper-parity graph is essential.

I hope this helps you to get better and understand some concepts

Edit: link to hyper-parity: https://docs.google.com/document/d/1Onc2lu5iW_ODnT71oN5lgpVVjVqbvUMzaylzxgykx0I/

Yea just a quick look

The closest to an as-executed version would be:
R2 Rw2 U R2 Rw2 U2 R2 Rw2 U R2 Rw2

Basically, every time you want to do M2, write R2 Rw2 instead. Written rotationlessly, it'd be what Garf wrote above.

This is a bad H perm to use in FMC though; consider also:
L R U2 L' R' F' B' U2 F B (10 moves; always cancels at least 1, or at least 2 if there's a surrounding quarter turn)
R2 S2 R2 U R2 S2 R2 (9 moves + AUF; always cancels at least 1, or at least 2 if there's a surrounding half turn)
R2 S2 R2 U L2 S2 L2 (same as above; unlikely to cancel more moves but good to know)

(I hope it's obvious enough, but in case it's not: you learn the alg with S2, and write it down with F2 B2.)


You should maybe start using proper scrambles. Seems a bit much of a coincidence that the two scramble sequences you've posted are both 15 moves.

 Twizzle link 

 Setup 

L U F' R B2 F' L2 R' D' B2 U B D L U'

 Moves 

R F2 R' U' F U' B R' B' R B2 R B R' D' B' D B D' B' D B' U B U' Rw' U L U F U' F' y U2 R U' R' U' F' U F y2 L' U L U F U' F' y R2 U2 F R2 F' U2 R' U2 R' U' M' U R U' Rw' R2 U2 R U2 R2 U2 R2 U2 R U2 R2 U'

(edit: I missed this earlier, but your solution still has an M' in it)

Your cross took 6 moves, which, fine, whatever. You then spent 19 (!!!!) more moves to finish solving the green layer.

Here's a 6-move xcross on green. You did mention you use xcrosses, right?? (If I were colour neutral, I would have gone for this xcross in a speedsolve. This is not an FMC-specific thing!)

 Twizzle link 

 Setup 

L U F' R B2 F' L2 R' D' B2 U B D L U'

 Moves 

U' D2 F U' B L2

The biggest weakness of the beginner LBL method you're using is that it involves a lot of breaking up and restoring. You're constantly destroying your earlier progress in order to solve additional pieces. Once the cross is solved, it minimally takes 3 moves to insert any first-layer corner. Once the first layer is solved, it minimally takes 5 moves to insert any second-layer edge, often at least 7 moves. That's a lot of wasted moves that can be avoided with a not-beginner method.

(Exercise for the reader: in what situations can you preserve the first layer and solve a second-layer edge in 5 moves? Answer: R2 U2 R U2 R2)

The big four (speedsolving) methods don't suffer this problem nearly as strongly.

CFOP: You still have to break up and restore the cross constantly, but it's still much less bad than beginner LBL because you're solving two pieces at once. Instead of 3 (solve corner) + 7 (solve edge) = 10 moves to solve one slot, you solve it in 4-7 moves most of the time.

Petrus: This problem only shows up during the EO step and the last layer. EO turns out to not be a major detriment because it's often possible to solve it in 2-5 additional moves in many different ways.

ZZ: Like Petrus, except that EO is done at the very beginning so there are no restrictions there, so the problem only shows up during the last layer.

Roux: This problem only shows up during CMLL, and even then it's not that big of a deal because you can often handle CMLL with insertions (an FMC technique) or with short algs. (Arguably, the need for slice moves during LSE might also count, but the efficiency win in F2B makes up for that when compared against the other big speedsolving methods.)

At your level, you don't need to practise FMC to get better at FMC. You need to practise literally any of the big four speedsolving methods.

I used 15 moves scrambles because I think its enough and I use the WCA scrambler. Thank you for the advice. Hopefully my 80ish average can go down.
*also it's practice not practise

 

[Garf]

Scramble: L U F' R B2 F' L2 R' D' B2 U B D L U'
Solution: R F2 R' U' F U' B R' B' R B2 R B R' D' B' D B D' B' D B' U B U' Rw' U L U F U' F' y U2 R U' R' U' F' U F y2 L' U L U F U' F' y R2 U2 F R2 F' U2 R' U2 R' U' M' U R U' Rw' R2 U2 R U2 R2 U2 R2 U2 R U2 R2 U'

I didn't use slice moves and used the w thing for wide moves
I just want a tutorial for FMC so I don't have 70 move solutions.
I use beginner F2L and 2LLL

Rotations aren't allowed for FMC.

 

[xyzzy]

Rotations aren't allowed for FMC.

??
Rotations are allowed!

It's just that you very rarely see experienced FMCers using rotations because we find it a waste of time and it introduces additional state (the orientation of the cube) that we have to keep track of.

(I would agree that it's a pretty darned good idea to stop (ab)using rotations when writing FMC solutions, but that's not exactly a priority to work on at this point.)

I used 15 moves scrambles because I think its enough and I use the WCA scrambler.

It's almost the same amount of effort to just use proper scrambles, so stop doing whatever you're doing and just use proper scrambles.

Let me throw this back at you: why is 15 enough? Why not 14? Why not 13? Etc. etc. you get the idea. Why draw the line at 15 moves, when you can just not draw a line?

*also it's practice not practise

American English is not the only English. It's not even the English English, for that matter.

PS: Please stop quoting entire posts when replying. Quote only the relevant parts, and delete the rest.

 

[Mrauo]

Reconstructions of my 22.67 Mo3 from Martorell Open 2023

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

(D R B' U) // EO (4/4)
(R' L' B') // DR-2e3c (3/7)
(L B L B) // DR (4/11)
R U2 R' // HTR (3/14)
(L2 B2 R2 F2 D2 F2 U2 B2) // finish (8/22)

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

22 moves

Spoiler: R' U' F U2 L B D B2 F2 L2 U2 B2 L U2 L' F2 U' L' U R F U B' R' U' F

B' D R' F U // EO+DR-4e4c (5/5)
R' U2 D2 L F' // DR (5/10)
L D2 F2 L D2 R F2 L U2 D2 L U2 R' // finish (13/23)

Solution: B' D R' F U R' U2 D2 L F' L D2 F2 L D2 R F2 L U2 D2 L U2 R'

23 moves

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

B' R U' L'// EO (4/4)
F' // DR-4e4c (1/5)
U' D B2 U' R2 D B // DR (7/12)
L2 F2 R2 * B2 U'^ // AB3c+slice (5/17)

* R2 D L2 D'^ R2 D L2 D' // 3c (8-2/23)
^ w //slice (0/23)

Solution: B' R U' L' F' U' D B2 U' R2 D B L2 F2 D L2 U' B2 D F2 D' L2 D'

23 moves

 

[THE_V01DL0RD]

What are some FMC methods I can use.
I wanna learn so I can get at least a 50 mean.

 

[SkewbIsTheBestEvent]

What are some FMC methods I can use.
I wanna learn so I can get at least a 50 mean.

CFOP is the best FMC method for Sub-50. You can also use blockbuilding for F2L

 

[THE_V01DL0RD]

CFOP is the best FMC method for Sub-50. You can also use blockbuilding for F2L

So 2LLL?

 

[SkewbIsTheBestEvent]

So 2LLL?

That would be needed, yes

 

[THE_V01DL0RD]

Scramble: B' U' L2 F U' L F' D2 R L' F U2 B2 R2 D2 F2 D2 F
Solution:
F L D2 (Cross)
x' D2 L U2 L' U2 B' U B (F2L #1)
U' R U2 R' U' R U2 R' (F2L #2)
y' U' F' U' F (F2L #3)
y2 U F' U' F (F2L #4)
y' R U R' U' R U' R' F' U' F R U R2 (OLL)
U L' U2 R U' L R' U L' U2 R U' L (PLL)
U2 (AUF)

 

[THE_V01DL0RD]

Scramble: B2 D F2 L2 D' R2 D' R' B2 D2 L2 U2 R2 B2 L B U2
Solution:
U2 F2 U F2 (Cross)
x' F U2 F2 U F (F2L 1)
y' B' U' B L' U2 L R' F R F' (F2L 2)
y2 U R U2 R' U' F R' F' R (F2L 3)
y F' U2 F U F' U' F (F2L 4)
F' Rw U R' U' Rw' F R (OLL)
y2 B2 L U L' B2 R D' R D R2 (PLL)

 

[THE_V01DL0RD]

This is unrelated, but where can I watch One Piece?

 

[Silky]

CFOP is the best FMC method for Sub-50. You can also use blockbuilding for F2L

This is actually terrible advice.

What are some FMC methods I can use.
I wanna learn so I can get at least a 50 mean.

Petrus is going to be the best method to learn method wise, at least at a beginner level.

 

[THE_V01DL0RD]

I think CFOP is pretty good. F2L is frusturating.

 Twizzle link 

 Setup 

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

 Moves 

R F2 R' L B' R' U R F
x' F U' F' U R U R' U R U' R'
y2 R U' R' U R U2 R' U R U' R'
y R U R' U' R U' R' U Dw R' U' R
y' F' U' F U F' U' F
R2 D' R U2 R' D R U2 R
y2 R' U2 R U2 L U' R' U L' U L U' R U L'
U

 

[tseitsei]

I think CFOP is pretty good. F2L is frusturating.

CFOP is insanely bad for FMC solves. That said it can definitely get you to 50 move solves.
But if you really want to learn FMC you should definitely learn Petrus (or any blockbuilding method) first. Blockbuilding is propably the most fundamental skill needed in FMC and you dont learn that by using CFOP. Petrus also has another very useful concept to understand: edge orientation.

And when you want to learn more you can start incorporating more stuff in addition to petrus (not so much the case for CFOP). You can later add NISS and starting with EO and Heise and insertions etc. quite "naturally" if you are using Petrus/blockbuilding.

Also Petrus just by itself is much better in FMC than CFOP.

I would say just learn Petrus first, then something like Heise, then insertions, then NISS

 

[THE_V01DL0RD]

But if you really want to learn FMC you should definitely learn Petrus (or any blockbuilding method) first. Blockbuilding is propably the most fundamental skill needed in FMC and you dont learn that by using CFOP. Petrus also has another very useful concept to understand: edge orientation.

Question 1. Is block building the thing where you do a 2x2 then a 2x2x3 and Solve the rest of the cube?
Question 2. Is 3 style edge orientation?
Question 3. What was your NR for FMC, 3BLD, 4BLD and MBLD?

 

[tseitsei]

Question 1. Is block building the thing where you do a 2x2 then a 2x2x3 and Solve the rest of the cube?

Yes. It's much more efficient than solving CFOP pairs since you always have to take a cross piece out and put it back in when inserting pairs.

Question 2. Is 3 style edge orientation?

No, 3 style is just another name for using commutators to solve corners during BLD solve (closely related to insertions in FMC).
Edge orientation is just orienting all of the edges in a way that they can all be solved without using F and B turns (or R and L or U and D turns). In traditional Petrus you first build a 2x2x3 and then orient the remaining edges. But you can also orient edges right at the beginning and do blockbuilding with oriented edges.
Edge orientation is also the first step of THE best FMC method currently where you do EO->DR->HTR, but that is way more advanced.

Question 3. What was your NR for FMC, 3BLD, 4BLD and MBLD?

Here is a link to my WCA profile where you can see my official PBs: https://www.worldcubeassociation.org/persons/2012LEHT01
However nowadays I am way worse in all the events besides FMC.

 

[THE_V01DL0RD]

I got a mean of 59.67.

 

[THE_V01DL0RD]

has anyone everdone 2x2x2 fmc

 

[SkewbIsTheBestEvent]

has anyone everdone 2x2x2 fmc

its kinda pointless

 

[THE_V01DL0RD]

its kinda pointless

What about 4x4x4 FMC