• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Kit Clement

Premium Member
Joined
Aug 25, 2008
Messages
1,606
Location
Portland, OR
WCA
2008CLEM01
YouTube
Visit Channel
A typical but lucky FMC attempt

L2 D2 B D2 // 2x2x2 Block
U' B2 D' B U B' // 2x2x3 Block
R D R' D2 // F2L '1
F' R F R2 F' R2 F //F2L
R2 B' R' B' D B D' R B //OLL
R' F2 R U D' F2 U' D R F2 R2 //3e

41 ETM

Ending F2L differently,

L2 D2 B D2 // 2x2x2 Block
U' B2 D' B U B' // 2x2x3 Block
R D R' D2 // F2L '1
D R D' R' F D' F' D //F2L
D R2 D2 F D F' D R2 D' //OLL
D2 r' D R' D R D' r D2 B R' B' // 3c3e

43 ETM

critique this please

First off, including the scrambles always helps. I made one with cube explorer: R2 D2 B R B R2 L2 U' F U R2 B' R B (14 optimal? Are you generating purposefully easy scrambles?)

First: why did you AUF on OLL? You could have cancelled 3 moves between F2L/OLL and another 3 between OLL/PLL for a 35:

L2 D2 B D2 // 2x2x2 Block
U' B2 D' B U B' // 2x2x3 Block
R D R' D2 // F2L '1
F' R F R2 F' R2 F //F2L
F' R' F' U F U' R F //OLL
F' R F' R' F' R' F' R F2 R' //PLL

Going earlier, note that you have a 2x2x2 block far sooner than you say:

L2 //2x2x2 block

After this, you seem to want to extend the 2x2x3 normally, but it comes at making one of your pairs (blue-orange-yellow) difficult to preserve later. I think it's easier to work on this 2x2x1 extension first:

B2 D2 //+square

And then pace the blue-orange-white pair. You can now use this 1x2x3 to create a PS-F2L

U' B' U B' //PS F2L

You're now incredibly close to getting a skeleton, we can now focus on solving the green-red edge in a way that solves all edges:

D2 F D F' D' B //5c

This leaves 13 to 5 corners, which is a pretty good solve. Insertion finder gets a 25 for this when only using pure comms (24 if you use longer comms): https://fewestmov.es/if/8d5dd489479f724c7211bec0865266b7

Another interesting finish is by seeing some pseudo-slotting stuff at the PS F2L - notice you have a pseudo-pair built one move before the PS F2L-1 is built, so if you pick up after that step, you can first insert the edge, then insert the pseudo-pair and leave 3c in 16 moves:

L2 //2x2x2 block
B2 D2 //+square
U' B' U //F2L-1.5
D B' D B U F D2 F' U' D2 //3c

Insertion finder gives this an optimal finish of 23: https://fewestmov.es/if/779b0a88e58ddd1deaac4066f98eb262

I would try to give you advice using the F2L-1 that you gave, but the EO is just not good in that state to get good finishes. If we back up a bit to this point though:

L2 D2 B D2 // 2x2x2 Block

Notice that you're 1 move away from a 2x2x3. In your solution, you do:

U' B2 D' B U B' // 2x2x3 Block

Which is 6 moves to make a 2x2x3 instead, just to preserve a pair. Additionally, just doing the 1 move to finish the 2x2x3 and breaking the pair gives you much better EO (only 2 bad edges from Petrus orientation) so using 5 moves for a pair and bad edges is definitely not worthwhile. If we try that start instead:

L2 // 2x2x2 Block
D2 B D2 B2 // 2x2x3 Block

Notice that there are a lot of ways to finish the Petrus EO and get some pairs along the way too. One example:

D' F D2 F' // EO+pair

This is only 4 extra moves and is a lot easier to work with this EO. The parity is bad on this case though, so it doesn't lead to a great finish (that I can see, at least), but all other things considered equal, this is much better to work with than your 2x2x3 marked substep.

Also, if you don't know how to use NISS yet, you should. I didn't use it in any of these examples (mostly because this scramble was super easy already to find continuations on) but it will make your options for continuations much more plentiful and make it easier to work with pseudo-blocks.
 

trangium

Member
Joined
Jul 24, 2019
Messages
87
WCA
2019TRAN10
If I learn one method just for FMC, should it be Petrus, ZZ, Roux, or something else?
Most of the top solvers use Domino Reduction in most of their solves, so learn Domino Reduction. A guide for Domino Reduction can be found at
 

xyzzy

Member
Joined
Dec 24, 2015
Messages
2,355
Hey guys, I tried DR for the first time. I achieved DR but I don't know where to go
U R D2 U' F2 R2 D //p222 U... (cubing.net)

can anyone end this and get a sub 50??
I kinda doubt you can start from a 44-move DR and finish in only 6 more moves…

Here's my own DR attempt for that scramble:
U (B2 U') // EO on U-D axis (3/3)
F // orient most of red face & solve two S-slice edges (1/4)
(R L' B' L' B) // left with 3 twisted corners and one S-slice edge (5/9)
(L F' L' F') // easy DR case (4/13)
(R U2 R' L' F2 R L2 U2) // orange layer (8/21)
(U2 F' U' F U R F R' F' U2 R' U R U' L') // F perm (15−2/34)

(I'm still pretty bad at DR (both in getting to DR, and in continuing after DR) so take this with a grain of salt.)

If you're going to do DR, it's not that useful to have a purely blockbuilding start because that locks out most of the moves you'd want to use to influence EO. For example, with a 2×2×2 block in DBL, that leaves the U, R, F faces free, and you get to use the F move for EO. At first glance it might seem like you get to flip four edge pieces at once with F moves, so there's no difference with doing EO right at the start, but the difference is that now you can no longer directly move bad edges to the DF or LF positions to flip them because the block is in the way. Also, even if you are going to blockbuild, you don't need to build "real" blocks; you can just build blocks out of oriented pieces and ignore the actual colours on them.

The power of DR comes from how you can affect corner orientation in very few moves when you don't have to preserve the belt. In normal solving, Sunes seem like pretty fast algs (7 moves makes it the shortest ZBLL), but with DR, you have even shorter "algs" for changing corner orientation: R U R', R U' R', and R U2 R' (3 moves).
 

xyzzy

Member
Joined
Dec 24, 2015
Messages
2,355
Another question:

Should I solve a belt right after EO and preserve it while doing CO or should I try to solve the belt after solving CO?
Solve 3/4 of the belt after EO, while forcing 3 or 4 twisted corners, then solve the last belt piece together with CO. If you know NISS, it's usually not too hard to set up to one of the 3-move cases.
 

Kit Clement

Premium Member
Joined
Aug 25, 2008
Messages
1,606
Location
Portland, OR
WCA
2008CLEM01
YouTube
Visit Channel
Another question:

Should I solve a belt right after EO and preserve it while doing CO or should I try to solve the belt after solving CO?

Solving the belt directly is a waste of time given that you need to break it to finish CO. Using the simple triggers that xyzzy mentioned and setting up to those is the simplest way to go from EO to DR.
 
Joined
Aug 30, 2020
Messages
1,422
Location
On a long train journey, Smashin' PBs one a stop
YouTube
Visit Channel
R B2 U2 F //EO(Somewhat) D2... (cubing.net)
I got stuck again.....please help me
Btw I don't know NISS and I get confused whenever I sit down to learn it.

E: I got DR:

R B2 U2 F //EO(Somewhat)
D2 R' U2 F2 U' F' L' B L B' //belt + force CO on top+ break EO(accidentally)
U' R' L F R L' //fix EO
(F R' U2 R F') D2 (F R' U2 R F') D2 //DR
U2 R L' B2 R' L F' B L2 F B' //223
U F2 U2 F2 //47 to 6c2e(I highly doubt this) lol

I'm stuck here
 
Last edited:
Joined
Aug 30, 2020
Messages
1,422
Location
On a long train journey, Smashin' PBs one a stop
YouTube
Visit Channel
First sub 40!!!!

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

Solution:
R F U L2 R B U' R U2 R2 //223
U' r' U r2 B' r' //F2L-1
f R2 f' //LS
U' F U R U2 R' U' R U R' F' //OLLCP
L2 U F' B L2 F B' U L2//EPLL


can anyone critique this and give any other good continuations after the F2L-1?
(if you use NISS please indicate where to switch)

I didn't use WCA notation sorry

Also, I can't seem to wrap my head around insertions and NISS so can anybody help me understand it?
(this is after seeing J Perm's tutorial for the 50th time lol)

@xyzzy @Kit Clement
 
Last edited:

asacuber

Member
Joined
Nov 4, 2015
Messages
1,442
Location
India
WCA
2015SANT44
YouTube
Visit Channel
First sub 40!!!!

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

Solution:
R F U L2 R B U' R U2 R2 //223
U' r' U r2 B' r' //F2L-1
f R2 f' //LS
U' F U R U2 R' U' R U R' F' //OLLCP
L2 U F' B L2 F B' U L2//EPLL


can anyone critique this and give any other good continuations after the F2L-1?
(if you use NISS please indicate where to switch)

I didn't use WCA notation sorry

Also, I can't seem to wrap my head around insertions and NISS so can anybody help me understand it?
(this is after seeing J Perm's tutorial for the 50th time lol)

@xyzzy @Kit Clement
Great job!
I was able to find this immediately, from your F2L
R F U L2 R B U' R U2 R2 //223
U' r' U r2 B' r' //F2L-1
f R2 f' //LS
f U' f' U' f U f' L' U L//Alternative OLL (cancels 2)
R' F R' B2 R F' R' B2 R2//PLL

which gives 36 instead of 39!
 
Joined
Aug 30, 2020
Messages
1,422
Location
On a long train journey, Smashin' PBs one a stop
YouTube
Visit Channel
Great job!
I was able to find this immediately, from your F2L
R F U L2 R B U' R U2 R2 //223
U' r' U r2 B' r' //F2L-1
f R2 f' //LS
f U' f' U' f U f' L' U L//Alternative OLL (cancels 2)
R' F R' B2 R F' R' B2 R2//PLL

which gives 36 instead of 39!
Thanks!! Looks like I should learn more algs and techniques(especially NISS, rNISS and insertions).
 

xyzzy

Member
Joined
Dec 24, 2015
Messages
2,355
f U' f' U' f U f' L' U L//Alternative OLL (cancels 2)
R' F R' B2 R F' R' B2 R2//PLL
Even better last layer:
R B U B' U' R' // EO
R B U' B' U' B U2 B' U' R' U' // ZBLL
15 moves for LL after cancelling the R' R.

Equivalently, on inverse:
(B U L U' L' B') // EO
(B L U' L' U L U L' U L U' L' B') // COLL
(U) // EPLL skip

(note: I'm not going to explain NISS here. I'll probably look at J Perm's video again because iirc it actually seemed like a pretty good explanation so I'm not sure which part of it you don't get.)

---

The above takes 18 moves for LSLL. Other options for LSLL I'd look into if this were my solve attempt and I had some time to spare:
option (i)
U' B U' B' // LS
This gives a pretty bad LL edge case (optimal to solve only the edges is 11 moves). None of the 6-mover or 7-mover EO algs give a nice ZBLL case afterwards either. On inverse, we have OLL into T perm:
U' B U' B' // LS (4/4)
(R B L' B L B2 R') // OLL (7/11)
(B2 D' R2 D B2 L2 D F2 D' L2) // T perm (10/21)

OLLCP on inverse gives a Z perm (with U2 AUF), so that's 4+8+12−2=22 moves for LSLL:
U' B U' B' // LS (4/4)
(R' U2 R U2 R B' R' B) // OLLCP (8/12)
(F2 B2 D' F2 B2 U' R2 F2 B2 L2 F2 B2) // Z perm (12−2/22)

You can even go for parity insertions, although this obviously doesn't count as an LSLL solution anymore, and it's not likely to result in anything good anyway.
… // F2L-1 (16/16)
U' B U' B' // LS (4/20)
U' // ab3c2e parity (5/21)
(My rule of thumb (targeting 30 moves) is that I don't bother with parity insertions for skeletons over 16-ish moves unless I'm desperate.)

option (ii)
On inverse, almost everything looks bad. It's an 8-mover F2L case, and there's no clean edge-only solution; you could do U B U B' U to solve a bunch of pieces, but that's just the same as the ab3c2e skeleton above. You can, however, join a pair to get a pseudoF2L then fix it immediately.

(U' B) B // F2L

Unfortunately, while this case resembles a 1LLL with an 11-move solution, it's not. Probably the best continuation there is to go straight into insertions (20 moves to 4c3e), but for an LSLL-y approach:
(U' B) B // F2L (3/3)
F' U' F @ U F R' F' R U // OLL (9/12)
@ = R' D' R U2 R' D R U2 // corner commutator (8−1/19)
Maybe there's a better corner insertion; didn't bother looking. Instead of doing a corner comm insertion, you can also just do an A perm at the end (which wastes 2 moves compared to this).

option (iii)
U2 B' R B2 U B' U' R' // VHLS (8/8)
B' U' B U' B' U2 B U2 R' U2 R U R' U R // some T 2GLL I don't know the optimal alg for (15/23)
This is pretty bad.

option (iv)
B U' B' U' B' R B R' // another VHLS alg (8/8)
R U2 R' U' R U' R' B U B' U B U2 B' U2 // another T 2GLL I don't know the optimal alg for (15−2/21)
See above remark.

That's about all I'd try for LSLL on this scramble before giving up and trying to find a different F2L-1/skeleton.

On an unrelated note, you should avoid using wide moves in your skeleton. Just write them as normal moves instead. (And yes, use WCA notation for wide moves if you're doing "serious" practice.)

PS: Mentions that are edited in after posting won't ping those people.
 

trangium

Member
Joined
Jul 24, 2019
Messages
87
WCA
2019TRAN10
I tried 5x5 FMC. I used the same method that I did for 4x4 FMC (HTR centers, finish centers while pairing edges, reduce to 3x3x3). Result: 102 moves!

Scramble: Bw Fw' B2 R Rw2 L Uw' Dw' Rw Uw2 Dw2 U Fw U2 Fw2 Bw2 U2 L' Lw U' Lw2 Rw F2 U L2 R Fw Lw2 L' Fw' Bw2 B' D B' Uw Bw2 Dw2 B2 Rw2 Lw' Fw Dw2 U2 D' Lw' Uw' D' U' R Lw2 F2 L' R2 Rw Uw B2 Lw' B Uw' Fw2

Rw B Dw' B Lw U D Lw' Fw U2 Bw' L Bw Lw // HTR L/R centers (14/14)
D' Fw' R D L Fw' D' L U' Bw' // HTR centers, 4-4-3, 2 wings (10/24)
R2 U' Fw2 // 3-3-3, 3 wings (3/27)
F' L' Dw2 // 3-2-2, 4 wings (3/30)
L2 F' D2 R Uw2 // 3-1-1, 6 wings (5/35)
D' B Rw2 // 2-1-2, 8 wings (3/38)
R' Uw U' B U B' Uw' // 10 wings (7/45)
U2 B2 Lw2 // 1-1-1, 12 wings (3/48)
U2 F U R' Bw2 // 0-1-1, 14 wings (5/53)
R2 D Rw2 U L' R2 U' Rw2 // 18 wings (8/61)
Dw L U' L' Dw' // 20 wings (5/66)
L' U2 D B L R2 U D R Dw2 // Reduction (10/76)
L' B2 L2 U L2 U B2 R2 D2 B2 R2 B2 U D' F2 D R D B2 U L2 U2 F' R2 B D // 3x3x3 stage (26/102)

3x3x3 stage explanation:
(D' B' R2 F) // EO (4)
L' (U2 L2 U' B2 D' R') // DR (11)
B2 L2 U L2 U (U') // HTR (17)
R2 D2 R2 // 2e2e + E-slice (20)

Skeleton: L' B2 L2 U L2 U R2 * D2 R2 U R D B2 U L2 U2 F' R2 B D
R2 B2 R2 D2 B2 R2 B2 D2 // Solve 2e2e (24)
Skeleton: L' B2 L2 U L2 U B2 R2 D2 B2 R2 B2 R2 * U R D B2 U L2 U2 F' R2 B D
* R2 U D' F2 U' D // Solve E-slice (26)

View at alg.cubing.net

The current UWR is 126 which is 24 moves more than my result. However, I'm not sure if this counts since this attempt literally took me 3 days.
 

Pyjam

Premium Member
Joined
Oct 8, 2010
Messages
2,142
Location
La Baule, France
Congratulations. Very impressive. sub-100 might be possible.

Last week-end, I established a new UWR for 6x6 FMC with 226 HTM.

Scramble:
R F Uw2 Fw Rw' B' Lw Dw' 3Uw Bw' D 3Fw' F 3Uw2 Dw' Rw' Dw B' 3Uw2 Rw F' B' Fw D Dw2 Rw' Dw2 R' B 3Rw2 Bw Uw' U 3Rw Lw' Fw2 Rw F' B' D' Rw' Bw F Lw' F' Dw' 3Fw' L' 3Fw' Dw' L2 3Uw' R2 B' 3Rw' Rw' Fw2 F' 3Fw2 Uw Fw 3Rw' Rw 3Fw2 L Lw 3Rw' 3Fw' B2 3Rw2 U2 F Lw' D Fw L2 F 3Rw 3Fw2 Bw

// CENTERS
U2 l' 3U2 F' 3r L u2 R2 3u' 2u U' 2F' d2 B' u2 D F 3r
z' x' F U' 3R' l D 3r U r' F 3r U 3r' U B' r2
3r U 3r2 B l U' l2 3l F 3d2 R2 F2 3d' U // 2 = 50
F' B' 3R B' r2 l' F' D L2 x'
r2 3r' U F2 3r' r' 4r2 F 3r F2 3r F B' r2 l' D' l' // 3 = 78
5f2 U' 3L' U l2 3l U' 3R U2 r l B 3l D2 3r // 4 = 94
F2 3r f' b L2 f b' l F' 3L F' l' F 3L' // 6 = 110

// EDGES
y' R2 D r' (D' L' D) 3l2 // 1 = 117
F R' F' U R' U' l' // 2 = 124
R2 F R2 F2 R F r2 3r' // 3 = 132
F R F' r' L F L' F' 3r' // 4 = 141
U' R' U B L' B' 3L' // 5+6 = 149
U' L U F L' F' l' // 7 = 156
F' L' F D R' D' l' r' // 8 = 164
F R' F' r 3r 4r2 D 3L // 9 = 173
U' R2 U 3L2 r U' L2 U r2 3L // 10 = 185
F' L' F r l' U' L2 U B L B' 4r // 12 = 197 (+87)

// 3x3x3
z2 F2 R B R' B F2 // EO
L' U' (B' D B U' B' D' B') U // left block (+corner cycle)
D' L B2 L' D' L2 // middle slices
U2 R' U R U2 R U // Solved = 226 (+29)


The current UWR is 126 which is 24 moves more than my result. However, I'm not sure if this counts since this attempt literally took me 3 days.
For me it's perfectly valid.
We're playing a serious game and we need time for it. ;)

It took me 3 days also.

I need to improved my skills for 3x3 FMC... :confused: :?
 
Last edited:

Kit Clement

Premium Member
Joined
Aug 25, 2008
Messages
1,606
Location
Portland, OR
WCA
2008CLEM01
YouTube
Visit Channel
scramble: U R2 D F2 D' L2 U B2 F2 R B R' F' L2 D2 L U' L' B R

z2 B2 D' U' B2 D2 //1x2x3
U D' F' U2 D R' U' D U' B' U B2 //F2L-1
R' U' R2 B' R' B D//F2L+EO
//25 to 3e2c2c

I'm stuck. There are no good insertions and the ZBLL is just too long.

Just something quick I found using your psF2L-1:

z2 B2 D' U' B2 //1x2x3 (4/4)
U D F' U2 D R' U2 D B' U B2 //F2L-1 (11/15)
B U2 B' U R' U' R U R' U R U2 D //3c (12/27)
 

RedstoneTim

Member
Joined
Jul 21, 2019
Messages
216
Location
Germany
YouTube
Visit Channel
scramble: U R2 D F2 D' L2 U B2 F2 R B R' F' L2 D2 L U' L' B R

z2 B2 D' U' B2 D2 //1x2x3
U D' F' U2 D R' U' D U' B' U B2 //F2L-1
R' U' R2 B' R' B D//F2L+EO
//25 to 3e2c2c

I'm stuck. There are no good insertions and the ZBLL is just too long.
A few general tips:
- Don't use rotations: Although they do not count towards the movecount (it's 24, not 25 moves), they make NISS and move cancelling (if they're used inside the solution) harder.
- Cancel moves: I'd recommend you to write down a skeleton without any comments. It becomes quite obvious which moves can be cancelled then: z2 B2 D' U' B2 D2 U D' F' U2 D R' U' D U' B' U B2 R' U' R2 B' R' B D, so it's actually just 22 now!
- Learn and use NISS: It can be really helpful for situations like these where you just cannot find anything better.
- If the finish doesn't look good, try to find a different start or change earlier parts of your solution.

Specifically for the solve, a few tips:
You can actually do what you did in one move less by replacing R' U' R2 B' R' B with R' F R' F' R2. You might know this from Dylan Wang's 21 NR, which could've been a 20 if he had used that alg.
If you do use your alg though, you can slightly modify it: R' U' R2 L' B' R' B L. This gives a better ZBLL case. You can actually reduce to 3c from here using a back antisune: R' U2 R U R' U R U'. This sadly only gives 32 to 3c, which ends in a 37, but I think it's still a good example of how knowing lots of different insertion algorithms can help you to force better cases, even if it seems like you couldn't do anything else.
For a solution with a lower movecount, you could either do what Kit Clement recommended above after your F2L-1 or go back further and do the following:

z2 B2 D' U' B2 D // Dr:y (5/5)
U F' U2 D R' U2 D B' // fE + Dr:y' (8/13)
U2 B' U B U2 B2 // EO() + DLf:y' + BLE (6/19)
U2 R' U R U2 R' U' R D // 3c (7/28)


I've oriented the edges a bit earlier in the solve, which created some blocks, which in turn lead to 3c. This is an example of what I meant with "change earlier parts of your solution" before.
 
Top