Random Cubing Discussion

[Anonymous]

I was just commenting. OCLL + PLL, COLL + EPLL, COELL + CPLL, they're all perfectly legitimate LL techniques for methods like Petrus. I was mostly talking about what the first guy said, which was something to the effect of "I was disappointed that the LL system was the same".

 

[KYLOL]

Just from my personal experience - The fastest way to do the final layer with the Petrus method is to force OLL skips ( Study a bit of Winter Variation for tips ). I get skips around 8-10% of the time if I had to make a quick estimate. I should clarify that I manipulate the Last layer very early, sometimes I can see that I have 2-3 LL corners already oriented so I will try to build the rest of my F2L while preserving them. It does become difficult because the move count starts to pile up a bit, but I think it still trims a bit of time in the end. And a quick update for anyone who cares - I'm average low 13's with the Petrus method. Hi Thom and Michael.

edit: typos

 

[Kirjava]

Oh hey Kyle. Be faster.

 

[oll+phase+sync]

The last 2 month I spend improving my Petrus times. (but I believe all this applies to ZZ as well)

1. I tried EJF2L (I an Im case of MGLS) to simplify Step 2 - (it's appliable in Step 4 , too) --- MY PROBLEM IS - Most time in Step 2 I need for finding the 3 pieces, not in solving them.

2. COLL + "Last 5 Edges" (<- From L2L4) {{to simplify Step 2 - (it's appliable in Step 4 , too) }} --- That really means only tracking 2 pieces instead of 3. (instead of the F2L EDGE you place any U Layer Edge) --- MY PROBLEM IS if I place th U-Layer Edge missoriented I have to use extra mind power for Step 3 (Edge Orientation)

3. Partial COLL - When I first read on this forum COLL recognition would be faster than PLL recognition I didn't believe that, but meanwhile I recognize many of my favorite fast COLLs the moment I see the OCLL case --- This definitivley makes me faster. (and I think it's a quite common strategy )

4. Phase recognition - that means in sometimes at the End of F2L (step 4) I just see what the edge case is (If its just 1 move I still aktivly phase). Now I just pick a suitable COLL (e.g. Sune or Antisune or Niklas), that way I'm able to narrow down the range of possible PLLs (often down to just 2)
This sounds compex, but even selecting and doing an N-perm is faster for me than full PLL recognition + execution for my favorite PLL.
- My personal feeling is that this way of solving is the fastest LastLayer stategy I can do currently.

EXPERIMENTALs

STEP2 MGLS: The idea is simple - plan ahead full step1+2 during inspection ... since this is pretty much impossible skip one Corner completely to make it easier - this is very similar to EJF2L - This would almost be a full step skip every solve - but it's above my processing power

My next MGLS idea was to merge Step 2 and Step 3 together (and again not solving the Step 2 corner at all ) - that means orienting all edges wihle placing both Step 2 edges, --- It feels a little bit like EOLINE to me (with the bonus that at least 4 edges are already oriented), maybe it would be a good simplyfication to only orient U and D edges during this merging step ???

I would be interessted if anyone tried things like this?

P.S. My Roux times improved even without traininig, so I didn't fullfill my signature goal

 

[Cyrus C.]

When doing the Step2 MGLS you have to make sure that the corner is a U layer corner.

 

[KYLOL]

Oh hey Kyle. Be faster.

I'm trying to get good at BLD. Leave me alone!
I'm too fat to be faster.

 

[oll+phase+sync]

When doing the Step2 MGLS you have to make sure that the corner is a U layer corner.

Right - Odds are 5 good corners against 2 bad - I would hope its not so much of a problem if you spot them early enough.

 

[qqwref]

Does anyone do the "L-perm" this way?

R' U' R B R' U' R U R B' R2' U R U
or: R' U' l U l' U' R U l U' l' R' U R U

 

[blade740]

How I recognize Square-1 Permutation using blocks and bars.

Also applies to 3x3 PLL and 4x4 (redux) parity PLL.

http://crunchatize.me/square-1-pll-recognition/

 

[qqwref]

Cool site, very nice diagrams (might also be useful for 4x4, arguably?). It's really too bad there isn't any way to represent the Ph perm with one letter, though... I like your E/W/Q recognition idea.

I wonder, for all you guys with the WCA database and SQL experience, has anyone (i.e. list the people who) improved their personal best single or average, in any event, by a multiple of 10.00 seconds? (Anything where the hundredths weren't recorded doesn't count, of course, since you can't tell if it was an exact multiple or not.)

 

[ben1996123]

R U' R2 U2 R r U R' U' r' U2 R U

pure 2 flip adjacent. Ive used it in F2L for ages but only just realised it was pure with a U at the end can sub 2 it easily.

 

[irontwig]

Pretty prå alg.

 

[Kenneth]

Supercube centre piece orientation comutator:

A : M' E M
B : U/U'/U2

A setup the R-centre to the U-layer, B orients it to any position you like, A' restores centres to their correct positions and B' orients the U-centre in the opposite direction compared to R.

Example, orients R and U' : (M' E M) U (M' E' M) U'

 

[DavidWoner]

2.06 F' R2 F2 U R2 U F' R2 F2 R' U'

z R' U R U' R' U' R U' R' / R2' F2 R2 U

I'd been wanting a new non-separating alg for that orientation case for ages, so I made it up during inspection

 

[Anthony]

I'd been wanting a new non-separating alg for that orientation case for ages, so I made it up during inspection

Like a boss.

 

[JackJ]

My solution to Erik's infamous 7.08.

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

F L' U' D
F R' F' R
U' y' R U2 R' U' R U R'
L' U L2 U' L'
U L' U2 L2 F' L' F
U2 F R U R' U' F' f R U R' U' f'
R2 u' R U' R U R' u R2' y R U' R' U2

F L' U' D F R' F' R U' y' R U2 R' U' R U R' L' U L2 U' L' U L' U2 L2 F' L' F U2 F R U R' U' F' f R U R' U' f' R2 u' R U' R U R' u R2' y R U' R' U2

 

[Anthony]

2.06 F' R2 F2 U R2 U F' R2 F2 R' U'

z R' U R U' R' U' R U' R' / R2' F2 R2 U

I just tried it and that's awesome. I hated doing F' L F L' F' L F L' for that case.

 

[Kirjava]

20:23:25 <+Kirjava> YESSS
20:23:31 <+Kirjava> 9.25 R2 D2 U2 L D2 B F' U2 F' U2 F' U' L2 F' L2 B' R D2 R2 L' F B R2 F2 U2
20:23:36 <+Kirjava> non matching
20:24:17 <+Kirjava> x2 y' U'R2UL z'
20:24:30 < Asdfgfdsa> Kirjava: pro master epic champion :O
20:24:42 < jernqvist> lol4pairs
20:24:48 <+Kirjava> r2UR'U'RUR2U'R2UR
20:25:05 <+Kirjava> RU2R'U'RU'R'
20:26:09 <+Kirjava> M'U(M'M2)UM2UM'U2LR
20:26:21 < statuek> ,seen jernqvist
20:26:23 <+Nibblr> jernqvist was seen speaking in #rubik 1m40s ago
20:26:26 < statuek> ,seen j`ey
20:26:26 < statuek> *
20:26:28 <+Nibblr> j`ey was seen speaking in #rubik 5h42m48s ago
20:26:34 <+Kirjava> 32 mvoes lol

 

[Xishem]

Could anyone give me an average movecount for 5x5 (or 4x4, but preferably 5) centers for reduction in STM? I'd like to know what fast people average.

 

[DavidWoner]

I just tried it and that's awesome. I hated doing F' L F L' F' L F L' for that case.

Well that's just sexy move written a stupid way. Previously I was using R U2 R' U' R2 U' R2'

But I like that sexymove version for preserving FL permutation, as long as you hold it at a different angle.