The New Method / Substep / Concept Idea Thread

[Teoidus]

I dont' think it'd be as good as Cardan

 

[Shiv3r]

I dont' think it'd be as good as Cardan

what's Cardan?

 

[Teoidus]

https://www.speedsolving.com/forum/threads/cardan-reduction-novel-ls-ll-approach.64042/

 

[Shiv3r]

https://www.speedsolving.com/forum/threads/cardan-reduction-novel-ls-ll-approach.64042/

mmk. Trying the NLS though, It turns out I know a LOT of the TELL algs already, cuz I know all the :
-Pure flips
-L3C(Comms, lel)
-PLL

Also, NLs does not have the problems of other LS methods, as for those you need to set up the Pair.
In fact, I can see why ZZ-CT is more Viable than the numerous ZZ-Ls methods out there(besides maybe ZZ-b): because the algs can easily be mirrored to any slot, which allows for freer blockbuilding, and since you solve the LS edge, then corner, you do not need to waste a look pairing up the pair, and also you do not use extra moves pairing it up.

 

[Shiv3r]

Okay, so. Random Idea:
Since a lot of ZZ users know Most of full COLL except for Sune/antisune, because of crappy recog, would it make sense to learn WV, but only the cases that with a normal insert would give you a sune?

This stemmed from an Idea that I had for never ever getting an S/AS COLL, by using 1 or 2 algs for all the cases that would usually give you a sune with an (U) R U' R' insert. I called it NICE(Never Insert & get S/AS COLL, Ever.)

 

[Neuro]

NLS itself has 35 algs and TELL (MUCH better name than Tri ZBLL IMO) has 50 algs, total of 85 algs. But a lot of NLS can be done intuitively, and the LL recog isn't too bad.

 

[Y2k1]

Okay, so. Random Idea:
Since a lot of ZZ users know Most of full COLL except for Sune/antisune, because of crappy recog, would it make sense to learn WV, but only the cases that with a normal insert would give you a sune?

This stemmed from an Idea that I had for never ever getting an S/AS COLL, by using 1 or 2 algs for all the cases that would usually give you a sune with an (U) R U' R' insert. I called it NICE(Never Insert & get S/AS COLL, Ever.)

Or... Look up how to recog s/as and just git gud with full COLL (really though the algs are really nice, and the recog for me is faster than sune+pll) Look up 3 sticker sune recog by ceitical cubing

 

[TDM]

Okay, so. Random Idea:
Since a lot of ZZ users know Most of full COLL except for Sune/antisune, because of crappy recog, would it make sense to learn WV, but only the cases that with a normal insert would give you a sune?

This stemmed from an Idea that I had for never ever getting an S/AS COLL, by using 1 or 2 algs for all the cases that would usually give you a sune with an (U) R U' R' insert. I called it NICE(Never Insert & get S/AS COLL, Ever.)

People don't dislike A/S COLLs because they're bad. They're not. It's just that for those cases, OCLL->PLL is faster. They're not slow; there's no need to avoid them!

 

[efattah]

LMCF revision 4:
https://drive.google.com/open?id=0B2QnZ3uD6I8kNkpHSURSbzluc2s

I have added new L6E sets (pages 13-20), and of interest to Roux solvers is the move count comparison of the L6E sets (to solve UR/UL + orient midges):

DFL set: 8.12

xDFL set: 9.00

rDFR set: 7.87

iDFR set: 8.50

This is really interesting; it means that if you choose the 2-look method for L6E, you are better off solving one of UL/UR in the wrong (opposite) location and disoriented, resulting in the rDFR set which has an average move count of 7.87 (and one case that is only 3 moves). Compare this to if you actually SOLVE UR or UL, you end up with the DFL set (8.12). The other sets (xDFL and iDFR) are the situations where you place UR or UL in place but disoriented, or in the opposite spot but oriented.

 

[Calode]

People don't dislike A/S COLLs because they're bad. They're not. It's just that for those cases, OCLL->PLL is faster. They're not slow; there's no need to avoid them!

What if you avoided them to setup to a zbll assuming you knew full zbll - sunes?

 

[Shiv3r]

What if you avoided them to setup to a zbll assuming you knew full zbll - sunes?

that may make sense. if you used the one where you learn the full WV case(by just looking would be about 8 cases if Im not wrong, 2 OCLL's * 4 AUF's, one of those cases is R U2 R')

 

[Sue Doenim]

What if you avoided them to setup to a zbll assuming you knew full zbll - sunes?

It would be better just to use phasing and ZZLL; you don't have to build the F2L pair, and there's less algs.
Also, wouldn't there just be two algs: insert normally or if doing so would give you a sune, insert with U2.
Also, what does sune mean? Is it named after someone like Niklas?

 

[Shiv3r]

It would be better just to use phasing and ZZLL; you don't have to build the F2L pair, and there's less algs.
Also, wouldn't there just be two algs: insert normally or if doing so would give you a sune, insert with U2.
Also, what does sune mean? Is it named after someone like Niklas?

both were coined By lars petrus.
you can see them here. lar5.com/cube
also Niklas wasnt named after a dude.

 

[crafto22]

Hey guys, just found some algs I started generating last year for ECE. It's basically just CO of both layers whilst preserving the E-slice. Algs are actually really nice, anyone know how we could make use of this?

https://docs.google.com/spreadsheet...764CGP4jgr8nBQjzbwWCkfiZ3Fo/edit?usp=drivesdk

 

[Shiv3r]

Guys, this has been my work in progress for a while now, here it is.
https://docs.google.com/document/d/1wyhyaUgKCZ7-bOrtdwgTQUlAHUPsN9zBLjhDvBG6VKY/edit?usp=sharing

 

[efattah]

Interesting idea for a Roux-Waterman hybrid, where you can decide in the middle of the solve:

1. Solve a 2x3x1 left block just like in normal Roux.
2. Start solving the right block. If things are going well finish with Roux.
2B. If the second block is not favorable, put in the two bottom right corners and any edges that are easy & convenient, and ignore the rest of the block's edges.
3. CMLL.
4. If you chose (2B) you are now in Waterman's phase, you can finish the solve with Waterman or you can keyhole the last edge or two of the right block and finish with Roux LSE.

 

[Neuro]

Continuing with efattah's post, I had this idea as a modifiation for Waterman, let me know your thoughts

1: FB// Same as Roux, place of left
2: DR// Solve a square in DR
3: L5C+// Solve L5C and last R edge. Many ways to do this, but I'd recommend having LRE on top and oriented while solving LDC and then using an as of now unnamed subset to solve LLC and LRE.

Zero rotations and makes it really easy to decide between Roux and Waterman to solve the puzzle. If RB is difficult then use Waterman and if it's easy use Roux. Only problem I see is the number of algs for L5C+, but it seems a lot better than the initial Waterman proposal. IDK about the efficiency though, I'd have to do some tests.

 

[efattah]

Continuing with efattah's post, I had this idea as a modifiation for Waterman, let me know your thoughts

1: FB// Same as Roux, place of left
2: DR// Solve a square in DR
3: L5C+// Solve L5C and last R edge. Many ways to do this, but I'd recommend having LRE on top and oriented while solving LDC and then using an as of now unnamed subset to solve LLC and LRE.

Zero rotations and makes it really easy to decide between Roux and Waterman to solve the puzzle. If RB is difficult then use Waterman and if it's easy use Roux. Only problem I see is the number of algs for L5C+, but it seems a lot better than the initial Waterman proposal. IDK about the efficiency though, I'd have to do some tests.

I like this! I would suggest a further refinement:
1. FB on left
2. Solve a square on DR
3. Solve last 5 corners in one step (probably involves an AUF set up move) ignoring FR edge
4. Solve L7E (ordinary LSE slots + FR)

I am generating the L7E algorithms as we speak for the advanced section of the LMCF document and so far the L7E algorithms look amazing. Super ergonomic [R, r, U, M] and the movecount average is looking to be [0-3 setup]+[10.7 move algorithm to solve L/R and orient midges] + [4.5 to permute midges] = 16.7 moves for L7E.

 

[Neuro]

I like this! I would suggest a further refinement:
1. FB on left
2. Solve a square on DR
3. Solve last 5 corners in one step (probably involves an AUF set up move) ignoring FR edge
4. Solve L7E (ordinary LSE slots + FR)

I am generating the L7E algorithms as we speak for the advanced section of the LMCF document and so far the L7E algorithms look amazing. Super ergonomic [R, r, U, M] and the movecount average is looking to be [0-3 setup]+[10.7 move algorithm to solve L/R and orient midges] + [4.5 to permute midges] = 16.7 moves for L7E.

PM me and maybe we can make this a thing. I wasn't sure about what to do for L5C+ and I see that there's probably better ways to do it. Not sure about L7E but does sound feasible

 

[Sue Doenim]

Quick idea: To lessen ZBLS alg count and improve ergonomics, orient last F2L edge during insertion of second-to-last F2L pair. Also, does anyone know where to find a set of algs for ZBLS other than Lars Vandenbergh's?