The New Method / Substep / Concept Idea Thread

[Thermex]

@Aerma Sue's right that would be a rediculous number of algs, but doing L5E instead (I guess this has been nicknamed "backbone") could have potential. The 2×2 method you came up with was basically just doing HD but permuting the bottom corners during LOLS wihich is much less alg-efficient.

 

[Neuro]

ZZ-b has more than 200 algs if you use the full 58 phasing algs.

Thanks for bringing that to my attention, I didn't think about the # of phasing algs just the ZZLL's! I may still keep it in my stats though simply because the number of algs is still well within reason to learn and the method can be pretty fast.

 

[Thermex]

Wait here's another megaminx LS+LL idea:

1. Put the DRF corner in its slot (correctly oriented) and orient the edges (~40 really short algs)
2. COLL (192 cases, problem is the recog might be complete garbage)
3. L6EP (EPLL+FR edge, again this might have terrible recog and a ton of cases)

 

[Shiv3r]

@Neuro Speaking of ZZ, I decided to do an FMC solve, and ended up doing ZZ-WV and getting a 45-move solve. Here's the solve(there are no rotations, so you will have to pan around to see what I'm doing)

 

[Shiv3r]

I came up with an Idea to skip OLL, it's basically 2look VLS and it sucks. but here it is, since I haven't talked about it before here:
-F2L-1
-insert the D layer corner while orienting all the edges(16 algs, twice as many algs as VHLS)
-BLE

 

[Sue Doenim]

I came up with an Idea to skip OLL, it's basically 2look VLS and it sucks. but here it is, since I haven't talked about it before here:
-F2L-1
-insert the D layer corner while orienting all the edges(16 algs, twice as many algs as VHLS)
-BLE

It's kind of like MGLS, but with less algs, right? It actually seems okay, at least for intermediate cubing.
My Hexagonal Francisco method from a while ago. It's sort of a 2GR HF combo thing.
Recap:
HF+
1a) 1x2x3 on D (<R,U,F,r,u,f>)
1b) Attach DLF and solve CP
2a) E-slice edges -1
2b) Insert last E-slice edge and force arrow case
3) L5C
4) L5E
Example solve:
Scramble: F2 B2 U2 R2 L F U' D2 L' D' B D2 F' R2 U2 B L2 F2 R2 U2 B
Inspection- z2
Block- F' U' f2 E F2 R' (6/6)
CP- R U R' D R U' R' D' (8/14)
E-slice- U R U' R' u' U' r U r' u2 U r U r' (14/28)
Last edge + 3-flip- E' r U' r' (4/32)
L5C- R U2 R2 U2 R' U R2 U' R U2 R (11/43)
L5E- M' U M u' M' U2 M' U M U2 M' U M2 (13/56)
56 moves. This is around ZZ-VH movecount, isn't it?
Note: I don't actually know how to fix CP, I just tried stuff until I found something that worked; Last E-slice edge alg was intuitive; L5C was from cube explorer; L5E was just a cancelled U-perm, so if this is to become a method, we should have all L5E arrow cases.

 

[xyzzy]

I came up with an Idea to skip OLL, it's basically 2look VLS and it sucks. but here it is, since I haven't talked about it before here:
-F2L-1
-insert the D layer corner while orienting all the edges(16 algs, twice as many algs as VHLS)
-BLE

It's JJLS.

 

[Neuro]

So for CFCE what I'm thinking is this: 3 cross edges, F2L (last one you either ensure a 4 flip on top or full EO), a CxLL subset that preserves EO and solves it for 4 flip and L5E.

F2L should be pretty intuitive and CxLL would theoretically have double the amount of CMLL and by putting all bad edges on the top there's no issue with symmetry. And L5E is pretty straightforward b/c you already have EO done. So basically you can get pretty similar TPS/lookahead as you can with CFOP with a theoretically lower movecount while having pretty low # of algs and simple recognition. My biggest gripe would be with the last F2L pair and EO recog but it's really not that bad. Not to mention the other alternative is learning like 16 algs per CxLL case. Any thoughts on this? I'll add a solve soon

 

[Shiv3r]

So for CFCE what I'm thinking is this: 3 cross edges, F2L (last one you either ensure a 4 flip on top or full EO), a CxLL subset that preserves EO and solves it for 4 flip and L5E.

F2L should be pretty intuitive and CxLL would theoretically have double the amount of CMLL and by putting all bad edges on the top there's no issue with symmetry. And L5E is pretty straightforward b/c you already have EO done. So basically you can get pretty similar TPS/lookahead as you can with CFOP with a theoretically lower movecount while having pretty low # of algs and simple recognition. My biggest gripe would be with the last F2L pair and EO recog but it's really not that bad. Not to mention the other alternative is learning like 16 algs per CxLL case. Any thoughts on this? I'll add a solve soon

didn't crafto22 just propose this exact same thing a while back?

 

[Hazel]

I know Ivy Cube isn't an official event or anything, but if it was someday that I think I might have come up with an okay advanced method for it:
1) Solve 3 centers around a corner
2) Solve the rest with one alg

 

[Hazel]

So for CFCE what I'm thinking is this: 3 cross edges, F2L (last one you either ensure a 4 flip on top or full EO), a CxLL subset that preserves EO and solves it for 4 flip and L5E.

F2L should be pretty intuitive and CxLL would theoretically have double the amount of CMLL and by putting all bad edges on the top there's no issue with symmetry. And L5E is pretty straightforward b/c you already have EO done. So basically you can get pretty similar TPS/lookahead as you can with CFOP with a theoretically lower movecount while having pretty low # of algs and simple recognition. My biggest gripe would be with the last F2L pair and EO recog but it's really not that bad. Not to mention the other alternative is learning like 16 algs per CxLL case. Any thoughts on this? I'll add a solve soon

I've heard that this only saves maybe 1 move for cross, and even makes F2L LESS efficient :/

 

[Shiv3r]

I

I know Ivy Cube isn't an official event or anything, but if it was someday that I think I might have come up with an okay advanced method for it:
1) Solve 3 centers around a corner
2) Solve the rest with one alg

The Ivy cube is just a pyraminx without the tips and styled differently
the method you are talking about is WO for pyraminx.

 

[Neuro]

Has this been proposed before?

1. 3 cross edges, last one is any misoriented U-layer edge (5 moves)
2. F2L-1 (20 moves)
3. Last pair whilst ensuring 3 U-layer edges are unoriented (~7 moves using 41 "new" algs, or ~10 moves with 2 "new" algs)
4. COLL (10 moves)
5. L5E with the easiest EO case, probably possible to one-look almost every time (~10 moves)
Total would be 52 moves with 41 new algs, which is better than regular CFOP, all while making cross easier and use of insepction time more efficient.

Basic algs for inserting the last pair:
Connected pair, edges need to be flipped: y R' F' R U2 M' U' M
Disconnected pair, edges need to be flipped: R' D' r U r' D' R2 U R'

@Shiv3r Not exactly, he forced an MO edge into the last cross slot and used an arrow case while solving L5E with only one possibility for L5E that's harder to recognize.

What I think I'd do differently now is to do 3 cross edges while forcing a in either an oriented U layer edge OR last cross edge, do F2L while forcing either a cross OLL or EO done (should be pretty intuitive), and CxLL that either does EO or preserves it and L5EP. Easy cross, F2L may be a problem but it should still be fast, CxLL is really easy as proven by COLL, CMLL, etc., and L5EP is pretty easy to recognize and it's entirely <M,U> or <R,U> with a rotation. Also easier lookahead into F2L may help to cancel out the slowish EO portion later in the solve.

EDIT: Alternatively, you can do the EO influence, and do COLL and use 2 sets of L5E. I think that both variants would be good but perhaps the EO+CxLL variant is better (recognition of misoriented L5E probably isn't that great even with rotational symmetry)

Solve with v1 45 STM
Solve with v2 43 STM

I've done several solves with the 1st variant and every solve has been under 55 STM so far, should be below CFOP's average movecount even with the awkward EO influence

 

[Sue Doenim]

Here's kind of an alternative way of doing WVCP.
1) EOF2L-1
2) Solve last F2L corner and solve CP (26 algorithms)
3) Solve last F2L edge and solve corner orientation (27 algorithms)

Also, I've thought of a better way of detecting CP during last slot. When an F2L corner is out, 3 U-layer corners are visible. If you know your COLL recognition well enough (from all angles), you can look at these three corners and know what the last U-layer corner would be were it in the F2L corners place (e.g. do R U R' on a solved cube. If the U-layer corner in DFR were in ULF, it would have to have it's U color sticker on the F face, bringing about a headlights case.). You can then tell the "COLL case" from the colors you can see, and identify the pieces that are switched. ( e.g. if you do R U R', the "COLL case" formed by the corners switches UBR and UFR when applied.). Here, you take this swap, and update it by swapping the 2 corners before the F2L corners, moving in a clock wise direction (e.g. after R U R', you would swap UBR and UFR.). This swap is the byproduct of inserting the F2L pair. You then take both swaps, and account for both while inserting your last F2L pair, giving you an <R,U> solvable corner subset.

With this recognition method in place, I'd like to propose what I would call 2GLS: a VHLS type alg set where CP and EO are solved, leaving you with a 2GLL case to finish.

 

[Neuro]

Sorry for spam but here's a quick idea I came up with that should be a better alternative than full COLS. Basically you can do WV, SV, BLE, and 1 set in which the corner is in it's slot but the edge isn't. For most of the LS cases you can get to any one of these cases in ~3 moves assuming you wish to use it. Altogether it has 129 algs that are of course all 2 gen. I can see it being quite useful for ZZ, particularly OH. Let me know your thoughts!

EDIT: My bad, there are actually 2 sets for corner in slot -edge not including BLE! So 5 sets of 27 gives 135 algs, still pretty reasonable IMO.

 

[Hazel]

Potential 2x2 method, not super advanced like EG but maybe viable:
1) Solve a "layer". One of the layer pieces doesn't have to be the correct one, as long as it's oriented properly.
2) OLL
3) 2x2 variant of TTLL

There's also 3 variants of this I thought of:

1) "layer"
2) CLL (solving permutation of the 3 corners on U layer that are supposed to be on U layer, but same algs as CLL)
3) use the same alg every time to solve TTLL

or

1) "layer"
2) a form of CLL that would make it so you could do a R2 U R2 U' R2 immediately after and solve the cube, or cancel into it.

or

1) "layer"
2) L5C

 

[Thermex]

@Aerma I didn't necessarily understand everything you just said, but I'm pretty sure your first idea was this?

1. Layer, but one piece from the opposite layer can be in it
2. OLL
3. TTLL on 2×2 (6 algs)

I actually think I might teach some of my friends this; I think anyone planning on learning HD at some point could use this as a stepping stone.

A new "pathway" of methods I might teach some of my friends:

1. Beginner's LBL (LBL with only sunes and T-perms)
2. LBL (layer, OLL, PLL)
3. Beginner's VOP (the method you described earlier that ends it TTLL)
4. VOP
5. HD

 

[Hazel]

Finally, a method I proposed that I don't have to put in the trash can

 

[VenomCubing]

Real quick thought: maybe during the insertion of the last f2l pair, a solver could make sure that at least one last layer corner is oriented. Then for the last layer, the LL edges could be oriented and permuted to be solved with an oriented corner (one algorithm), leaving the last three corners to be solved, so L3C would be the last step. Any feedback?

 

[shadowslice e]

Real quick thought: maybe during the insertion of the last f2l pair, a solver could make sure that at least one last layer corner is oriented. Then for the last layer, the LL edges could be oriented and permuted to be solved with an oriented corner (one algorithm), leaving the last three corners to be solved, so L3C would be the last step. Any feedback?

That's fish and chips or snyder LL. It's not too bad anD better in terms of movecount though you can get really horrible cases like 3 twisted corners