The New Method / Substep / Concept Idea Thread

[Sue Doenim]

If you solve the corners and you solve 2 edges that are opposite to each other, then you have an lse similar to roux. Assuming you used csp, this would be 90 algs or less. Also, it would only be around 8 algs I think if you solve one of the edges with an m2 first.

That looks like a way better choice. 70 algs, actually, probably way better recog. Also looks nice for progression, EPLL+L5E, 16 algs, then swapped EPLL, 5 more, swapped L5E, 12, then the 36 others.

 

[Arc]

Here's an interesting idea of a method.

1. CPFB - 1x2x3 on left while reducing the remaining 6 corners to the permutation subgroup that contains solved.
2. Magic - Orient the F2L edges (DR, FR, BR) and the UL/UR edges, and insert the UL/UR on D.
   
3. F2L - F2L just like Petrus/Briggs/2GR/ZZ/Leor/lot's of stuff.
4. OLL - Solved CP subset of OLLCP, 57 algs. Can be done with SuneOLL.
5. L4EP - Insert UL/UR with an M2 and permute the remaining 4 unsolved edges (Roux step 4c).

 

[Cale S]

Of course. Have you continued with this method? Are you satisfied?

I generated algs for it but I actually forgot about it until now

 

[Metallic Silver]

I wonder, is this squan method "humanly" possible?

1. CSP
2. Orient Both Layers
3. Permute Both Layers

 

[Metallic Silver]

Or why not Square-1 Columns First?

1. Cube Shape/CSP
2. Solve all Corners
3. Solve all Edges

 

[Thom S.]

I wonder, is this squan method "humanly" possible?

1. CSP
2. Orient Both Layers
3. Permute Both Layers

Not really in the near future. OBL can easily be used but Even-Parity PBL is about 800 Algorithms(If I recall it correctly) and as you need a really solid recognition, it would be a slow, hard process

 

[Thom S.]

Or why not Square-1 Columns First?

1. Cube Shape/CSP
2. Solve all Corners
3. Solve all Edges

Too hard recognition, no fluid CS-Step 2 transition

 

[Hazel]

I wonder, is this squan method "humanly" possible?

1. CSP
2. Orient Both Layers
3. Permute Both Layers

I'm not a squan person, but how would a method like this be on 2x2?
1: separate pieces into their correct layers
2: orient both layers
3: PBL

 

[Neuro]

I'm not a squan person, but how would a method like this be on 2x2?
1: separate pieces into their correct layers
2: orient both layers
3: PBL

Just use Ortega. Separation wouldn't be a good step, and there's be a lot of (probably not good) orientation algs. Ortega is easier and probably much faster

 

[Thermex]

Here's an interesting idea of a method.

1. CPFB - 1x2x3 on left while reducing the remaining 6 corners to the permutation subgroup that contains solved.
2. Magic - Orient the F2L edges (DR, FR, BR) and the UL/UR edges, and insert the UL/UR on D.
   
3. F2L - F2L just like Petrus/Briggs/2GR/ZZ/Leor/lot's of stuff.
4. OLL - Solved CP subset of OLLCP, 57 algs. Can be done with SuneOLL.
5. L4EP - Insert UL/UR with an M2 and permute the remaining 4 unsolved edges (Roux step 4c).

This looks like a pretty decent idea, and could probably be a decent transition method from CFOP to Roux. Problem is I just don't really see how this is better in any way than Vanilla Roux. Also, how exactly would step 2 be done? (maybe you could show some example solves.)

That looks like a way better choice. 70 algs, actually, probably way better recog. Also looks nice for progression, EPLL+L5E, 16 algs, then swapped EPLL, 5 more, swapped L5E, 12, then the 36 others.

I'm still not understanding what the second and third steps are, are they part of a template for a method or actually concrete steps? Either way, here's three random CSP -> Blocks -> Corners -> Edges style squan methods possibly similar to yours:

a.) An extremely simple Roux n' Screw/Lin variant (might have been proposed before):
1. CSP
2. FB (1x1x3 on DL)
3. "SB" (DBR+DR)
4. L5C (5 algs)
5. LSE (Either algorithmic or intuitive works)

b.) Bope (pretty much same steps as the 3x3 method I proposed a while back)
1. CSP
2. FB+ (the left half of the slice on the D-layer)
3. LSC (solve the 6 remaining corners, 50 algs)
4. L6E (this step solves the DF, DR and U-layer edges all at once. Probably around 100 algs or so)

c.) I've proposed this idea before, it seems to me like the fastest of the three but it could be too alg-heavy:
1. CSP (what?!)
2. Solve the orientation of every piece on the D-layer except the DFR and FR pieces (Ultra fast+intuitive)
3. Solve the corners with HD NLL (36 fairly short algs)
4. Solve the edges (~200 algs if I'm calculating this right)

Any of these seem viable? Which should I try to continue to develop?

 

[Sue Doenim]

I'm still not understanding what the second and third steps are, are they part of a template for a method or actually concrete steps? Either way, here's three random CSP -> Blocks -> Corners -> Edges style squan methods possibly similar to yours:

a.) Extremely simple Roux n' Screw/Lin variant (might have been proposed before):
1. CSP
2. FB (1x1x3 on DL)
3. "SB" (DBR+DR)
4. L5C (5 algs)
5. LSE (Either algorithmic or intuitive works)

b.) Bope (pretty much same steps as the 3x3 method I proposed a while back)
1. CSP
2. FB+ (the left half of the slice on the D-layer)
3. LSC (solve the 6 remaining corners, 50 algs)
4. L6E (this step solves the DF, DR and U-layer edges all at once. Probably around 100 algs or so)

c.) I've proposed this idea before, it seems to me like the fastest of the three but it could be too alg-heavy:
1. CSP (what?!)
2. Solve the orientation of every piece on the D-layer except the DFR and FR pieces (Ultra fast+intuitive)
3. Solve the corners with HD NLL (36 fairly short algs)
4. Solve the edges (~200 algs if I'm calculating this right)

Any of these seem viable? Which should I try to continue to develop?

What my method is at this point is:
1) CSP
2) Build 2 opposite D layer blocks, force Vs
3) NLL
4) L6E
The 2 D layer blocks must be made of a pair of opposite edges and corners. A green edge with a green red corner and a blue edge with a blue red corner would not work, for example.
Method A looks a lot like Yau-1. B seems really promising, but I think the alg count would be more like 6!/2, or 360. That's why having opposite D-layer edges for L6E is important; it reduces symmetries a ton. That could be worked around, though. If you solve any block before L6C, you would have less than 76, I believe. C is the same kind of deal, by my raw calculations you would have 4*3*2 (number of cases for the 3 oriented D layer edges) *5*4*3 (remaining L5E) = 1440 algs. That's a lot. I'd look most at B.

 

[Arc]

This looks like a pretty decent idea, and could probably be a decent transition method from CFOP to Roux. Problem is I just don't really see how this is better in any way than Vanilla Roux. Also, how exactly would step 2 be done? (maybe you could show some example solves.)

Step 2 is done very similarly to Leor/Briggs but slightly differently because you don't care about the EO of some edges.
Here's an example solve.

 

[Cale S]

I'm not a squan person, but how would a method like this be on 2x2?
1: separate pieces into their correct layers
2: orient both layers
3: PBL

I generated the algs for that second step when both layers are OCLLs
bit.do/rolrol

I don't think it's good

 

[Thom S.]

a.) An extremely simple Roux n' Screw/Lin variant (might have been proposed before):
1. CSP
2. FB (1x1x3 on DL)
3. "SB" (DBR+DR)
4. L5C (5 algs)
5. LSE (Either algorithmic or intuitive works)

If your Second block would also solve the DB Edge, this would be Yau-1

b.) Bope (pretty much same steps as the 3x3 method I proposed a while back)
1. CSP
2. FB+ (the left half of the slice on the D-layer)
3. LSC (solve the 6 remaining corners, 50 algs)
4. L6E (this step solves the DF, DR and U-layer edges all at once. Probably around 100 algs or so)

Seems better than a.) as you can easily see your LSC case while inserting the DB edge. I'd be interested if you can develop a way to alter your Algorithms to insert D Layer Edges-like CP for Vandenbergh.(just an Idea for this is that every algorithm doesn't directly solve them- it's a two-phase one, 1-Setup 2-Interchange, this way, the different part would be the blockbuilding during 1. Could work, but I doubt it is relevant for Speedsolving)

c.) I've proposed this idea before, it seems to me like the fastest of the three but it could be too alg-heavy:
1. CSP (what?!)
2. Solve the orientation of every piece on the D-layer except the DFR and FR pieces (Ultra fast+intuitive)
3. Solve the corners with HD NLL (36 fairly short algs)
4. Solve the edges (~200 algs if I'm calculating this right)

I actually like this. I'm generally not a fan of permuting pieces directly after CS as it results in pauses. My best guess is to insert D Layer edges as you could force good LSE/L5E cases

 

[Thermex]

What my method is at this point is:
1) CSP
2) Build 2 opposite D layer blocks, force Vs
3) NLL
4) L6E

I was thinking of including this same method in my list. It looks pretty good to me.

Method A looks a lot like Yau-1.

your Second block would also solve the DB Edge, this would be Yau-1

Good point, though I feel like not solving the DB edge here improves ergonomics a lot. I still probably won't pursue this anyway.

B seems really promising, but I think the alg count would be more like 6!/2, or 360. That's why having opposite D-layer edges for L6E is important; it reduces symmetries a ton. That could be worked around, though. If you solve any block before L6C, you would have less than 76, I believe. C is the same kind of deal, by my raw calculations you would have 4*3*2 (number of cases for the 3 oriented D layer edges) *5*4*3 (remaining L5E) = 1440 algs. That's a lot. I'd look most at B.

Did you account for AUFs? For method B if you bring the edge that goes at DF to UF you reduce the number of cases to 5!/2 (cases with DF in U-layer) + 13 (L5E) + 13 more (swapped L5E) + 4 (EPLL) = 90 cases. For method C there should be more like 14^2 cases since there are only 14 cases for the bottom 3 layer and 14 L5E cases. Correct if I'm wrong but I'm not sure if you included AUFs in your statistics. I'll probably start developing B and C once I'm done with a couple of LxE methods for 3x3 and Twisty EG.

 

[Sue Doenim]

Let's try calculating it by D layer edge positions.
Both solved: 4
Swapped: 5
1 solved: 12
1 swapped: 12x2=24
Adj. on U: 12x2=24
Opp. on U: 12
81 algs. Looks like you're a lot closer.
Let's look at C.
Cases for back 3:
Solved
Adj.x4
Opp.x2
U permx8
O permx2
H perm
Z permx2
W permx4
Total: 24
24x12 L5E +49 PBL = 337 algs. Not quite 1440, but a bunch.

 

[Thermex]

Let's try calculating it by D layer edge positions.
Both solved: 4
Swapped: 5
1 solved: 12
1 swapped: 12x2=24
Adj. on U: 12x2=24
Opp. on U: 12
81 algs. Looks like you're a lot closer.

Cool! This really isn't that much considering PSC is only about 50 algs. I'll probably start working on this method next year.

Let's try calculating it by D layer edge positions.
Let's look at C.
Cases for back 3:
Solved
Adj.x4
Opp.x2
U permx8
O permx2
H perm
Z permx2
W permx4
Total: 24
24x12 L5E +49 PBL = 337 algs. Not quite 1440, but a bunch.

Your calculations look correct to me, I forgot to count the "parity" cases on the D-layer. Here's a replacement method that's pretty close to the original:
1. CSP
2. Solve the orientation of all but one corner on each face (pretty easy to do, could also be done algorithmically)
3. Solve corner permutation on both layers while inserting the remaining edge into the D-layer (48 algs)
4. EP (49 cases)

 

[pizo45]

complex reduction method (CRM)

1. solve all edges in the e slice and orient edges in the u and d layers (intuitive)
2. permute the edges while orienting corners on L (algs)
3. permute the corners while orienting the corners on R (algs)
4. permute the remaining corners

this has the mathematical advantage of permuting and orienting something at the same tiem for every step, just like in the snyder method, i think it would be able to get 40 moves average really really easy

 

[James Hake]

complex reduction method (CRM)

1. solve all edges in the e slice and orient edges in the u and d layers (intuitive)
2. permute the edges while orienting corners on L (algs)
3. permute the corners while orienting the corners on R (algs)
4. permute the remaining corners

this has the mathematical advantage of permuting and orienting something at the same tiem for every step, just like in the snyder method, i think it would be able to get 40 moves average really really easy

Could you provide an example solve? These steps seem rather complex.

 

[shadowslice e]

complex reduction method (CRM)

1. solve all edges in the e slice and orient edges in the u and d layers (intuitive)
2. permute the edges while orienting corners on L (algs)
3. permute the corners while orienting the corners on R (algs)
4. permute the remaining corners

this has the mathematical advantage of permuting and orienting something at the same tiem for every step, just like in the snyder method, i think it would be able to get 40 moves average really really easy

So essentially a corners last variant? I really doubt that it could be very efficient because you have to work around the solved edges which makes it really inefficient. Also, you permute edges while orienting corners algs will not be very good as the edges would work more naturally in the {U,D, double turns} movegroup but the corners would be more {R,U,D,L,F} (which incidentally is not very good either as most algs are 2 to 4 gen for a reason).