The New Method / Substep / Concept Idea Thread

[sub20oneday]

not idea but a question:

why no eoll/cpll wouldnt that be fast?

 

[JTWong71]

not idea but a question:

why no eoll/cpll wouldnt that be fast?

The problem with that is that really nobody wants to force an E-Perm, which is 16 moves. So then you may think what if we solve 1 Corner with that, then you are left with A-Perms, but that adds a bunch of algorithms. Lastly, recognition is a little harder than COLL in most people's opinions due to EO being solved, you cannot just look at the top face.

This is interpreting EOLL as Edges of Last Layer, but I would normally think it is Edge Orientation Last Layer, but seeing CPLL making it Edges of Last Layer.

 

[wir3sandfir3s]

Recognise one of 40 thousand cases. Simple

Have you looked at Briggs2? It's the only kinda viable CP first method I've seen and even it isn't so great.

CP taking 40k Algs?? Uh, then maybe we can find a way to do it intuitively...
Yes, I am a fan of Briggs2. The difference with mine is that it is solved first step so you don't have to slow down, and if you are only turning LU or RU or maybe even FU at a time TPS will probably be pretty high...

 

[JTWong71]

CP taking 40k Algs?? Uh, then maybe we can find a way to do it intuitively...
Yes, I am a fan of Briggs2. The difference with mine is that it is solved first step so you don't have to slow down, and if you are only turning LU or RU or maybe even FU at a time TPS will probably be pretty high...

If we are to try this, instead of E Moves for the L-Side Edges, you can just use u2 (Wide).

 

[wir3sandfir3s]

If we are to try this, instead of E Moves for the L-Side Edges, you can just use u2 (Wide).

You can use u move in place of E moves if you want, you will just have to AUF if you are going to switch sides you are block building.
Btw, I imagine this will take even more time than F2L to get used to, let alone get good at...

 

[JTWong71]

You can use u move in place of E moves if you want, you will just have to AUF if you are going to switch sides you are block building.
Btw, I imagine this will take even more time than F2L to get used to, let alone get good at...

Yeah, I think that normal ZZ would be better.
A Roux block on LD will likely be what the First Step is, along with Corner Permutation,
due to the restrictions and what pieces won't be shifted.

 

[wir3sandfir3s]

Yeah, I think that normal ZZ would be better.
A Roux block on LD will likely be what the First Step is, along with Corner Permutation,
due to the restrictions and what pieces won't be shifted.

Who said this was based on ZZ? It's just one of those weird methods I propose occasionally.
anyway, I thnk that the simple CP first step can be applied to pretty much every method, besides Roux. Also, another cool thing is that you can ignore centers because you will be moving the middle layers anyway.

 

[JTWong71]

Who said this was based on ZZ? It's just one of those weird methods I propose occasionally.
anyway, I thnk that the simple CP first step can be applied to pretty much every method, besides Roux. Also, another cool thing is that you can ignore centers because you will be moving the middle layers anyway.

I know you aren't, but it does also have EO in the beginning.

Part of the thing is that CFOPers say that EO is hard to detect in the beginning, but of course after practice it becomes a lot easier. Having to detect another somewhat hard step will make it difficult to inspect in less than 15 seconds.

 

[crafto22]

Hey folks, I haven't posted in this thread in a while, but I have a new idea so yeah here it is. So basically I do a lot of 2-gen solving just for fun and I figured out how to do some 2-gen winter variation. I don't know 2GLL so I used this winter variation thing to skip 2GOCLL and go straight to EPLL. anyway so this got me thinking, what if we did some sorta thing where we solved F2L-1 with EO done and then did some alg to pair up the last CE pair and solve CP. Then we could just do 2GWV and go straight to EPLL. I don't know if this could possibly be a good idea but just in case it does result in a lower total move count it may be worth trying.

 

[Teoidus]

Is this possible?

2x2x3 (~13 STM)

Attach 1 CE pair (4 options) forming F2L-1, without one cross edge (~3 STM)

2gen reduction + DF in 1 step (possible...? ~10 STM?)

2gen LSLL (6 STM LS + 12 STM 2GLL)

=> ~46 STM

 

[Dane man]

Not sure if this has been suggested (probably has), but here is a silly idea I just had in passing.

Do cross, then the first two slots of F2L. Insert the third slot and orient the edges simultaneously (maybe intuitive in some way?). Insert the last slot and orient corners. PLL

If the third slot and fourth slots don't take too many extra moves, then the extra moves might be lower than the average OLL. If so, this could have potential.

 

[wir3sandfir3s]

Not sure if this has been suggested (probably has), but here is a silly idea I just had in passing.

Do cross, then the first two slots of F2L. Insert the third slot and orient the edges simultaneously (maybe intuitive in some way?). Insert the last slot and orient corners. PLL

If the third slot and fourth slots don't take too many extra moves, then the extra moves might be lower than the average OLL. If so, this could have potential.

L2S with CFOP.

 

[James Hake]

Not sure if this has been suggested (probably has), but here is a silly idea I just had in passing.

Do cross, then the first two slots of F2L. Insert the third slot and orient the edges simultaneously (maybe intuitive in some way?). Insert the last slot and orient corners. PLL

If the third slot and fourth slots don't take too many extra moves, then the extra moves might be lower than the average OLL. If so, this could have potential.

I have also had some ideas similar to this and might use a similar concept in petrus. Use something similar to VHLS to orient all the edges and while inserting the last pair. Then you can use winter variation for the last slot and get an OLL skip. Or you can just insert normally and use COLL, ZBLL, or just have one of 7 easy OLLs

 

[Sion]

I know it isn't talked about, but what about a corners first method for clock? It's my second favorite event being 3x3, and I was dabbling with this idea:

It's still an idea, but maybe it will work:

Solve corners while Simultaneously solving the center clocks.

Orient Edges (just get all the edges and center clock to face the same direction with one alg)

Permute Edges (get all the oriented edges to Face Noon with one alg)

This would take a rather high amount of algorithms, but I haven't seen any real speed clock methods out there.

 

[Teoidus]

2-gen redux

1. FB (9/9) -> <Rr, U>
2. 2x2 in BDR (9/18)
3. EO+CP with algorithm (9/27) -> <R,U> + (r U2 r')
4. FD edge + LS (9/36) -> <R,U>
5. 2GLL (14/50)

= 50 STM

The block can also be made by expanding a 2x2x2 into a 2x2x3 + pair, or by placing 3 heise-like squares.

Almost all algorithms are <Rr, U, F> and consist of combinations of very familiar triggers (R U R' U', R' F R F', F' U F). Average movecount is 7.4 STM with stdev 1.67. Many algorithms are absurdly short. Examples:
R' F R F'
F R' F' R
F' U F
r' F R F' M'
M' U M

They are available (though not yet polished) here. I'll be posting something on recognition later on (the only worry I have with this method, though I think it will not be a problem with enough practice).

Because of the magic of rotational symmetry, step 3 has 55 algorithms instead of 2^6/2 EO * 6 CP = 192. This brings total algorithm count to 55 reduction + 84 2GLL = 139; even less if you decide to just use Sune/ASune -> UPerm (which would still meet the 2GLL 14 STM movecount estimate).

3 example solves with alg.cubing.net links here (49 STM) and here (49 STM) and here (41 STM).

 

[wir3sandfir3s]

2-gen redux

1. FB (9/9) -> <Rr, U>
2. 2x2 in BDR (9/18)
3. EO+CP with algorithm (9/27) -> <R,U> + (r U2 r')
4. FD edge + LS (9/36) -> <R,U>
5. 2GLL (14/50)

= 50 STM

The block can also be made by expanding a 2x2x2 into a 2x2x3 + pair, or by placing 3 heise-like squares.

Almost all algorithms are <Rr, U, F> and consist of combinations of very familiar triggers (R U R' U', R' F R F', F' U F). Average movecount is 7.4 STM with stdev 1.67. Many algorithms are absurdly short. Examples:
R' F R F'
F R' F' R
F' U F
r' F R F' M'
M' U M

They are available (though not yet polished) here. I'll be posting something on recognition later on (the only worry I have with this method, though I think it will not be a problem with enough practice).

Because of the magic of rotational symmetry, step 3 has 55 algorithms instead of 2^6/2 EO * 6 CP = 192. This brings total algorithm count to 55 reduction + 84 2GLL = 139; even less if you decide to just use Sune/ASune -> UPerm (which would still meet the 2GLL 14 STM movecount estimate).

3 example solves with alg.cubing.net links here (49 STM) and here (49 STM) and here (41 STM).

Cool, very cool... Except CP
Maybe combine this with my idea and solve CP in beginning? Just an idea.

 

[JTWong71]

Something Shiv3r and I thought of...

After a 2x2x3 Block + EO are solved, you then solve the rest of F2L, except the the FR Edge.
Then you use an algorithm to solve Corner Permutation and that F2L Edge simultaneously.

The algorithms are pretty good, averaging 8.167 Moves with only 6 algorithms, a couple are just inverses of each other and are the standard EO preserving algorithms.

This is something that can be easily learned, and used for when the F2L Case comes up as you just need to know standard CxLL Recognition, as the FDR Corner is already solved.

Algorithms

 

[Sion]

Something Shiv3r and I thought of...

After a 2x2x3 Block + EO are solved, you then solve the rest of F2L, except the the FR Edge.
Then you use an algorithm to solve Corner Permutation and that F2L Edge simultaneously.

The algorithms are pretty good, averaging 8.167 Moves with only 6 algorithms, a couple are just inverses of each other and are the standard EO preserving algorithms.

This is something that can be easily learned, and used for when the F2L Case comes up as you just need to know standard CxLL Recognition, as the FDR Corner is already solved.

Algorithms

Can you do something with this so it is compatible with PCMS? It can help recognition of cases eons faster than a normal CxLL case.

 

[wir3sandfir3s]

Something Shiv3r and I thought of...

After a 2x2x3 Block + EO are solved, you then solve the rest of F2L, except the the FR Edge.
Then you use an algorithm to solve Corner Permutation and that F2L Edge simultaneously.

The algorithms are pretty good, averaging 8.167 Moves with only 6 algorithms, a couple are just inverses of each other and are the standard EO preserving algorithms.

This is something that can be easily learned, and used for when the F2L Case comes up as you just need to know standard CxLL Recognition, as the FDR Corner is already solved.

Algorithms

Are you seriously too lazy to take the extra 3-6 moves to pair it?

 

[JTWong71]

Are you seriously too lazy to take the extra 3-6 moves to pair it?

It is just for when you are solving and the F2L case just comes up.

Also, the Average movecount is around 8 when the standard algorithm is 9 Moves long.