The New Method / Substep / Concept Idea Thread

mDiPalma said:

i think it's a really good method.

2x2x2
2 F2L pairs.
pair and PLS
permute U-layer corners while inserting a D-layer edge
last 5 edges per columns first.

or

2x2x2
3 F2L pairs
CLL (of some sort)
insert 1 D-layer edge while orienting U-layer edges.
EP5 (really good alg set)

Click to expand...

Where are last five edges algs? I've looked before and can't find them.

ALSO:
not really a full idea, but when I tried and failed to solve a square one intuitively (just got it and didn't know any method) it was not in a square shape but I managed to solve everything but three pieces: two edges and a corner. I determined it would be impossible to solve them with a commutator because of the irregularities in shape making too many turns impossible, but it got me thinking: would a "direct solve" method be competitive? It would be something like this:
Solve cube shape
Get as many pieces as possible solved relative to each other, maybe one or two blocks misplaced and a few pieces totally unsolved (even in the wrong layer)
Solve the rest with algs.

Like I said, this is totally undeveloped, I'm just asking if something like that could be fast.

lerenard said:

Where are last five edges algs? I've looked before and can't find them.

ALSO:
not really a full idea, but when I tried and failed to solve a square one intuitively (just got it and didn't know any method) it was not in a square shape but I managed to solve everything but three pieces: two edges and a corner. I determined it would be impossible to solve them with a commutator because of the irregularities in shape making too many turns impossible, but it got me thinking: would a "direct solve" method be competitive? It would be something like this:
Solve cube shape
Get as many pieces as possible solved relative to each other, maybe one or two blocks misplaced and a few pieces totally unsolved (even in the wrong layer)
Solve the rest with algs.

Like I said, this is totally undeveloped, I'm just asking if something like that could be fast.

Click to expand...

Sounds similar to the roux and screw method.

molarmanful said:

Advantages over CFOP and Roux:
----------
1) MORE WAYS TO START. If you use CFOP or FreeFOP, you can make a Cross - 1E and solve F2L pairs or something like that in any order. If you use Roux, you can build 1 or 2 1x2x3 blocks and finish up the F2L pairs if you haven't already. If you use Petrus, you can build a 2x2x2 and/or a 2x2x3 block, then finish up F2L pairs.

2) AUTOMATIC OLL SKIP (for those who use PLS). Okay, who doesn't love skips? Besides, an OLL skip increases the chance of a PLL skip. MOAR SKIPZ!

3) PREDICTABLE. It is super easy to predict your last layer by the time you start EFL+LLEO. Because EFL+LLEO uses only slice moves, corner orientation is not affected, meaning you can predict either your OLL/COLL case if you don't use PLS, or whether you have headlights PLL, no headlights PLL, or an EPLL case if you use PLS.

Click to expand...

1) If you started with Roux blocks, there's no reason to switch to Pang because Roux is perfectly good by itself. I can see this as a variant of CFOP or Petrus though. But for CFOP, that cross edge slot will be a huge problem if you cant take an edge out of there quickly, in case you need it for F2L. Plus it hinders lookahead, plus it's VERY difficult to take advantage of the open space if you dont have either edges oriented or a free M slice. And in either of those 2 cases it'd just be ZZ or Roux.

2) You don't skip OLL. You have PLS for OLL and the EO step for OLL. That's 2 steps for OLL and therefore nothing was skipped. And PLL skip chance doesn't change.

3) Okay if we continue to talk about this method, just assume that everyone knows PLS. If anyone's gonna use this method properly there's no reason to not know it. I agree that PLL will be much more predictable as an advanced user of this method can quickly see which stickers are gonna join together, which is probably enough to get a good sense of PLL after glancing at CP.

And going back to #1, more ways to start is not an advantage to this method over CFOP because you can just FreeFOP it anyway. The main difference between Pang and FreeFOP in the F2L stage is you ignore that one cross edge. It doesn't really make a difference, so in that respect, FreeFOP is just as good as Pang.

lerenard said:

Where are last five edges algs? I've looked before and can't find them.

Click to expand...

yeah, they are Athefre's algs. i think his site is down. but here are the algs

http://s140.photobucket.com/user/Athefre/media/EP5Final.png.html

molarmanful said:

You seem to be the biggest critic of my method; you've been blasting it from the day I proposed it to this site. I'd like to ask how you would improve this method to your liking. Or do you think that this method is hopeless and that I should just abandon my work? I've heard enough about your misgivings; I have similar doubts. Now it's time I hear other things.

Click to expand...

Yeah, going back through the thread, I seem to come off as antagonistic. Sorry about that, I've been having a really bad week.

Honestly, I'd probably turn it into Roux or one of the methods MdiPalma suggested (probably the second one). Basically, both of the methods he recommends address the main concern I have with the method, that being, the sheer number of hidden pieces at any one time, which makes lookahead very difficult. Smiles explained that part better than I can.

If you can come up with a way to deal with those hidden pieces, you've made a fairly decent method. I don't think a Columns First method is ever going to be better than any of the big 4, but I'd love to be proven wrong.

supercavitation said:

Yeah, going back through the thread, I seem to come off as antagonistic. Sorry about that, I've been having a really bad week.

Honestly, I'd probably turn it into Roux or one of the methods MdiPalma suggested (probably the second one). Basically, both of the methods he recommends address the main concern I have with the method, that being, the sheer number of hidden pieces at any one time, which makes lookahead very difficult. Smiles explained that part better than I can.

If you can come up with a way to deal with those hidden pieces, you've made a fairly decent method. I don't think a Columns First method is ever going to be better than any of the big 4, but I'd love to be proven wrong.

Click to expand...

Oh, it's fine. Thanks for the feedback, though.

Roux F2L?

Is roux f2l (I don't now a good name) fast?
I mean that first do the roux left and right side and then permute the last two or one corner with M and U.

Edit: Sorry I mean the last two or one edges
So that you have yor f2l done and than oll and pll or something like that.

This has been suggested many times before and isn't a good method. Either use CFOP, or use Roux. Don't try combining them like this.

Ok thanks, I'm cfop user and I will not learn roux so I think a mix of them. But ok

This makes me think of that cyoubx video about rotationless F2L in CFOP being a silly way of doing Roux. Anyone able to find the link?

sneaklyfox said:

This makes me think of that cyoubx video about rotationless F2L in CFOP being a silly way of doing Roux. Anyone able to find the link?

Click to expand...

this?

https://www.youtube.com/watch?v=Js1LZscj18A

Thomasthecuber said:

Is roux f2l (I don't now a good name) fast?
I mean that first do the roux left and right side and then permute the last two or one corner with M and U.

Edit: Sorry I mean the last two or one edges
So that you have yor f2l done and than oll and pll or something like that.

Click to expand...

i don't know if this is much worse if at all to cfop f2l. i think it is less moves and no rotations, but sometimes hard to lookahead and M moves. so maybe it's just as good. of course you can't do last slot and edge control stuff as easily.

GuRoux said:

i don't know if this is much worse if at all to cfop f2l. i think it is less moves and no rotations, but sometimes hard to lookahead and M moves. so maybe it's just as good. of course you can't do last slot and edge control stuff as easily.

Click to expand...

harder to look ahead. one of CFOP's big advantages is how easy it is to look head.

but 2 D edges + OLL + PLL i assume is slower than CLL + LSE.

the only advantage i can see out of doing RouxF2L is if you do EO during the last 2 edges (and corner recog) and finish with ZBLL.

GuRoux said:

i don't know if this is much worse if at all to cfop f2l. i think it is less moves and no rotations, but sometimes hard to lookahead and M moves. so maybe it's just as good. of course you can't do last slot and edge control stuff as easily.

Click to expand...

Would it be more moves if you have to then add in the two cross edges afterwards?

TDM said:

Would it be fewer moves if you have to then add in the two cross edges afterwards?

Click to expand...

i thinks so. those edges take about 6-7 moves? roux f2b is 25 for me in a speedsolve.

GuRoux said:

i thinks so. those edges take about 6-7 moves? roux f2b is 25 for me in a speedsolve.

Click to expand...

Oops, I meant to say more, not fewer

mDiPalma said:

yeah, they are Athefre's algs. i think his site is down. but here are the algs

http://s140.photobucket.com/user/Athefre/media/EP5Final.png.html

Click to expand...

Thanks! Most of them are intuitive, and the rest are easy to understand and remember.

Here is my idea: after F2L is done, you do an algorithm that permutes corners of the last layer and flip all edges. After that you use 2gen algorithms to complete the rest. So for example:

instead of doing OLL and PLL, do F U R U' R' F' then you will have this 2gen case:

This can be useful for OH...

That's CPEOLL/2GLL. You can find algs for the CPEOLL step here.

oops thank you i didnt know that