CPEOLL (Fridrich add on method)

[hawkmp4]

CrF for cross+F2L? C+F?

 

[Kenneth]

This line isn't objective and doesn't belong in the wiki, regardless of whether or not it's true: "It is a great mystery why these methods are not popular at all, it is fast, effective and easy but there are very few people that uses columns."

I haven't convinced myself that Cross+F2L is faster than Columns First when both are mastered, but I don't think it's hard to see why Cross+F2L is the more popular speedcubing method. Solving the cross first reduces the number of possible edge, and hence also pair, positions. This makes it easier to make the solution for each case as an automatism (either by directly learning algorithms or as a result of repeating intuitive solves). And automatism (i.e. braindeadness) is great for speedcubing.

Don't bother, I tried pointing this out a while ago.

Mabye I put that in bad wordings but, there are billions of variations for CF or LBL, and on top of that there are loads of akward methods used, columns is obvious compared to many of those and still there are no users.

If you don't like some of my writings in the wiki, then just remove it, I don't own the wiki, nor the articles I put there. It is something that belongs to the community. I'm well aware of that when I write to there, if I had problems with people editing my writings I would not put it there. If it is a huge change that destroys hours of work for me, then please talk to me first (at the wiki talk page), mabye I can explain why I did like I did.

I'd be interested to hear what Roux solvers think of Columns First variations. They're a better judge than Cross+F2Lers like me.

I wouldn't call myself a Roux solver (well, I can do it, I've done a sub20 average) but I've done Columns First enough to be moderately fast at it (around 20). I think it certainly has some potential but the biggest problem is the relatively large number of rotations. S moves can be done but they are awkward and I don't know any way to do fingertricks with them (but hey, I said that about M moves too a while back, and look at what I know now), so in practice whenever you need to do those turns you have to do a y rotation. This doesn't happen just during the L8E step but during the columns step as well, because occasionally an edge you want is sitting in DL or DR, which is not only awkward to look for but (again) awkward to execute. So I think this method has potential, sub20 is definitely possible and sub15 probably is, but it won't be as fast as Roux or Fridrich (with good F2L technique) without some modification.

After practice you learn to save on rotations, Dw is a pretty obvious substitute for U + y. Putting down pairs you can use Rw U R' instead of F R' F' R or y U' L' U L. Solving pieces opposite to the previous instead of an adjacent in L8E if you have a option to do so can save two rotations.

The rotations are basicly only y, you loose on the grip shift but looking ahead is not a problem (exept for pieces that are stuck in D or are sitting on the backside but that is not due to rotations).

 

[rubiknewbie]

The problem is that these occur rarely. My experience is that rarely practiced algorithms become forgotten algorithms. And you will not orient corners just to do this since orient edges gives better options.

 

[Kenneth]

True, but 50% of the 10 cases that are not ELL or PLL have "intuitive" solutions (PLL with setup) so for those you don't need to learn a specific alg. So there is only five left, not so much to remember.

(One of those algs, the one that solves the "T" is useful for classic Pochmann BLD, it orients the edges while doing the T-perm, at least the one I put there, others may change permut for more LL edges)

And besides that, 1:27 is not "rarley"

Compare to H-PLL for example, occures 1:72, and still you know an alg for it =)

 

[qqwref]

After practice you learn to save on rotations, Dw is a pretty obvious substitute for U + y. Putting down pairs you can use Rw U R' instead of F R' F' R or y U' L' U L. Solving pieces opposite to the previous instead of an adjacent in L8E if you have a option to do so can save two rotations.

The rotations are basicly only y

Dw can work, but i don't tend to use it because it doesn't seem to set itself up nicely. Besides a lot of the time when I'd want to use an S move I especially do not want a U or U' before it. The putting down pairs tricks don't help since an experienced F2L solver would know how to insert a paired pair without y rotations already, my concern is with S moves. The third tip is useful. I agree it's only y rotations in the method, same as Fridrich, but there are definitely more and it's better to not have to do them.

Here is a similar method I suggest:
- While you build the four pairs, also place DL and DR. I guess this can be done most effectively by putting the five pieces on left first (can be done in inspection?) and then the five on right.
- Now fix corners, might as well do it now, you can do it in one step and neglect the M slice.
- Last six edges. Orient first, then permute (I personally like placing UL/UR and then doing the M slice, but if you're really fast at ELLs you could place DF/DB first I suppose).
It's definitely more optimized than the normal columns first method

 

[Kenneth]

Depends on what the user is used to I guess because I have trouble solving in sub 30 using Roux but not with columns where I average about 27 atm... But it also depends on the scramble, if I have a easy 1x2x2 to start with then of course Roux is an option.

As you know I tried a few methods and the columns I use now is the fastest for me so far. But of course, being my method it is adapted to what I can do (decently) fast.

---------------
Edit, don't like S moves, so you keep doing y? I just found this : setup R' U' R' U R .. inverse it to solve

Well, won't solve all positions but... one down =)

Honestly I prefer (y) Lw' U L ..

 

[Tnghia]

In my opinion, it will only work if you have already learn oll and pll. It may take more time to look than the old way.

 

[TK 421]

Omg...

 

[irontwig]

In my opinion, it will only work if you have already learn oll and pll. It may take more time to look than the old way.

That's what "Fridrich add on" means.