Improvements on the 2-look OLL

[sjoerdtheone]

Hi guys, I just thought of this but I don't know whether it's widely known and used or whether I'm just posting this, even though everyone already discovered this (without posting?).

Okay so it's about the 2-look OLL which basicly means that you learn the algorithms for OLL when you come across a "cross" or "plus" on the u-face.
Now, normally when I solve a cube, and I've finished F2L and I see any non-cross figure on top, I simply use FRUR'U'F' to get a cross (sometimes needed twice).
But the problem I found using this method is that it's hard to predict what your next algorithm needs to be, because the corners don't stay in the same orientation after FRUR'U'F'.
I found a way around this: when you get to a non-cross OLL, look at the orientation of the corners, and remember which, and from which side you will need to apply the algorithm to fix ONLY the corners.
After that orient only the edges of your top-face by using either:
- RUR'U'M'URU'r' or
- M'UM U2 M'UM or
- MURUR'U'M2URU'r' (use this from any side if there are NO edges correctly oriented)

You select the algorithm based on what kind of pattern you see. If there is a bar in the u-face colour from UL to UR then you use the first.
If there is a L shape of any kind, put one correctly oriented edge of the L shape in the UF position, and the other in the UR position. You do this by simply rotating the cube, or Adjusting the U-Face (AUF).

When you have executed one of these algorithms, you are now left with only the corners to be oriented, which you can do with the algorithm that you thought of in the first step.

Hope this helped and any feedback is appreciated

 

[Boulet]

And what is the improvements? It is still a 2-look OLL, no?

 

[brunson]

Now, normally when I solve a cube, and I've finished F2L and I see any non-cross figure on top, I simply use FRUR'U'F' to get a cross (sometimes needed twice).

I would call that a three look OLL.

 

[Crzyazn]

...I find this a little pointless.

If you have trouble recognizing OLLs now after flipping edges, who knows what will happen when you try to learn ALL the OLLs.

I also find MU algorithms a little cumbersome, 2-4 sec for flipping edges with MU as opposed to 1-1.5sec for the FRU algs (but mostly this is an issue of practice?)

 

[sjoerdtheone]

Now, normally when I solve a cube, and I've finished F2L and I see any non-cross figure on top, I simply use FRUR'U'F' to get a cross (sometimes needed twice).

I would call that a three look OLL.

Well, it's two look because from the first look you can see whether you need to perform the alg. twice or only once.

The improvement is that it is technically a two look OLL but it's done with one look, and two algorithms.

Edit: and if you don't like the algorithm, you can use either of these instead, after doing a y2 or U2:
-> r U R' U' r' R U R U' R'
-> F R U R' U' F' U2 F R U R' U' F'
but the the one mentioned first is the one I use..
(I got these straight off this website, Collection of OLL algorithms)

 

[DAE_JA_VOO]

Now, normally when I solve a cube, and I've finished F2L and I see any non-cross figure on top, I simply use FRUR'U'F' to get a cross (sometimes needed twice).

I would call that a three look OLL.

Well, it's two look because from the first look you can see whether you need to perform the alg. twice or only once.

The improvement is that it is technically a two look OLL but it's done with one look, and two algorithms.

No, because after that, you have the cross and need to do ANOTHER alg to orient the corners.

 

[Crzyazn]

a "look" is a step, no?

i think you've taken the definition a bit to literally.

 

[sjoerdtheone]

Okay, if a look is a step, then it is two looks. If a look is a look (which I thought) it is one look. But this method does reduce time between these looks.

 

[Boulet]

For me, the firt look is almost recognized instantaneously (3 cases: line, elbow, just center). Then, there are 7 cases for the second look, wich are fast to recognized (1 sec. or less). Your "one look" seem to required at least the same time then the traditionnals 2-look because there are more things to look for. Are you really faster with this method?

Micael

 

[sjoerdtheone]

Well, since I have just thought of it, I haven't timed an average with it yet but I think I will soon..
I don't think it makes a big difference (at least not seconds) but I thought it was a nice idea to take a closer *look* at.

 

[AvGalen]

a look is a look, not a step.

but basically you are still using 2 look OLL, only you do both looks (edges, corners) at the same time. You might save 0.01 because of that, but you will lose more because the algs are longer/slower. So it is not a good idea for speed, just a nice alternative

 

[Lofty]

I would define a "look" as when you can close your eyes and finish the step. If as soon as you finish OLL you can orient the edges with your eyes closed even tho you have a dot and are doing two algs it will be one look.
Anyway If you have trouble recognizing corner orientation you really should practice that...

 

[brunson]

If I can perform my OLL and anticipate the PLL so I don't have to look, is that a 1 look LL?

 

[Dene]

I dear say, learning all this, when it would be much better in the long term just to learn the OLL's...

 

[Jai]

Instead of doing a whole algo to handle the edges, then doing an algo to handle the corners, do a set-up move for the edges (a slice move), solve the corners, then undo the set-up. http://speedsolving.com/forum/showthread.php?t=2664 This is just a simple version of Kenneth's XCLL.
I honestly don't know how it works, though, because I don't do BLD and I don't know much cube theory, so if you're going to ask, Kenneth is your man for that.

 

[sjoerdtheone]

Problem with your method, Jai, is that there are still equally many algorithms as with normal OLL... (Although it might be a little more intuitive)
"my" method has about 10 algorithms I think ?

 

[AvGalen]

Problem with your method, Jai, is that there are still equally many algorithms as with normal OLL... (Although it might be a little more intuitive)
"my" method has about 10 algorithms I think ?

Your method has the same number of algs as regulare 2 look OLL (3+7), only the first 3 algs are longer and slower. The idea of better look-ahead for the second OLL is nice, but the cons outhweigh the pros as I mentioned earlier.

 

[sjoerdtheone]

yeah I think you're right on that one

 

[Lucas Garron]

I published compound OLL 2 years ago.

I don't think it's a terribly unique idea, but I hadn't seen anybody else mentioned it, and I would absolutely have used it if I were going for full Fridrich.
But it's good to know that these ideas are viable enough to recur...

Anyhow, the best time to use this is for learning OLL, I think:
A) 2-look, 2-alg OLL
A) 1-look, 2-alg OLL (compOLL)
B) 1-look, 1-alg OLL

First, you try to learn enough algs to do A for every OLL case. Once you're done with PLL and improving at F2L, you should learn the compOLL algs and execute all unknown cases as B. Any cases you learn should automatically become C. Then, as you gradually learn, all your algs become C. You'll be one-look every time, but your average OLL exec time keeps decreasing. And it'll be nice when you start replacing long-ish compound cases with short ones.

 

[sam]

Hi,
I do kind of what lucas does with his compOLL but i use very easy 7/8 move algs (the T perm ones) or other easy ones to turn each case into a sune/anti-sune. if anyone has specific cases they would like to know or just more about this method PM me.

I average 17 with it and full PLL as well as F2L. and i'm color neutral. so yeah.

-Sam