To turn, rotate, or learn a new algorithm. That is the question.

[bsarlo]

I have chosen and memorized a set of algorithms for my OLL and PLL steps. Most often during a solve, the last layer is not oriented and I either need to do an U/U'/U2 or I will rotate the cube y/y'/y2. In your expert opinion, which is generally the fastest, most efficient way? This may be two questions and may also depend on the specific cases. But generally,which do you do? Do you perform a U and then the algorithm with an extra U' at the end? Do you rotate the cube y and perform the algorithm, having the cube solved at the end? Would it be best to learn another algorithm for that orientation? OR...is the time difference so minute that it is just a matter of personal preference?

 

[Kirjava]

I prefer to not pick any specific thing and just do whatever seems natural at the time - I appear to do both.

I tend to do AUF on 3x3x3 (CFOP) and rotations on 4x4x4 though.

It's like how on reddit today there was a thread asking in what order/how do you count on your fingers. I don't have a set way because it's not something I've thought about that much - and it's probably not that important.

Some cases have good algs for other angles, some cases don't. It's worth learning the good ones, for sure.

Although you probably want someone who uses CFOP as their main method to weight in on this too.

 

[Cubenovice]

I prefer to do cube rotation and or / U so the PLL will result in a solved cube.
This enables you to go full speed during the PLL without having to add an additional look

 

[bsarlo]

Sorry, should have clarified that I was referring to 3x3x3, but I guess this does apply to all cubic puzzles. I forget about the bigger puzzles (and smaller ) because I have not been cubing long. I own the different sizes, but spend all my free time on my 3x3x3s. I really want to get better and don't want to waste my practice time on other puzzles. Plus, whenever I spend time on my 4x4x4, I slow on my 3x3x3 because I get my Algs mixed up or it takes me longer to remember the case. I just need more time to get the muscle memory and Algs ingrained.

The Algs I tend to struggle with are the A Perms for the PLL step. I usually do an U turn because it takes me longer to get the grip back after I rotate.

 

[5BLD]

I prefer to AUF. Especially in Roux. If I rotated I could be messing up the S slice!
In CFOP and Petrus I AUF because I find it fasta.

 

[Kirjava]

Especially in Roux.

This issue really doesn't apply to Roux at all.

 

[5BLD]

It could, if say you were using COLL. But you're right. It doesn't really...
Well as I said I AUF for CFOP.

 

[Kirjava]

Why would COLL make a difference?

 

[5BLD]

Doesn't mess up blocks so you can do a y after if you rotated a quarter turn.

 

[Kirjava]

There's an important difference between CLL for Roux and PLL for CFOP - CLL doesn't require AUF after it's performed.

Any rotations you are doing before or after CLL are just a waste of time.

 

[5BLD]

Yup, you're absolutely right... therefore, as you said, it's much better to AUF.

 

[Kirjava]

No. In Roux AUFing is correct - not better.

In CFOP it is unclear which is better.

 

[collinbxyz]

I usually tend to do some kind of U turn, especially if the alg that I use for the case that comes up ends with a U/U'/U2 move, then it's the same amount of moves, just a different last one, if that makes sense.

 

[EricReese]

I don't rotate, its a waste of time. I just do U until I recognize the PLL, then perform, then AUF. But it is only like that with some PLLs. With those that require rotations, I auf first. Such as A perm. Its awkward to rotate back then AUF. However with N perms they are easy enough where I don't need to AUF even though there is a rotation.

Sort of situational but I see no reason to rotate. I find it faster to auf after PLL...

 

[bsarlo]

Thanks to the few of you that voted. My goal was to get a sense of whether, when faced with the choice, people rotated or "AUF". Which, by the way, I had to look in the Wiki for what that meant. I'm definitely a newb - Avg last 10 = 51.57. I'll get there!

 

[uberCuber]

For me, rotation vs. AUF depends on the case. There are some cases where performing an AUF is just inconvenient after the alg, whether because of an x rotation or because of the fingertrick with which I finish the alg (i.e. A, E, F, V), so I rotate for those cases. For most cases though, I AUF before and after the alg.

 

[Sa967St]

For PLLs it really depends on the case. For deciding whether to do U/U'/U2 or y/y'/y2, it depends on which is more efficient based on the current hand position/grip and if it leaves an AUF at the end of an alg, whether it's an awkward U/U'/U2 or not. For learning new algs, it depends on the how fast the alg can be executed; there's no point in learning a slow alg just to know PLLs from lots of different angles. For example, if a lefty T perm is slower than y2+regular T perm, then there's no point in doing lefty T perms in speedsolves.

For OLLs, I tend to mix it up when it comes to U vs y. New OLL algs are generally easy to learn, especially when mirroring ones you already know or just performing them from different angles (e.g. R U R' U R U2 R' or y2 L U L' U L U2 L'). When it comes to learning completely new OLL algs, to me it's only worth it if it affects CP differently. I generally wouldn't learn two completely different OLL algs for the same OLLCP, even if they are from different angles.

 

[bsarlo]

...For learning new algs, it depends on the how fast the alg can be executed; there's no point in learning a slow alg just to know PLLs from lots of different angles. For example, if a lefty T perm is slower than y2+regular T perm, then there's no point in doing lefty T perms in speedsolves...

I understand where you are coming from and I agree. Some algorithms, like some moves, are just easier/faster. For me, and most right handers I'm assuming, my U is almost always going to be faster than my B.

Just want to throw this out there, kinda devil's advocate:

In my case, most algorithms are pretty slow since I am still learning. I feel like I should learn the lefty ones as well as the righty ones if it is going to save me an extra two turns or rotations. ?newb? Especially if the cube is already oriented after F2L. Also, considering how much muscle memory is a factor, I would think I should be able to train my left hand to work. So far so good, I think. It is just taking forever to memorize them and I get them mixed up sometimes. Granted, it may not be as fast as my right right away, but two cube rotations seem really slow to me. Plus, what makes a "slow" alg? If I practice that slow algorithm over and over, won't it eventually become a fast algorithm? That's assuming, of course, that the algorithm is efficient in number of moves. I fully understand that I could never do a 10 move algorithm faster than an 8 move algorithm. I just think that if I train my hands to do the moves over and over again, they will not be "slow" moves.

??

What do you think? Please help me learn here. I know there are a ton of cumulative cubing years in this forum and I would really like some real experience. Maybe someone has actually tried to memorize different algorithms for the same OLL/PLL and can weigh in?

Thanks!

 

[Sa967St]

Just want to throw this out there, kinda devil's advocate:
-I feel like I should learn the lefty ones as well as the righty ones if it is going to save me an extra two turns or rotations. ?newb? Especially if the cube is already oriented after F2L. Also, considering how much muscle memory is a factor, I would think I should be able to train my left hand to work.

The lefty T perm vs. y2 + righty T perm wasn't a great example I guess. Some people can do it really fast, but I can't so I would never use it in a speedsolve.

Granted, it may not be as fast as my right right away, but two cube rotations seem really slow to me. Plus, what makes a "slow" alg? If I practice that slow algorithm over and over, won't it eventually become a fast algorithm? That's assuming, of course, that the algorithm is efficient in number of moves. I fully understand that I could never do a 10 move algorithm faster than an 8 move algorithm. I just think that if I train my hands to do the moves over and over again, they will not be "slow" moves.

If you really want practice to get fast at an awkward algorithm, yeah sure go for it.

 

[Kirjava]

Plus, what makes a "slow" alg? If I practice that slow algorithm over and over, won't it eventually become a fast algorithm? That's assuming, of course, that the algorithm is efficient in number of moves. I fully understand that I could never do a 10 move algorithm faster than an 8 move algorithm.

So you think Eb'SDF'B'ES will always be faster than URU'R'URU'R'UR in the end because it's less moves?