Edges, then corners method - sloww?

[sweetandy]

I'm learning the Fridrich method (albeit slowly). Presently, however, I use a system that I found on the Rubik's website, kind of. At least, I use their few, few algorithms.

I solve the white cross, then the middle layer's edges, then the yellow layer's edges. Then, I slowly rotate the corners into place from the yellow side. Of course, if the white side of a corner cube is on the yellow face, I can't rotate it in and it requires yet another use of the algorithm. So at the very least, I have to do this eight-step algorithm four times, likely much more, just to get the first two layers (with the edges of the last layer already solved). Found on the Rubik's website.

R'ULU'RUL'U

That flips the three corners on the top layer of the cube in the way you're holding it. So when I come to the last layer, I find a cube in the right position, but not necessarily the right rotation, and do the three-rotate thingy until they're in the right place (not nec. right rotation). Then I use the corner-flipper-down rotation to flip any corners the right way. If there are three, four, or two diagonally that are not in the right rotation, then I have to do it twice. Again, it's slow, long, and found on the Rubik's website.

LUL'ULU2L'R'U'RU'R'U2R

Those are really my only two algorithms for my current method, except the few tricks I have for solving the edge pieces.

All in all, my average time with this method is about 1:17. My record is 59s, and I think that's darned near a cap for this slow method.

Does anybody have any advice for me, or is "Fridrich" the answer to this? I mean, I can get the first two layers in about 40-50 seconds right now, and only know four or so algorithms, but then there's permuting and orienting the last layer, where I'm pretty much lost. IS Fridrich the answer?

 

[tenderchkn]

The logic of that method escapes me. If you place an edge before a corner in a cross methdo, you're building the middle layer before the first layer is completed. That means to place the corner, you must take the edge out, pair it with the corner and put both in. Hence, the inefficiency. Even if this is done using algorithms, that is in essense what is happening.

Full Fridich, or even beginner Fridich eliminates the need for any algorithms for the first two layers. Either the corner/edge is paired and placed simutaneously, or the corner is placed, paired with the edge and put back in. Then, the last layer can be solved with a few quick algorithms. Check out one of the many beginners tutorials and you'll see.

Couple notes: R'ULU'RUL'U (Nikolas) also PERMUTES the same three corners it rotates. RUR'URU2R' (Sune) and its mirrors/inverses only rotates three corners, and can be executed much faster. Experiment with it.

 

[Johannes91]

Edges first can be very fast IMHO, but doing corners that way is absolutely ridiculous. Using commutators is much more effective, check out Ryan Heise's excellent guide here. It's for solving last 3 corners, but you can use it to solve 2-3 corners at a time.

Originally posted by tenderchkn@Feb 24 2007, 07:43 AM
Couple notes: R'ULU'RUL'U (Nikolas) also PERMUTES the same three corners it rotates. RUR'URU2R' (Sune) and its mirrors/inverses only rotates three corners, and can be executed much faster.

It's Niklas. Sune also disturbs edges...

 

[Colton7909]

I was playing with this method and i found it to be about as fast as the beginners method if you know a few shortcuts, and if you get lucky it's faster. I use something i like to call substitution, but i have no idea if that's the real name for what i'm doing. I'll try to make a video about how i do the edges first method and some shortcuts. Even though it can get faster it's no where near as fast as Fridrich. With the edges first algorithm i average about 45-55 seconds, and with Fridrich i usually get around 30-40 seconds. So if speed is your goal I recomend Fridrich.

 

[waffle=ijm]

how about no. and you just bumped a 2 year old thread.

Keep in mind the date.

 

[cubeninjaIV]

Fail.

 

[Me]

but have he had made a new thread about it then you would ridicule him for not searching for this topic. I commend you Colton7909 for using the search feature and adding something to a two year old thread.

 

[Edward]

but have he had made a new thread about it then you would ridicule him for not searching for this topic. I commend you Colton7909 for using the search feature and adding something to a two year old thread.

But its not like hes asking a question. I think his post deserved a thread.

 

[waffle=ijm]

New thread would have been better because of this sentence

So if speed is your goal I recomend Fridrich.

so it's a response to the original poster who hasn't been on since
02-25-2007 01:05 PM

hence new thread would have been nicer so we could discuss.

 

[rubiknewbie]

It's better not to teach a 2-year old poster as you run the risk of teaching to someone who may be a sub-15 sec cuber now .

 

[Dirk BerGuRK]

...hence new thread would have been nicer so we could discuss.

You can't discuss it because it's an old thread?

 

[Ranzha]

BUT WE DIGRESS!
Remember what this thread is about!
I find the best way for this approach is to solve cross, the middle layer (using R U R's and L' U Ls) and then using one of three algs to solve the first layer corners. Then 4LLL (learning full PLL and COLL now).
I find this method a good step from Beginner's Method in that when I use this, my F2L is about 20-25 seconds. Then, with a 15-18 LL, I average about 40.
When executed properly, you know where the corners go after performing an algorithm and you can both look ahead and easily transition to the next case.
This way, the F2L looks flawless, but its move count is bad, is all.

--Ranzha V E

 

[nigtv]

Something that I found extremely useful when learning the beginner method was learning how to mirror and inverse algorithms. Even if I only remembered sune or niklas or whatever, could still get many useful out of that. Lets you not memorize and keep track of quite as many algs when youre beginning...