OK So I was thinking of this today, best speedcubing 2x2 method out there, i just want to know how many algs you would need... ok you have a 2x2 in your hand, on the U layer, you have a 1 white, it can be anywhere on the top half of the cube you know?

Looking down on the top of the cube, you don't have to see the white, it can be on the LBF or R face, but has to be on the U layer, get it?

ok, so if you have 1 white on the U, you have 3 on the D face right? perform x2, then you have 3 on top, 1 on D Right?

ok, so 3 moves R'D'R you have all whites or w/e color you want on 1 layer.

If you have 2 on each layer, D and U, if they are touching each other you need at most a D2 or a U2, and then a L2 or a R2. OK, easy right?

if you have 2 diagonal from each other on top and bottom, put them so that they corrospond, like the FRU is blue, and the DRF is blue, like the LFU and the DLF is green, BLU and DLU is blue, and RUB is green, an R2 or L2 would get that set up. Now if on the U layer you have 2 blue and 2 green connectiong like 2 lines, and then checkerboard on the bottom, like F corners, blue, F U that is, BU is green, DFR is blue, FDL is green, BDR green, BDL green that would be B'D2 B'.

SOOOOOOOOOOOOOOO, step 1, get a set of opposite colors on their layers, not faces! well, faces ok, but all you need is one colors "below the belt" of the 2x2, and the opposite color would be "above the belt" get it?

This i am assuming can be seen in preinspection and executed very quickly

Step 2!!!!! Use 1 Algorithim to permuate and orient the 4 cubies on above the belt cubies, this alg CAN screw up the bottom cubies, therefore they can be shorter, and or using finger shortcuts, (only R and U moves or something)

Step 3, perform x2 then use 1 alg to permuate and orient the cubies remaining unsolved.

Simple RIGHT!?!?!?

I am just wondering, how many algs would this take? and does anyone have them written out somewhere? so review, first step, able to be seen in preinspection like max of 4 moves am i right?

then just 2 algs, simple and fast!!!!

has this already been invented? is it on a webpage? any info comments anything on this?

Sort, orient both layers simultaneously, then permute as in Guimond?
Sounds good; if I weren't trying a different method, I would go calculating the orientation algs...

No basically he is saying you put opposite colors on the correct layer (either upper or lower layer). They can be incorrectly oriented. One example is to have all the white corners as the upper layer, and the yellow corners as the bottom layer. Then you orient AND permute the upper layer, do x2, and orient AND permute the last layer. It's a three step method that would be very quick if it's feasible. Only problem is the number of algs. I don't know how many, but I would guess 50-100, probably more... That's my educated guess for whatever its worth

88 per layer

The last layer can be solved using CLL, that's 42 algs (43 including solved). I guess that the first layer takes 3*43-1=129 algs (you could use CLL for 1/3 of the cases, but they wouldn't be optimal because we don't care about the last layer). I don't think this is a good idea, too much memorization and too little intuition.

so...171 algs in all?

lets see....for the 3x3, 120 in all if i am right, for oll,pll, and f2l, no cll, so it would be somewhat close..

ya, i thought it would be a crap load of algs.

Yes, unless I made a mistake.

For Fridrich, yes.

It's not as bad as ZB but... man that's a large number. It's probably faster to do Guimond or a layer method anyway, if you have good lookahead and recognition, because the recognition would be very difficult to learn with so few stickers and so many possible algs. A lot like ZB, in fact.

Would anyone consider calculating it out to see what the average movecount would be?

Krig-Akkersdijk-Wojtyla method, if I have the names right, is the best, if we don't count optimal. It goes as following:

1. Complete a face.
2. Do the rest.

According to Gunnar, it averages 13 moves, meaning anyone can average under 5-6 sec with it. However, there are 120 algorithms.

That does sound pretty fast, although the recognition could be hazardous, because you're looking at a CLL-like pattern plus one of three bottom-layer permutations...

it woudl be the optimal method for computer solving, computer looks at the cube faces, and then can generate the solve in like under a second...rubot mini?!?!?

I think there's a program out there (maybe on the yahoo speedsolving group) which will always give you an optimal 2x2x2 solution (it knows God's algorithm).

Sorry I gave up reading after the first three paragraphs didn't show any attempt to get to the point.

Ya I'll admit, it was bad.

In under hopefully 5 moves get 1 opposite pair's stickers on their respective layers/faces. white on top, yellow on bottom, doesn't matter orientation or permuation.

step 2. use 1 alg to permuate and orient 1 layer.

step 3. same thing to the last layer.

step 4. look at your stackmat in awe at the low low time.

I think it should never take more than 2 moves.