Crazy, Weird, Bizarre, and Unknown Systems

[dChan]

Originally posted by Tomarse@Apr 19 2007, 07:53 AM
I think dan harris and Lars vanandndnererh what ever his name is are making one together, but no sign of it yet =] cubestation.co.uk

lol, it has already been created- t is the VH system. It is a stepping stone from Fridrich to ZB.

Someone throw out a weird idea next and let's see if we can get it working! The best idea can possibly become a forum project for us to work on!

 

[AvGalen]

Throw out a weird idea?

Make 1 face (not complete layer) at the bottom
Transfer all top-layer edges from the middle layer to the top layer
OLL Top (+Parity fix)
PLL Top (+Parity fix)
PLL Bottom
Fix all middle layer edges at once

Weird enough?

 

[watermelon]

Why not just solve the first layer correctly as the first step and save yourself a PLL later on!

 

[AvGalen]

Because it is possible to do the first face in your head during pre-inspection. A whole layer isn't.

Just try to make 1 face using fewest moves and to make one layer using fewest moves.

 

[dChan]

Originally posted by AvGalen@Apr 20 2007, 02:00 AM
Throw out a weird idea?

Make 1 face (not complete layer) at the bottom
Transfer all top-layer edges from the middle layer to the top layer
OLL Top (+Parity fix)
PLL Top (+Parity fix)
PLL Bottom
Fix all middle layer edges at once

Weird enough?

Hey Av, that idea is dangerously close to what I'm doing right now.. I'm watching you... h43r:

 

[AvGalen]

I was trying to think of something weird, but possible. This method has 6 steps (Full Fridrich has 7) but will use way to many algorithms for 3 and 4 because of the parities. This MIGHT be possible in 20 seconds, with a year of training. But I don't think it will be faster then Full Fridrich.

Could you describe yours dChan?

 

[dChan]

Originally posted by AvGalen@Apr 21 2007, 12:39 AM
I was trying to think of something weird, but possible. This method has 6 steps (Full Fridrich has 7) but will use way to many algorithms for 3 and 4 because of the parities. This MIGHT be possible in 20 seconds, with a year of training. But I don't think it will be faster then Full Fridrich.

Could you describe yours dChan?

I don't have a name for it yet so I'll just call it M,U/D(Middle, then U AND D-layers).

1. Solve middle layer edges
2. Piece decomposition(decomposition of pieces into their proper U or D layer, not orientation or permutation- in one algorithm)
3. Orientation and permutation of U and D edges(1 or 2 algorithms)
4. Orientation of U and D corners(1 or 2 algorithms)
5. Simultaneous U and D-layer permutation(Double layer permutation ,or DLP, in one algorithm)

There are 5 steps, which can all be achieved in 1 or 2-looks depending on how advanced you want to be with this system.

It just sort of came about when my sister scrambled the cube and everything else was scrambled except for the middle layer edges- they were perfect. So I've been working on the solution for about 2 weeks now. It can be solved with the algorithms we already know(namely the H, Z , and T PLLs) using setup moves. It requires special algorithms though to be efficient.

Tell me what you think.

 

[dChan]

bump

Another idea:

1. Solve all 12 edges(intuitively? Possible Ortega-style))
2. Orient all 8 corners(7 x 2 algorithms)
3. Permute all corners(1 algorithm of x cases)

This would literally be 3 steps- 4 if you orient U/D corners seperately(1 of 7 cases x 2 layers).

Edit: Maybe use COLL instead of steps 2 and 3. Or an advanced COLL that permutes and orients ALL 8 corners. In this way the method would be 2 steps.

Edit: I just realized that the corners still need to be decomposed into their proper layer so COLL is not an option. Maybe something new- AC(All Corners)?

 

[AvGalen]

dChan: Your first idea

1. Solve middle layer edges
2. Piece decomposition(decomposition of pieces into their proper U or D layer, not orientation or permutation- in one algorithm)
3. Orientation and permutation of U and D edges(1 or 2 algorithms)
4. Orientation of U and D corners(1 or 2 algorithms)
5. Simultaneous U and D-layer permutation(Double layer permutation ,or DLP, in one algorithm)

is not doable for a human being:
Step 1 can be done during inspection.
Step 2 requires an insane amount of algs. This would be like learning all "opposites centers" on a 5x5x5 with the extra difficulty of leaving middle layer edges in place.
Step 3 cannot be done in 1 step. It can be done in 2 steps (using ELL-algs?) or only Orientation in 1 step. Orienting and Permutating edges in 2 layers are just to many algs.
Step 4 is possible 8*8 possibilities (allowing U and D corrections) and there will be mirrors (probably about half of them) among those possibilities.
Step 5. Possible.

I think it is safe to say this method would not work. You cannot just "do a step you normally do with F2L correct" and say "do it on U and D at once".

Method 2 is not defined enough. "Solve 12 edges", sure.
Orient all 8 corners is nog 7x2 algs. One corner in the U layer and 1 corner in the D layer could be misoriented for example
Permute all corners in 1 alg? You should really do some math on the amount of possibilities for this. Blindfolded cubers that use cycles (like me) would love to be able to do this in 1 alg.

Methods like "AC" are only possible if you are a computer or have no time-limit.

 

[dChan]

By "1 or 2 algorithms" I meant 2 sub-steps or 1 if you have a library of algs or are constantly using ACube. This method would require some research with programs to find the proper algs. So will the second one. I say '1 algrotihm of x cases' because the number of possibilities is not defined.

Just some weird ideas anyway. They would be hardly a speedcubing method unless someone were to learn hundreds of algs to make each step 1 look. Step 2 would require 2 looks and can be accomplished with only a T, H, and Z alg. That is how I solve it this way. I also use a commutator to orient.