• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Not a one in 12 chance to assemble the cube correctly?

SenileGenXer

Member
Joined
Mar 7, 2012
Messages
206
Location
Virginia, USA
We know there are 12 orbits of the cube. It is assumed that the probability of randomly assembling the cube and having a solvable puzzle is 1 in 12.

Have we checked that?

I was thinking of all the ways it's possible to misassemble the cube.

1) Clockwise corner twist (1 in 3 chance)
2) Counter-Clockwise corner twist (1 in 3 chance)
3) Edge flipped (1 in 2 chance)
4) Two edges swapped (1 in ? chance)
5) Two corners swapped (1 in ? chance)

The last part is a mistake you could make assembling the cube. Now because of 4-cycle moves like the T-Perm and F-Perm two corners swapped is indistinguishable from two edges. They will show up as the same PLL parity error.

If you make mistake #4 and mistake #5 randomly assembling the cube they cancel each other out - there is no observable mistake. A cube assembled with two edges swapped and two corners swapped is solvable with a F-Perm/T-perm.

I was thinking that instead of 12 orbits there were 14 orbits but two pairs of orbits were conjoined. That is the solvable orbit is conjoined with the two edges and two corners swapped orbit. The two edges swapped orbit is conjoined with the two corners swapped orbit. I hypothesized that this produces a 2/14 chance of randomly assembling a solvable cube. 1/7 when simplified.

To test it I disassembled a cube into a brown paper paper bag, shook up the bag, reassembled the cube randomly and tried to see if it was solvable. Kept track of 50 random assemblies. If my theory was correct I should get 7.1 solvable cubes. If the conventional theory was correct I should get 4.1 solvable assemblies.

I got six.

Data available here.
 
Last edited:

Carrot

Member
Joined
Feb 9, 2009
Messages
1,910
WCA
2008ANDE02
YouTube
Visit Channel
for 4) and 5) combined you have these scenarios:
even corners + even edges = solvable
odd corners + even edges = unsolvable
even corners + odd edges = unsolvable
odd corner + odd edges = solvable
Each of above case has the same probability, hence 2/4 chance of an unsolvable permutation of the pieces.

also, number 1) and 2) are wrong.
there's 1/3 chance of getting corner orientation correct, all 2 corner twists can be made into 1 corner twists, therefore 2/3 for 1) and 3/3 for 2).

E: Also, your dataset only has 56% corner fails, and you KNOW it should be like 67%, therefore you are kind of bound to get more solvable cubes than expected. All in all, just no, it is 1/12.
 
Last edited:

SenileGenXer

Member
Joined
Mar 7, 2012
Messages
206
Location
Virginia, USA
1/3 chance of getting a clockwise corner twist. 1/3 chance of getting a counter clockwise corner twist. 1/3 chance of getting it correct. More along the lines of what I was thinking. Post updated.

Even and odd edge/corners sounds right but are you sure about their probability?
 
Last edited:

Carrot

Member
Joined
Feb 9, 2009
Messages
1,910
WCA
2008ANDE02
YouTube
Visit Channel
1/3 chance of getting a clockwise corner twist. 1/3 chance of getting a counter clockwise corner twist. 1/3 chance of getting it correct. More along the lines of what I was thinking. Post updated.

Even and odd edge/corners sounds right but are you sure about their probability?

Yes I am.
 

Hypocrism

Member
Joined
Aug 4, 2012
Messages
316
WCA
2009ADLA01
One of the permutation sets can be entirely ignored (because an odd permutation of edges or corners is not by itself impossible - see T perm). The other has a 1/2 chance of being solvable (given the edge permutation mod 2, the corner permutation must be the same mod 2).

Just like we don't count the orientation of all the corners when calculating 1/3 chance of corner orientation being solvable - we only need to take into account the "8th" corner because given the position of the other 7, it is the 8th which determines the cube's solubility.
 

Stefan

Member
Joined
May 7, 2006
Messages
7,280
WCA
2003POCH01
YouTube
Visit Channel
If you make mistake #4 and mistake #5 randomly assembling the cube they cancel each other out - there is no observable mistake.

There's not just no observable mistake, there's no mistake. You can't make mistake #4 and mistake #5.

I get the feeling this thread is a hoax, especially with the 50 assemblies experiment. Way too much work for a way too small sample size. Hard to believe you actually did that :D
 
Last edited:

AvGalen

Premium Member
Joined
Jul 6, 2006
Messages
6,857
Location
Rotterdam (actually Capelle aan den IJssel), the N
WCA
2006GALE01
YouTube
Visit Channel
I like how he checks theory with reality, but "his math" of "12 becomes 14, so 1/12 becomes 2/14" makes no sense to me.
also....that samplesize.
Has anyone ever done a computer simulation where all pieces are put randomly and then checked for solvability? That I would expect a few million states to be checked in a second (and the 1/12 to be clearly showing)
 

TMOY

Member
Joined
Jun 29, 2008
Messages
1,802
WCA
2008COUR01
I like how he checks theory with reality, but "his math" of "12 becomes 14, so 1/12 becomes 2/14" makes no sense to me.
The flaw in his reasoning is pretty obvious: he's assuming his 14 differentevents all have the same probability of happening, which unfortunately is not the case.
 

SenileGenXer

Member
Joined
Mar 7, 2012
Messages
206
Location
Virginia, USA
I like how he checks theory with reality, but "his math" of "12 becomes 14, so 1/12 becomes 2/14" makes no sense to me.
also....that samplesize.
Has anyone ever done a computer simulation where all pieces are put randomly and then checked for solvability? That I would expect a few million states to be checked in a second (and the 1/12 to be clearly showing)

My thinking was looking at the orbitals. I don't know if anyone has named and numbered them but I think they work out like this:

1) You can assemble it correctly.
2) you can assemble it with a corner twist (+1)
3) you can assemble it with a CC corner twist (or -1)
4) You can assemble it with an edge flip
5) Edge flip plus corner twist
6) Edge flip plus CC corner twist
7) You can assemble it with a PLL error
8) PLL error and a corner twist
9) PLL and a CC corner twist
10) You can assemble it with a PLL error and and edge flip
11) PLL error plus edge flip plus corner twist
12) PLL error plus edge flip plus CC corner twist.

These are the orbitals generated by the 1/3 * 1/2 *1/2 math. the 12 states that exist. That's the result but that's not how a cube is assembled. I was thinking instead of PLL error there were a differentiation between corner PLL error and edge PLL error as you assemble it and before it's solved. A proto-orbital with both edge and corner parity was conjoined to the solved orbital - 2 chances to assemble it correctly instead of only 1/12 results.

In the cold light of morning I can't figure out where the 14 came from. If I make a list of proto-orbitals there are 24. 2/24 = 1/12
 
Last edited:

Christopher Mowla

Premium Member
Joined
Sep 17, 2009
Messages
1,184
Location
Earth
YouTube
Visit Channel
en.wikipedia.org/wiki/Rubik%27s_Cube said:
519 quintillion[SUP][28][/SUP] possible arrangements of the pieces that make up the Cube, but only one in twelve of these are actually solvable. This is because there is no sequence of moves that will swap a single pair of pieces or rotate a single corner or edge cube. Thus there are twelve possible sets of reachable configurations, sometimes called "universes" or "orbits", into which the Cube can be placed by dismantling and reassembling it.
Clearly only one third of corner orientation positions are reachable, but if we twist a single corner we can reach all other 3^8 - 3^7 possible corner orientation positions with legal moves.

Therefore, shouldn't there just be (2)(2)(2) = 8 "cube universes instead of (2)(2)(3) = 12 "cube universes"? It seems there has been a confusion of the orientation operations with the orientation positions.
 

Stefan

Member
Joined
May 7, 2006
Messages
7,280
WCA
2003POCH01
YouTube
Visit Channel
if we twist a single corner we can reach all other 3^8 - 3^7 possible corner orientation positions with legal moves.

If we twist a single corner one way we can reach other 3^7 possible corner orientation positions with legal moves.

Instead of the cube, imagine the integers, and instead of turning, imagine adding or subtracting 3. Then you have three "universes":
{..., -6, -3, 0, 3, 6, ...}
{..., -5, -2, 1, 4, 7, ...}
{..., -4, -1, 2, 5, 8, ...}
If you're in one "universe" (i.e., at one number in that "universe"), you can make moves (add or subtract 3) as much as you want, you stay within that "universe". Just like the three corner orientation "universes".
 
Last edited:
Top