One-Answer Puzzle Theory Question Thread

[shadowslice e]

So I just thought it might be useful to have a one question answer thread instead of having a new thread everytime someone want to ask a simple question.

Anyway, my question is: where can I find a copy of the pruning tables?

Thanks

SSE

 

[Kudz]

I got one question:
What is probability of LL skip after scrambling cube 2gen and solving it 2 gen too. EP is done and CP too.
Will it be: no. WV cases x no. EPLLs?

 

[shadowslice e]

I got one question:
What is probability of LL skip after scrambling cube 2gen and solving it 2 gen too. EP is done and CP too.
Will it be: no. WV cases x no. EPLLs?

I think it would be 1/(CO cases*EP cases)/4/4 or around 1 in 40 I think though that may be a bit high. Someone correct me if I'm wrong please.

 

[ch_ts]

Anyway, my question is: where can I find a copy of the pruning tables?

Short answer: people make their own pruning tables

Longer answer: (I assume you're talking about pruning tables for a computer solver) Yeah, people make their own. If you're thinking of a table that could be read by humans and be useful for a humans, (for example, case xxx is solved by the alg xxxxx) this is not that kind of table. Pruning tables would be an array (probably) that had estimates of distances from solved. For example, it might tell you that position #5032 takes at least 6 moves to solve, and I might have a million (or whatever) of these cases. I say "at least" because it might take more moves. As an example, a pruning table might tell me that a particular corner configuration takes 6 moves to solve when I ignore the edges. If I do include the edges, then it is likely to require more moves. In general, you don't want to overestimate the distance from solved, you only want to underestimate (but you do want to be as close as possible) Pruning tables can be made by starting from the solved position and exploring outward from there and recording how many moves it took to get the positions. And I might have several pruning tables together, maybe one for corners while ignoring edges, maybe one for edges while ignoring corners, maybe some combination of corners and edges, etc.

 

[Cale S]

I got one question:
What is probability of LL skip after scrambling cube 2gen and solving it 2 gen too. EP is done and CP too.
Will it be: no. WV cases x no. EPLLs?

Probabiity of 2GLL skip?

CO is done 1/27 and EPLL is done 1/12 so 1/324

 

[shadowslice e]

Short answer: people make their own pruning tables

Longer answer: (I assume you're talking about pruning tables for a computer solver) Yeah, people make their own. If you're thinking of a table that could be read by humans and be useful for a humans, (for example, case xxx is solved by the alg xxxxx) this is not that kind of table. Pruning tables would be an array (probably) that had estimates of distances from solved. For example, it might tell you that position #5032 takes at least 6 moves to solve, and I might have a million (or whatever) of these cases. I say "at least" because it might take more moves. As an example, a pruning table might tell me that a particular corner configuration takes 6 moves to solve when I ignore the edges. If I do include the edges, then it is likely to require more moves. In general, you don't want to overestimate the distance from solved, you only want to underestimate (but you do want to be as close as possible) Pruning tables can be made by starting from the solved position and exploring outward from there and recording how many moves it took to get the positions. And I might have several pruning tables together, maybe one for corners while ignoring edges, maybe one for edges while ignoring corners, maybe some combination of corners and edges, etc.

Yeah, that's what I meant when I said pruning tables I was just wondering if there are any available online for CO+EO (or solving 1 face at a time) or other stuff like that cause I didn't want to have to go through the process of creating my own (call me lazy but... )

I'll make my own if there aren't any but I was wondering if anyone had one first or some sort of template for one of that kind. I'm sort of going to see what may happen if you combine two tables and what effect it has on the rough distance to solved.

 

[Lucas Garron]

While a separate thread isn't really a problem, I encourage you to post in the more relevant threads/forums instead. That makes these questions easier to find, and might get faster replies for people paying attention to those specific threads.

Anyway, my question is: where can I find a copy of the pruning tables?

This fits better in the software area subforum.

Anyhow, ch_ts has already answered this, but also note that there are different possible pruning tables for any given puzzles. For example, you could have a pruning table for EO or for CP. Or a combined one that takes both into account.

I got one question:
What is probability of LL skip after scrambling cube 2gen and solving it 2 gen too. EP is done and CP too.
Will it be: no. WV cases x no. EPLLs?

This should probably go in the probability thread (thriving with regular replies since 2010).

 

[shadowslice e]

What's God's number in STM/QTM/ATM?

 

[Cale S]

What's God's number in STM/QTM/ATM?

QTM: 26
STM: not known, but it's 18-20
ATM: not known

 

[sqAree]

What's God's number for the 3^4 and if not known, maybe for the 2^4?

 

[kbrune]

What the maximum number of move for optimal solving of any given 2x2x2 block on a 3x3?

 

[shadowslice e]

Are there any ways if classifying cubes according to invariants like how some knots can be expressed as polynomials?

 

[techgeek1129]

I hear it everywhere like 2gen F2L and stuff. What does it mean?

 

[Ranzha]

I hear it everywhere like 2gen F2L and stuff. What does it mean?

In general, 2-generator ("2-gen") means using only two types of turn. The most common example is <R,U>, which means using R-turns and U-turns only (R, R2, R', U, U2, U'). The next most common is <M, U>.

 

[techgeek1129]

In general, 2-generator ("2-gen") means using only two types of turn. The most common example is <R,U>, which means using R-turns and U-turns only (R, R2, R', U, U2, U'). The next most common is <M, U>.

thanks

 

[vm70]

How did anyone find out the PLL+Parity algorithms for even cubes? Did they find the algorithms by adding parities one by one to every single standard 3x3 permutation?
I'm not talking about the standard H, Z, etc.
This is about Opp. Edges, Adj. Edges, Opp. Corners, Adj. Corners, O, W, P, D, I, C, Q, K, Ξ, and ϴ.

EDIT: Too many perms! Just shortening to letters.
https://www.speedsolving.com/wiki/index.php/4x4x4_Parity_Algorithms

 

[AlphaSheep]

How did anyone find out the PLL+Parity algorithms for even cubes? Did they find the algorithms by adding parities one by one to every single standard 3x3 permutation?
I'm not talking about the standard H, Z, etc.
This is about Opp. Edges, Adj. Edges, Opp. Corners, Adj. Corners, O+, O-, W, P[a], P, P[c], P[d], D[a], D, I, C[a], C, Q[a], Q, Q[c], Q[d], K[a], K, Ξ, and ϴ.
https://www.speedsolving.com/wiki/index.php/4x4x4_Parity_Algorithms

Three methods that I can think of:

• As you suggest, combine the 3x3 and parity algs. Some times you can insert one into the other and cancel moves. This tends to give long algorithms though.
• You can generate them using computer programs like ksolve.
• You can derive them manually if you have a good understanding of what causes parity and can come up with short sequences of moves to affect pieces in certain ways, and then combine these sequences in the form of commutators and conjugates to get the desired effect. Someone like @Christopher Mowla can probably explain this a lot better than I can.

 

[Franklin]

Ok I think this is where I put this question but
What is the definition of Orient and Permute?
I use to think orient = that piece is where it's suppose to be but might need to be twisted and
permute = solved
but that is clearly not the case

 

[GenTheSnail]

Ok I think this is where I put this question but
What is the definition of Orient and Permute?
I use to think orient = that piece is where it's suppose to be but might need to be twisted and
permute = solved
but that is clearly not the case

Your definition of oriented (piece is where it's suppose to be but might need to be twisted) is when a piece is permuted.
Oriented is when an edge can be solved with only <R, U, L, D> moves, and when a corner by only <U, D, R2, F2, L2, B2> moves.

For instance, OCLL (Orient Corners of the Last Layer) is when you twist all the corners so the *yellow* is facing up. The pieces are not necessarily in the right place though.
PLL (Permute the Last Layer), is when all the pieces have *yellow* facing up, and you put them in the right place.

 

[MethodNeutral]

The easiest way to just swap two corners is to use Niklas, right? It's only five moves, just enough to swap them. How can I do this with edges? I ask this because I've been looking into the Tripod method. The last layer for this already has a 2x2x1 block solved, and I was interested in seeing if I could use a short alg (like Niklas) to swap the remaining two edges while keeping all of the solved cube so that I'm left with just a commutator.

(I understand that it's just cycling corners so that they appear swapped)