Swapping two corners on a 4x4x4

[Jason Baum]

Hi everyone,

I've been practicing 4x4x4 BLD stuff a lot recently, and I think I'm about ready to start attempting full BLD solves. One more alg and I think I'll be ready. I need a supercube-safe algorithm for just swapping two corners. Any help would be greatly appreciated!

 

[cmhardw]

Jason, on a super-cube 4x4x4 the parity of the corners must match the parity of the X-centers. Therefore to swap two corners means you must also perform an odd permutation on the X-centers.

What this means for 4x4x4 BLD is that it is usually easier to solve the centers before you decide to fix parity in the corner permutation. Or another option you have is something Daniel Beyer proposed. Fix corner parity before you solve the centers, but most 3x3 algs that do this rotate the U center 1 quarter turn. You could memorize the affect of this quarter turn for your parity fix alg, then turn the U layer such as to undo the effect on the U layer centers. Now solve centers as normal how you memorized them. After that restore the U layer by undoing the U layer turn you did at the start of your centers solve.

Either way would work, but just know that it's impossible to have a supercube safe alg that swaps two corners only. An alg that swaps 2 corners must also perform an odd permutation on the X-centers, leaving you with the above two options for 4x4x4 BLD on a regular 4x4x4. Obvioulsy this case is more interesting when BLD cubing on a 4x4x4 supercube though, and takes a bit more thought to fix correctly ;-)

Chris

 

[hait2]

sorry to jason for a bit of a derail (if it bothers i could find a more suitable thread), but i suppose this is a good a place as any to ask this:

today i got myself a nice little xmas present - a 4x4x4 cube (i couldn't find a 5x5x5 unfortunately, so this will have to suffice for big bld for me)

i figured out how to solve it (since i've done a 5x5x5 before -- once) except the OLL parity (i actually figured out a crappy alg for pll parity. i looked up a better one on cubestation)

etc. etc. so now i am going to do it bld
but i'm facing 2 major problems in memorization

first: there's 2 of each edge and 4 of each centre. the edges are distinct from one another as i could tell by the insides. my question - how can i tell which edge is which during memorization?

same question for centers (but i'm not sure if it applies.. )

any light on this would help me a lot i hope to have done a 4x4x4 bld before the new year
and i'm really sorry if it's a totally newbie question ;(

 

[Jason Baum]

Chris- Thanks for the reply, that makes sense. I was about to go searching for an algorithm that swapped two corners when I realized I can just do the standard PLL parity alg and a T permutation. I practiced it a bit with various scenarios and it works fine. So I know everything I need to know now in order to solve a 4x4x4 blindfolded! Now I just have to actually do it. Last night I tried blindsolving just the edges and made it all the way through but was off by a 3 cycle, but I am very confident that I'll be able to do this very soon.

hait2- For the edges, the best way I've found is to just mentally trace where the pieces have to go. Like you said, each edge has its own unique spot, so you can track every edge this way. It's weird at first, but you get used to it (I'm still getting used to it). For the centers, check out Mike Hughey's guide here, it helped me out a lot: http://www.speedsolving.com/showthread.php?t=2207

 

[masterofthebass]

Jason, my way of doing corners is actually not corner safe (Y perms). I just do corners and fix parity after I do the centers. This way, I just solve the corners, and then I'll be left with edge parity. Then I just swap UL and UB using the normal parity algorithm.

hait2, using r2 makes this a little bit easier, because you can mentally do the set-up to see which piece will be put into the buffer. After a bunch of solves, you eventually learn which piece is which. For the centers, all of the pieces are interchangeable. I would look at Mike Hughey's tutorial for the centers, as it helped me immensely. A simple memo working also is provided.



EDIT--
Apparently Jason and I were writing responses at the same time...

 

[hait2]

thank you very much for the quick responses!

oh yeah one more thing.. i was looking at scrambles from jnet timer and i just realized: i dont know what lowercase letters mean. is it a double layer turn or an inner-layer turn? initially i thought it'd be a double layer turn and then Rr would be inner, but i remembered reading something on this forum about multislice or something..

i jus wanna make sure i scramble correctly

 

[masterofthebass]

On JNetCube, the lower letters mean only the inner slices. THis is the major flaw that I have with JNet. I just use the WCA Scrambler.

 

[hait2]

so i suppose the accepted standard is lowercase means double layer turn, and there's no special notation for inner layer turns at all (so like inner r' would be r'R for example) ?

and mike's tutorial was really really helpful. im probably being far too optimistic, but i hope to learn all necessary concepts today, sleep on it, and attempt tomorrow! i expect a miserable failure but what the hey. if i'm doing everything correctly, i'm not pushing myself hard enough! '

edit: wait nevermind, make that the day after. forgot about boxing day.. i probably won't have any time tomorrow

 

[Mike Hughey]

Just out of curiosity, for those who just solve centers before swapping corners on the 4x4x4 BLD, has anyone come up with a faster algorithm for swapping corners than doing the T-perm followed by the PLL parity algorithm? I figure that's an awful lot of moves; surely there's a faster algorithm that could be memorized for this? Then again, I guess most of you speed demons can do this in less than 4 seconds, so it probably doesn't bother you much. It takes me probably 15 seconds to do this, so a few moves shorter might really help me a little.

 

[Jason Baum]

Wow, I'm so mad right now but kind of happy at the same time. I again tried blindsolving just the edges of the 4x4x4 cube and should have gotten it successfully but did one of the dumbest things. The very last cycle I had was (uRB->rBD), so I did R' U B2 as my set up move and performed the alg r U2 r U2 F2 r F2 l' U2 l U2 r2 to interchange them. Then for some inexplicable reason, to undo my set up move I did U' R B2. As soon as I took the blindfold off I realized what I had did. I went back and undid that last cycle and sure enough that was the only thing that was wrong. So I guess the good news is I came THAT close to successfully doing the edges BLD... but it's so frustrating to have something like that happen.

Oh well. Tomorrow I'll work on the centers a bit more and maybe try just blindsolving the centers. I know I can do the edges now, so if I can successfully do the centers BLD tomorrow maybe I'll go for a full solve! This is a lot of fun!

 

[hait2]

good luck jason!

i've experimented with commutators more extensively and screwed up a bit. so the cube was sorta messed up in a bunch of places, but not everywhere.

anyway i forced myself to solve it cycle wise, freestyling commutators (all sighted of course), starting from the centers and then edges. it took a long time (mostly because i kept screwing up and messing up the cube even a bit more!)
toward the end i started to get the hang of it, and for the past few cycles i didnt make mistakes. i did finish eventually.

this is totally different from my 3cycle method for 3x3x3 though, i didn't really think about the cube in this way before.. it's a bit of a challenge
im goin to sleep now. hopefully my mind will think it over
i think i will get the techniques down soon.. memorizing is a whole other story.
i figured out how to trace the edges to determine which is which btw, great tip jason

 

[Mike Hughey]

Wow, I'm so mad right now but kind of happy at the same time. I again tried blindsolving just the edges of the 4x4x4 cube and should have gotten it successfully but did one of the dumbest things...

I understand exactly how you feel Jason, but this is a great achievement. Not only did you almost get it, but you know exactly what you did wrong. You're essentially already 100% there. I'm guessing it's a matter of no more than a week before you get your first successful overall solve, and it could be as little as a day or two. (Assuming you don't take a break in there.) I'm looking forward to hearing about it!

As for the setup move mistake, it's an easy thing to do - I don't know why. That's why I'm usually really really careful when I'm swapping the last two edges. I've taken to memorizing the moves that each piece makes of the two that are swapping in the setup moves. So if I'm swapping G and Q in my memory scheme, I'll memorize "G goes to E, then to B, then Q goes to A" to get them next to each other. Somehow that's easier for me to remember when undoing the setup moves.

 

[Jason Baum]

Okay, I just successfully did the edges blindfolded! I tried doing centers before that attempt and was way off, so I think I just need more practice with my eyes open, but I think it's safe to say that I have the edges part of the solve down now. I'm really excited, I never thought I would ever be able to do this puzzle blindfolded, and now I'm so close. I'm very slow though, I think my first completion of the whole cube will be well over an hour. Is this a normal first time? I don't want to get super fast with this, but I would like to have a decent time in competition someday (somewhere around 20 minutes would be nice).

 

[cmhardw]

My first ever 4x4x4 BLD solve was on a 4x4x4 supercube and, including memorization, took me 5.5 hours. So to be near an hour, and possibly sub-2 hours, sounds fast to me for timing of a first ever solve ;-)

Chris

 

[Mike Hughey]

Of course, keep in mind that Chris did that many years ago, when no one had developed these nice systems that we're able to just steal and use now, so Chris's perspective is rather warped on this. He thinks I'm improving fast, to let you know how warped it is.

However, Jason, you will be VERY surprised how fast you can improve at this. Once you get good at doing the commutators sighted, your speedsolving ability will kick in, and I'm sure you will be as fast as I am within weeks or even days. At least, that's what everyone else on here has done. And you're a faster cuber than most of them.

My first attempt at a 4x4x4 BLD was 88 minutes, and my first successful one was 63:42.73. So I would also consider an hour a normal first time. But if you miss a few of them before you get one right, I'm guessing you could easily get down to 30 minutes by the time you get your first one.

I suspect that by the time you have a chance to try this in a competition, you won't be satisfied with any less than 15 minutes. I think 20 minutes will be quite easy for you, as long as you get lucky and get one right. Even I'm averaging well below 20 minutes now.

By the way, how fast are you these days at 3x3x3 BLD? (Your most recent time on speedcubing.com is 2:30, but it's over 2 years old!)

 

[Jason Baum]

By the way, how fast are you these days at 3x3x3 BLD? (Your most recent time on speedcubing.com is 2:30, but it's over 2 years old!)

That 2:30 is still my best time actually. Ever since I got that time, I have RARELY practiced BLD until last week or so. I guess I'm just starting to get back into it after not having practiced it in a long time. I've seriously done more 3x3x3 BLD solves in the past couple of days than I have in the past year or two. I'm starting to get decent at it again, I'm a lot more consistent and I've had a couple sub 3 solves. I usually average somewhere around 3:30. I'd like to average sub 2.

 

[cmhardw]

Jason let us know how your full attempt goes when you do it, and good luck!

Chris

 

[pjk]

Congrats Jason on doing this. I plan to start working on 4x4 BLD soon as well. And I have also been working on 3x3 BLD. I avg around 3:00, which is pretty sad considering that I memo in just over a minute. But hopefully I can avg 2:30 or so by April. Dan Knights, and a couple of us from the Denver area are working on a competition in Denver for April 12th. If you by any chance have the time/money to come to Denver, you should. Dan Knights should give you some 3x3 competition.

Sorry to get off topic. Once again, congrats.

 

[Mike Hughey]

Patrick, that's interesting. I also memorize in just over a minute for the 3x3x3 BLD, and average around 3 minutes total time. And just like you, I consider this to mean I need to work a lot more on execution than on memorization (although clearly memorization could use some work as well).

 

[ROOT]

im about to work on 4x4 BLD, but before i memo im trying a full solve with it all written down, then im doing it seperately then fully. and anyone else notice that your times improve much faster in bld than regular. i just started bld 2 months ago and im already getting sub-3 times.