[Help Thread] Big Cubes Discussion

[Queenofkings]

Go to 5:10 in the video. He uses a 6x6 for his example, but it works for any big cube.

Thanks for the suggestion, but that disrupts the edges. I’m looking for a way to move the center pieces without having to redo the edges.

 

[Nmile7300]

Thanks for the suggestion, but that disrupts the edges. I’m looking for a way to move the center pieces without having to redo the edges.

It is much better and faster to solve the centers before the edges on big cubes. The only real exception is the Yau method and even then, that's just a few edges, not all of them.

 

[xyzzy]

I used to solve the 5x5 by doing the 4x4 first (ignoring the middle slice) then solving the 3x3 (big corners) to get the edges. That would bring me to the attached photo where I had a technique to swap the center edges until they were complete, finally solving the 3x3 with big centers to finish. That was years ago and I’ve forgotten what technique I used - muscle memory is giving me: u M u’ M’ after rotating the outer slices to get the configuration I want but of course that algorithm messes with all the centers so even if I get one right the others are screwed up. Does anyone have a suggestion? All of the center fixes I can find are for before the edges are done.

Commutators. Some examples:
3R U 3R' 2U 3R U' 3R' 2U'
3R U 2R U' 3R' U 2R' U'

 

[DNF_Cuber]

You can basically do niklas but with slices for the Ls and Rs

 

[Queenofkings]

It is much better and faster to solve the centers before the edges on big cubes. The only real exception is the Yau method and even then, that's just a few edges, not all of them.
[/QUOT

It is much better and faster to solve the centers before the edges on big cubes. The only real exception is the Yau method and even then, that's just a few edges, not all of them.

Yeah that’s how I do it now, I used to use the 4x4 method I described when I taught other people how to do it it b/c it was easier for them sometimes.

 

[ABCubeTutor]

I used to solve the 5x5 by doing the 4x4 first (ignoring the middle slice) then solving the 3x3 (big corners) to get the edges. That would bring me to the attached photo where I had a technique to swap the center edges until they were complete, finally solving the 3x3 with big centers to finish. That was years ago and I’ve forgotten what technique I used - muscle memory is giving me: u M u’ M’ after rotating the outer slices to get the configuration I want but of course that algorithm messes with all the centers so even if I get one right the others are screwed up. Does anyone have a suggestion? All of the center fixes I can find are for before the edges are done.

my new method solves this, and also offers a new way to solve the 5x5x5. https://www.speedsolving.com/wiki/index.php/ABCube_Method

not to pat my own back, but there is a new method for solving big cubes, and it undoes parity and eliminates memorizing algorithms. it is https://www.speedsolving.com/wiki/index.php/ABCube_Method it is not great for speedcubing, but excels at large cubes, it does the centers last because the parities can be erased by a simple center slice turn,

 

[TheSlykrCubr]

what should i average on 6x6? I use Redux btw

average under 1:10 on 4x4 (Yau)
and under 2:30 on 5x5 (Redux)

 

[xyzzy]

what should i average on 6x6? I use Redux btw

average under 1:10 on 4x4 (Yau)
and under 2:30 on 5x5 (Redux)

Probably 4:30-5:00.

Rule of thumb is that compared to 555 times, 666 times are a bit under double and 777 times are a bit under triple.

 

[abunickabhi]

Probably 4:30-5:00.

Rule of thumb is that compared to 555 times, 666 times are a bit under double and 777 times are a bit under triple.

Yeah true, 6x6 times are double of 5x5 times.
and 7x7 is triple of 5x5 times.

 

[Existential Shrimp]

I was struggling a lot with l4e on 5x5. I watched a Cubeskills video on it, and it was really helpful, but it was still primarily focused on slice flip slice. Is there another method for l4e that I should be using?

 

[ZB2op]

Any 5x5 centers tips.

 

[KWCuberWasTaken]

hi, so I am wondering what are the best commutators for the 6x6 last two centers, mainly for the last 2 center edges (2 on each center need to be solved)

thanks a lot!

edit: didnt know how this worked, this was my first time posting something. i will ask in a different forum

 

[MJS Cubing]

Hi, welcome to the forums! This video may help, not sure what level you are at.

You may have better luck asking this question in a mega thread, and you can find these by searching in the search bar. Here's a big cube discussion thread that I'm sure will help you.

[Help Thread] - Big Cubes Discussion

I am using 32223 right now, but I have wanted to test Yau method also. Where can I find a tutorial?

www.speedsolving.com

 

[KWCuberWasTaken]

Hi, welcome to the forums! This video may help, not sure what level you are at.

You may have better luck asking this question in a mega thread, and you can find these by searching in the search bar. Here's a big cube discussion thread that I'm sure will help you.

[Help Thread] - Big Cubes Discussion

I am using 32223 right now, but I have wanted to test Yau method also. Where can I find a tutorial?

www.speedsolving.com

thanks! will do!

 

[KWCuberWasTaken]

okay i probably sound like a complete beginner asking this but what are some commutators for 6x6/7x7 last two centres? mainly for the very last part. also this is my first reply in a thread so please tell me if I did something wrong

 

[KWCuberWasTaken]

Any 5x5 centers tips.

build blocks. i would start with 2x2 or 1x3 blocks as far as i know. a mistake i made when starting out was i made a cross on the centre and then inserted the corners. make sure not to do that

 

[BenChristman1]

okay i probably sound like a complete beginner asking this but what are some commutators for 6x6/7x7 last two centres? mainly for the very last part. also this is my first reply in a thread so please tell me if I did something wrong

Go to about 5:10 in this video. He uses a 6x6 as an example, but it works for any big cube.

 

[KWCuberWasTaken]

Go to about 5:10 in this video. He uses a 6x6 as an example, but it works for any big cube.

thanks so much man. appreciate it!

 

[carcass]

Any 5x5 centers tips.

Use T shapes to your advantage, not enough solvers see this as a quick way to get a 2x2x3 block

 

[xyzzy]

(transferring to this thread)
(also I say "you" a bunch of times but I mean that generically rather than carcass specifically)

Here is a trick to avoid triple parity on 6x6. Finish all of your edges, and if you have parity, don't do it. Pick a cross color so the parity edge will end up in your last layer. If you aren't color neutral, do a Uw Dw' Flip Uw' Dw or something like that to make the edge yellow. Then, when you get to OLL, you can be sure if you have inner or outer parity.

The value of this "trick" is predicated on triple parity being the worst thing ever, and since that is a cowdung premise, this trick turns out to be not really that useful in practice. Like, I use it in maybe 10% of my solves and it's not completely useless, but its value is way overblown by beginners who think it's some super fancy trick.

So this trick lets you "avoid" "triple parity". But concretely, what's happening here?

1. What is triple parity? It's when you get parity during edge pairing (parity of the inner wings and parity of the outer wings differ), then you get OLL parity (odd parity both for inner wings and for outer wings), and finally PLL parity.

2. What is meant by "avoiding"? The simplest interpretation is just that you won't encounter the above scenario.

Okay, fine, so this trick does allow you to avoid "triple parity". So what is the alternative? To encounter the triple parity scenario, we need these conditions to line up once the centres are solved (**):
(i) Inner wings have odd parity.
(ii) Outer wings have even parity.
(iii) Edge groups and corners together have odd parity.

These events are independent (*) and each happen with probability 1/2 (**), so even without going out of your way to avoid triple parity, it only happens one-eighth of the time anyway. When this does happen, with the basic "beginner" method, you end up doing three parity algs. That sucks, right? (No, it doesn't; you just have an irrational hatred of parity algs because they took you a long time to learn. This is also besides the point.) But going back to my original question: what is the alternative?

(i) You see that there's one edge group left to fix, so you decide to leave it to the end. (No parity alg here!)
(ii) You are restricted in your choice of cross colour. Actually not that big of a deal because it's not like you could choose the best cross in one second anyway, and the restriction can help you look ahead into cross better.
(ii.5) If you're not colour neutral (or at least, quad CN), sometimes you'll get the messed up edge on an inconvenient colour. That sucks. Maybe you'd want to slice-flip-slice to transfer it to a yellow edge group instead. Not as bad as a parity alg, but slice-flip-slice is not a free operation.
(iii) After finishing F2L, you count the number of bad edges and see that you have inner parity. (We're focusing on the situation where you'd otherwise have gotten triple parity. It's impossible in that situation to have outer parity here.) Gosh, the inner parity alg stinks. So many slice moves!
(iv) Do you use OLL parity tricks like using a trigger setup to force OLL skip on the 1-flip L cases? Too bad, you can't use them in conjunction with the inner parity alg.
(v) Oh, and after you've solved inner parity, you still have to deal with PLL parity. No change there.

Congratulations, you replaced two bad parity algs with one super-bad parity alg, and one of the three parity algs you originally had to do didn't even change at all. Granted, inner parity is not as bad as having to do both edge pairing parity and OLL parity, but it's barely any better.

Oh, do you use a dumb version of the inner parity alg where you have to do U2 l r' at the end? Even worse; you just eliminated whatever little benefit this trick had.

(*) Almost, but the deviation from independence is negligible. (If you want to go into the nitty gritty details, you do have to take into account that humans do not solve twisty puzzles by rigidly following a flowchart.)
(**) Assuming you don't use the parity L2E algs. This is actually important! Read on.

------

Now, if you just take everything I wrote in the above section at face value, you might come to the conclusion that at least the trick isn't actively harmful if you're using a proper inner parity alg. Which is true, but not the right perspective. The right perspective is that once you start learning the parity L2E algs (here, have one: r U2 r U2 F2 r F2 l' U2 l U2 r2), the value of this trick plummets even further. It's never worth it to force this trick when you have a parity L2E case other than the OLL parity case, and that's also the rarest L2E case. The cost-benefit analysis is a bit more involved here, so bear with me for a moment.

Oh, and in case you haven't noticed, PLL parity is actually irrelevant to the discussion here, so I'll be ignoring that henceforth. (It's kind of funny that that's the case because this trick is always marketed as "triple parity avoidance" and then it turns out one of the three parities is unavoidable.) The following four scenarios are equally likely (1/4 each).

Scenario A: Even inner parity, even outer parity.
Do edge pairing as usual. No edge parity, no OLL parity.

Scenario B: Odd inner parity, odd outer parity.
Do edge pairing as usual. No edge parity, yes OLL parity.

(These two scenarios above do not involve the trick at all.)

Scenario C: Even inner parity, odd outer parity.
Do edge pairing up to L2E. Pair up the inners too, if they're not already. (***) Depending on your edge pairing strategy for L4E, the probability of getting the OLL parity case will differ (e.g. it's almost zero with AvG-like edge pairing), but it'll be somewhere between almost-0 and 1/6. At least 5/6 of the time you'll be getting something else, which is the interesting subcase.
If you use the trick: You do slice-flip-slice here. Get outer parity when you reach OLL later.
If you don't use the trick: You do the parity L2E alg here. No parity when you reach OLL later.

Scenario D: Odd inner parity, even outer parity.
As in Scenario C, do edge pairing up to L2E, and there'll be a ≥5/6 chance of getting a not-OLL-parity L2E parity case.
If you use the trick: You do slice-flip-slice here. Get inner parity when you reach OLL later.
If you don't use the trick: You do the parity L2E alg here. Get OLL parity when you reach OLL later.

In Scenario C, you're clearly better off not using the trick. Scenario D is a bit fuzzier, but if you look at the average move counts, using the trick costs around 8 (slice-flip-slice) + 22 (inner parity) = 30 moves, while not using the trick costs around 15 (parity L2E) + 17 (OLL parity) = 32 moves. The numbers will vary somewhat depending on your exact choice of algs, but my point here is that even if you do save on move count by using the trick here, it's a tiny saving. You can't quickly distinguish between Scenario C and Scenario D, so conditioned on being in C or D, by using the trick, there's a 50% chance you're wasting moves big time (Scenario C) and a 50% chance you're saving maybe a little bit (Scenario D). The trick turns out to be a net negative here.

(***) Ha ha, would you believe it if I said it can get even more complicated? Ha… ha… (I'm not writing up any more than this today, but hint: inner parity algs other than OLL parity.)

------

tl;dr Were you expecting a summary here? Hell no, I spent Time writing it so you should spend Time reading it too. triple parity avoidance trick overrated, do not mindlessly (ab)use