K4 Final

[Kirjava]

The updated/fixed/final/correct guide is now available publicly at http://kzx.ath.cx/.

It's basically been rewritten since the last one since I've changed the method, and that the last one was a bit ****.

Been passing it around for a while, but might aswell announce it.

k4e.txt is an example solve.

Will be the last incarnation until I discover something new

(PstScrpt: K4 == Direct solve 4^3 method)

 

[pjk]

I know Richard is averaging sub-1:10 with a hybrid k4 method, and had a 55.xx today when we were racing.

 

[Jason Baum]

Awesome, I've been waiting a while for something like this. How many algs does it take you on average to solve the LL edges? I might start learning/practicing this, it seems like a fun method.

By the way, on the opposite edges cases for step 7, the last alg is
ru2lD2l'U2lD2l'r'. It should be rU2lD2l'U2lD2l'r'

 

[Kenneth]

Blocking up all faces exept the last one in step 4, before F3L is ended must be pretty turn consuming. That's why I do F2B before the last four centres. that way I can do six edges before centres using short U/r/l operations to pair them up. But on the other hand, you got easier centres than in my method. But, again, I do two addtiional edges while bulidng last four centres.

I always end the solve in opposite edges because I do not like the adjacent parity cases, the opposite algs are much faster using only U2/F2/B2/D2 and the M slice layers, you only need to orient x and do all as U2/M turns, the same goes for the opposite commutators.

I would also like to hear about your average move count, my guess is sub 130

I was working on my guide for my direct method (see the guides section here at Speedsolving) but got totaly into the "floating edge" method for 3x3x3 that I'm currently working on (I write a thread in a day or two =) so the last parts have to wait

 

[Kirjava]

By the way, on the opposite edges cases for step 7, the last alg is
ru2lD2l'U2lD2l'r'. It should be rU2lD2l'U2lD2l'r'

Cheers, I'll change it when I get home. I sat down with a cube and typed all the algs out in one go, I've already had to fix another. (I think that's all of them now)

tbh, if you are competent enough at 4x4x4 solving, the guide is redundant and can be shortened to (2 Opp Centres > 1x3x4 > 4 Centres > First Layer > F3L > C*LL > ELL)

 

[death_kg]

Oh Thanks!
I finally find it...
I played Rubik's Revenge for 1 day and I wanted to study the K4 method.
THX!

ps Does anyone have a web page that teaching "how to solve a Rubik's Revenge"?

 

[rachmaninovian]

hehe anyone tried k6?wouldn't it be hard?ha ha

 

[Kenneth]

Would be the same but you have to do twice as many ELL steps. Or not really because often, if pices are paired you can do more than one in one alg using double layer slices.

Fun thing is that you can get the parity twice, one for wing edges and one for inner wing edges, but you can solve both in one alg with double slices.

 

[Stefan]

Some presentational suggestions:

- Don't use <pre>, this prevents line breaks and forces me to scroll to the right a lot.
- Add an overview page showing pictures of all stages.
- Why zip? I'd prefer simply browsing your site.

 

[Kirjava]

Alg's fixed.

Good suggestions, it's a zip because I was distributing it to a couple of people at first so I just uploaded it, extracted now.

I widened the IFRAME, any other changes are too significant to be done at this time.

I'm not a proper web developer, I just make it look right for my browser/resolution/OS

 

[brunson]

- Don't use <pre>, this prevents line breaks and forces me to scroll to the right a lot.

+1 </pre>

 

[death_kg]

I tried all LL algorithms and I made some picture to it, do you want it?
Actually , I didn't play 4*4*4 cube so much..
What is the most useful algorithm in those LL algorithms?
thx!

 

[DeathCuberK]

I find it hard to build blocks in the K4 method, probably because I am a fridrich user. Is it still considered K4 method if you solve centers, then pair edges to make a cross, F3L, then last layer? Is kinda more.... fridrichy?

 

[Kirjava]

You bumped the wrong topic