Really tough challenge for 3x3x3

[Stefan]

Both Ville and I used Corners, Centers, Edges

Me, too, now. Somewhat finished the method and did a handful of solves, best time so far was 4:12.96. More details/algs later.

Layer by Layer was never an option for me, but should be possible as well

Definitely. I checked out the old threads on the yahoo group with the keychain pulling version, back then I solved F2L intuitively (probably Fridrich-style) and then had quite alright somewhat "followable" algs (i.e. not just random five turns repeated thirteen times) for CO, EO, CP and EP:
http://games.groups.yahoo.com/group/speedsolvingrubikscube/message/15614

Anyhow, qqwref, are we going to see an avg12?

Definitely. At least from me.

 

[Stefan]

Now my record is 2:27.25, although that was a really exceptional solve. I'm still unsure about some situations and then I have to think during the solve, would like to get rid of that. More algs could help a lot, too, particularly nice 3-cycles of edges.

 

[qqwref]

Nice job Stefan... I don't think I could ever get times like that without much shorter algs.

 

[Stefan]

What are your algs?

Corners are trivial. Then I solve centers with these:
T perm twice for a 180 degrees center rotation
(r u) (r u' r' u' r' u' r u) to rotate centers around a corner
The latter is easy, the first (r u) creates some distortion and the rest just resolves the corners.

Then for edges I use this double J perm variation:
(r u r' f' r u r' u' r' f r2 u' r') y (l' u' l f l' u' l u l f' l2 u l) y'
My y is more like 45 degrees, i.e., I actually try to stay in real <r,u,f>. Anyway, this does a 3-cycle of edges. I'd like to be able to do it backwards, that would help my time, but I'm too lazy to learn new als so I just do it twice. I use this for everything, including flips.

 

[Stefan]

Getting more consistent:
2:49.54, 2:58.60, (4:00.73), (2:27.80), 2:39.65 => 2:49.26

Now I stop until someone beats me.

 

[Stefan]

Here's a sweet alg: r' f r f' u' r u r' f' u f u'

 

[byu]

Here's a sweet alg: r' f r f' u' r u r' f' u f u'

I'll use that as my scramble and try and solve from there

Total Time: 0.00

Beat that Stefan!

 

[AvGalen]

Can anyone point me to a working link for ksolve?

I found several links in http://www.speedsolving.com/forum/showthread.php?t=1201 but none of them work anymore

This topic also nicely details the differences between Stefan and me:
* I think I discovered a new challenge, Stefan dives deep into cube history and shows me I didn't
* I answer the "is every position reachable" by pointing out things I discovered during playing, Stefan gives/points at the theoretical explanation
* I just get the challenge solved with a KISS-method, Stefan actually searches for algs and practises them
* Some users don't understand the challenge, Stefan does a Pochmann (tm), I politely explain to them what the challenge is

 

[Gunnar]

Can anyone point me to a working link for ksolve?

http://apelgam.se/Rubik/ksolve.zip

For those who doesn't know, it's my brother Kåre who has made the program. You define the puzzle in a textfile, so it can solve pretty much any puzzle.

 

[Spooter]

Can you use inverts?

Inverts of what? Of the scramble? That would be stupid.

Uhhh inverts such as r' u' f'....

Uhm, have you ever thought about your question? Try to solve the scramble u without using u'.

Yah, actually you can, its called u3.....

 

[AvGalen]

Can anyone point me to a working link for ksolve?

http://apelgam.se/Rubik/ksolve.zip

For those who doesn't know, it's my brother Kåre who has made the program. You define the puzzle in a textfile, so it can solve pretty much any puzzle.

Thanks Gunnar

Can you use inverts?

Inverts of what? Of the scramble? That would be stupid.

Uhhh inverts such as r' u' f'....

Uhm, have you ever thought about your question? Try to solve the scramble u without using u'.

Yah, actually you can, its called u3.....

WOW, you don't read other peoples post before you decide to post yourself, don't you

 

[Ville Seppänen]

3:44,91, (5:22,86), 3:18,31, (2:59,89), 3:18,88 = 3:27.37
Dang, doesn't quite beat Stefan.

 

[AvGalen]

3:44,91, (5:22,86), 3:18,31, (2:59,89), 3:18,88 = 3:27.37
Dang, doesn't quite beat Stefan.

.....yet?

 

[Ville Seppänen]

Indeed: (3:10,12), 2:06,45, (1:49,05), 2:21,03, 2:08,25 = 2:11.91

And here an average of 12: 3:44,91, (5:22,86), 3:18,31, 2:59,89, 3:18,88, 2:09,20, 4:52,91, 3:10,12, 2:06,45, (1:49,05), 2:21,03, 2:08,25 = 3:01.00

 

[Stefan]

I wrote a scrambler now so that we have high quality scrambles: http://stefan-pochmann.info/spocc/other_stuff/tools/scramble3x3ruf/. It produces a two-page PDF file with 25 scrambles on each page. Every time you reload the page it should give you different scrambles. The reason for using 40 moves is explained on my tools page: http://stefan-pochmann.info/spocc/other_stuff/tools/.

 

[trying-to-speedcube...]

The scrambler sometimes gives canceling moves between the first and second line.

Not to say it's bad, but it isn't really like it should be, right?

 

[Stefan]

I love KSolve.

Single: 1:08.19

Average: 1:35.36
(1:08.19), 2:16.19, 1:38.78, 1:29.67, (2:21.91), 1:27.78, 1:25.06, 1:41.09, 1:51.26, 1:13.52, 1:10.52, 1:39.70

The move canceling bug should be fixed now, thanks.

 

[Stefan]

Did a few more solves just to get rid of the 2+ minutes times...

Single: 1:04.34

Average: 1:27.44
1:27.78, 1:25.06, 1:41.09, (1:51.26), 1:13.52, 1:10.52, 1:39.70, 1:23.47, 1:31.41, 1:29.09, (1:04.34), 1:32.72

 

[Stefan]

This topic also nicely details the differences between Stefan and me:
* I think I discovered a new challenge, Stefan dives deep into cube history and shows me I didn't
* I just get the challenge solved with a KISS-method, Stefan actually searches for algs and practises them

Well, point 1 is somewhat the reason for point 2. Because I had been working on this before theoretically but never much practically, this time I finally wanted to record some times and get good at it. Also, the algs I used yesterday (when I averaged about 2:50) were just algs I knew, like the Tperm or Jperm, but done with <r,u,f> instead of <R,U,F>. Finding useful combinations of these algs, and finding/learning the setup moves, those were the new and hard parts.

Me digging out the old discussions was btw not meant as "this isn't new, you moron". Just in case someone thought that (not Arnaud, I think he understood my intention). No, I knew there were old ideas but couldn't remember much myself, so I went looking for them. And I like how the discussions referred backwards several times in steps of about two years. I'll probably devote a page on my website to <r,u,f> solving, pointing to the earlier discussions, describing the equivalence with the keychain solving problem, explaining the parity issue, including my scrambler, and finally showing how I solve it now (and others, if they tell me how they do it and if it's fast or interesting).

 

[Stefan]

New average: 1:16.45
1:07.97, 1:14.53, 1:23.64, (1:30.41), 1:09.94, 1:26.80, 1:12.38, 1:11.89, 1:14.61, 1:23.00, (1:06.22), 1:19.77

C'mon, Ville! Catch me!