• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Random Cubing Discussion

Kirjava

Colourful
Joined
Mar 26, 2006
Messages
6,121
WCA
2006BARL01
YouTube
Visit Channel
btw, here's the scramble from Minh Thai's 22.95 world record solve.

U L2 D' B2 U' R2 B2 F2 D' F2 L2 R2 F R2 D L2 R2 B' L' D' R F'

A reconstruction is in the works :O
 

Cubenovice

Forever Slow
Joined
Feb 6, 2010
Messages
2,291
Location
Vlaams-Brabant (Belgium)
btw, here's the scramble from Minh Thai's 22.95 world record solve.

U L2 D' B2 U' R2 B2 F2 D' F2 L2 R2 F R2 D L2 R2 B' L' D' R F'

A reconstruction is in the works :O

My times on this scramble
CF: 2:38.00
CFOP: 45.50 = my average
LBL as my daughter does it, incl daisy: 1:14.84
ZZ: disqualification (first timed solve, was still checking out EO)
Dimzay EF: DNF Fail on corners (purposly NOT using any known algs)
 
Last edited:

jiggy

Member
Joined
Oct 21, 2009
Messages
454
Location
London (UK)
btw, here's the scramble from Minh Thai's 22.95 world record solve.

U L2 D' B2 U' R2 B2 F2 D' F2 L2 R2 F R2 D L2 R2 B' L' D' R F'

A reconstruction is in the works :O

My times on this scramble
CF: 2:38.00
CFOP: 45.50
LBL as my daughter does it, incl daisy: 1:14.84
ZZ: disqualification (first timed solve, was still checking out EO)
Ooo, that's a fun game!

00:21.81 (about average for me) SUCK IT THAI!!
 

Athefre

Member
Joined
Jul 25, 2006
Messages
1,248
(More) Intuitive Corner Method

Step 1: Three Pieces of a Face

Like Guimond, you'll first be placing three matching or opposite oriented corners on D.

Step 2: Orientation

Now using only F' and R moves, and adjusting U-layer, replace DFL and DBR with corners from U while also orienting two corners on U. After you have done this, you are left with a three move orientation. If you have three oriented corners on U, take the single misoriented corner on U and place it at DFL or DBR with an F' or R move, based on the way it is facing.

Step 3: Three Pieces of a layer

Place three corners of a layer on D in their correct positions using R2/F2/L2/B2/U/D. Have the free slot be at DFR when you are finished.

Step 4: Finish

Your goal here is to pair up all corners. You have one of four options:

1. If the two U corners counter-clockwise to the remaining D corner (that is on U) are opposite, place the remaining D corner at UFR and do R2UR2 (U'R2 afterwards to form the U and D layer).
2. If the two U corners clockwise to the remaining D corner are opposite, place the remaining D corner at UFR and do F2U'F2 (UF2).
3. If there is a pair on U, place the remaining D corner in it's correct position (DFR) using an F2 or R2 move - depending on which of those moves breaks up the U pair (you want to break it up while placing the remaining D corner). Now, after a y rotation, you are left in a position to do option 1 or option 2.
4. If all three U corners are correct or if all three U corners are opposite, place the remaining D corner at UBL and do R2UR2 (or F2U'F2). After a y rotation, you have one of the other options.

Example Solves:

Scramble - D' U R U2 D F2 D R2 F2 B2 L D2 B F' L2 D2 L2 R F R2 L' B2 L' F2 D2

Step 1: y2
Step 2: U'R U2 F'UF
Step 3: UL2 U'F2U'F2UF2 U'

Scramble - U' R2 F' D2 R' D2 R F2 L2 U L' D F2 D' R' B' L' R' B' L' F L B R' L

Step 1: x'y'
Step 2: U'F'U'RF'UF
Step 3: U2F2R2
Step 4: U'F2 y' UF2U'F2UF2

Scramble - F R' F' R F' L R' B' R D' B' F' L' D' U2 B2 D2 L F2 R B D' B R' F2

Step 1: x2
Step 2: U F' F' R F'UF
Step 3: L2D'R2
Step 4: U R2UR2 y U2R2UR2 U'R2

How this started - I saw Thoms post about how he thought there should be something really simple for 2x2x2 that just hadn't been found yet. So, I made it my goal to find something more intuitive than what is out there. I felt like having three pieces on D was great, it's simple. So I kept that step as the first. Having the second step be setting it up into a three move orientation seemed obvious. But, I noticed that working with the free slot at DFR and trying to turn it into that had problems (for a beginner). Often DFR would be oriented while working with the slot (leaving you with a face to have to undo) and there is too much trial and error. I wanted to keep DFR misoriented but still have two oriented corners on U. Then came the idea.

Of course, the orientation step in Guimond is intuitive to more experienced solvers. But, I'm hoping this is easy for a beginner to understand. Hopefully I haven't messed up and missed something that ruins this idea.
I'm considering making a topic about this, but first I would have to improve this post with more details and pictures. I thought I would give it a test here first.
 
Last edited:

Athefre

Member
Joined
Jul 25, 2006
Messages
1,248
Yes I have and it's just a coincidence that they are similar, trust me. The differences I see are:

- In yours it is creating pairs near the beginning, after the three D-layer pieces.
- In yours there is a separation of the layers after orientation and a final permutation of pairs.
- I wanted to avoid a long sequence of moves as the final step (yours isn't very long, I think it's great they you left it at one sequence). I wanted something repetitive, short (R2UR2 or F2U'F2), and easy to understand. I had other ideas for Step 4 (use one pair solving sequence, having one of the D-layer pieces be in the wrong position [for a final R2U'R2, etc. move]), but I felt that this was the simplest.

Most 2x2x2 methods are about the same. It's usually either make three pieces as the first step, or make a layer as the first step.
 

Cride5

Premium Member
Joined
Jan 27, 2009
Messages
1,228
Location
Scotland
WCA
2009RIDE01
Yes I have and it's just a coincidence that they are similar, trust me. The differences I see are:

- In yours it is creating pairs near the beginning, after the three D-layer pieces.
- In yours there is a separation of the layers after orientation and a final permutation of pairs.
- I wanted to avoid a long sequence of moves as the final step (yours isn't very long, I think it's great they you left it at one sequence). I wanted something repetitive, short (R2UR2 or F2U'F2), and easy to understand. I had other ideas for Step 4 (use one pair solving sequence, having one of the D-layer pieces be in the wrong position [for a final R2U'R2, etc. move]), but I felt that this was the simplest.

Most 2x2x2 methods are about the same. It's usually either make three pieces as the first step, or make a layer as the first step.

Step 1 basically has the goal of creating 3 on the bottom and an adjacent pair on the top (all of opposite colours). The pairs just provide intermediate sub-steps to make it easier. The idea in this step is to use just R/U moves. From there its just 3 moves to two oriented faces.

There is no 'separation' step, but instead it is done during solving of permutation. First a pair is built and held in DL, and then R2/U moves are used 'square-1' style to solve the rest.

I had a wee look at your website ... and there was quite an interesting 3x3 method idea. I kindof came up with something similar a while ago too, lol :p ... but I can see there are also differences there as well. Basically a Guimond-style 3x3 solve, starting with orientation, followed by separation, then permutation of individual layers.
http://www.speedsolving.com/forum/showthread.php?t=15534

I guess these kind of 'orient-first' methods can be quite good fun, but probably not so practical in speedsolving - mainly because of lookahead problems. Guimond is probably the exception because it is possible to do all the stuff which requires a glance ad D and B during inspection, leaving the rest of the cube solvable by looking only at U, F and R...
 
Last edited:

Athefre

Member
Joined
Jul 25, 2006
Messages
1,248
Step 1 basically has the goal of creating 3 on the bottom and an adjacent pair on the top (of opposite colours). The pairs just provide intermediate sub-steps to make it easier. The idea in this step is to use just R/U moves. From there its just 3 moves to two oriented faces.

There is no 'separation' step, but instead it is done during solving of permutation. First a pair is built and held in DL, and then R2/U moves are used 'square-1' style to solve the rest.


I actually find your corner method kind of difficult. I find myself "trying" things during the final part of Step 1 instead of understanding what I'm doing. Maybe I'm not reading correctly, or maybe I just need think longer before I make a move.

Cride5 said:
I had a wee look at your website ... and there was quite an interesting 3x3 method idea. I kindof came up with something similar a while ago too, lol :p ... but I can see there are also differences there as well. Basically a Guimond-style 3x3 solve, starting with orientation, followed by separation, then permutation of individual layers.
http://www.speedsolving.com/forum/showthread.php?t=15534


I saw that :) After, and during the creation, I looked around on speedsolving, the speedsolving Wiki, Twisty Puzzles, and the Yahoo! group to make sure there wasn't anything out there like it.
 

Cride5

Premium Member
Joined
Jan 27, 2009
Messages
1,228
Location
Scotland
WCA
2009RIDE01
Step 1 basically has the goal of creating 3 on the bottom and an adjacent pair on the top (of opposite colours). The pairs just provide intermediate sub-steps to make it easier. The idea in this step is to use just R/U moves. From there its just 3 moves to two oriented faces.

I actually find your corner method kind of difficult. I find myself "trying" things during the final part of Step 1 instead of understanding what I'm doing. Maybe I'm not reading correctly, or maybe I just need think longer before I make a move.

Once you have a pair in UL and you're holding one in DL, try this (using only R/U):

* If you see 3 on top 3 on bottom, do R ... if it leaves just 2 pairs then go back to 3 on top 3 on bottom and preserve a different pair in the top, then do R.
* If you have 4/2 and R doesn't create 3/3, then do R to create a 2/2, break up the pair on U to build an alternative pair.
* If you have 2/2 and can't create anything with R-moves, do U R to create an alternative pair on top, you should now be able to get 4/2
 

Athefre

Member
Joined
Jul 25, 2006
Messages
1,248
Once you have a pair in UL and you're holding one in DL, try this (using only R/U):

* If you see 3 on top 3 on bottom, do R ... if it leaves just 2 pairs then go back to 3 on top 3 on bottom and preserve a different pair in the top, then do R.
* If you have 4/2 and R doesn't create 3/3, then do R to create a 2/2, break up the pair on U to build an alternative pair.
* If you have 2/2 and can't create anything with R-moves, do U R to create an alternative pair on top, you should now be able to get 4/2

That definitely helps me, but I think I would have trouble getting a beginner to remember it. I know when I first started I struggled to learn the 7 sequences in the method that came with my cube. It was extremely annoying, making me want to give up sometimes.

Of course, my idea isn't perfect either. I think beginners would have a hard time remembering where they were supposed to place the lone D corner (that is on U) before they do either R2UR2 or F2U'F2.
 

Cride5

Premium Member
Joined
Jan 27, 2009
Messages
1,228
Location
Scotland
WCA
2009RIDE01
I guess the idea is, that if a beginner forgets what to do they can just try randomly turning R/U to create and destroy pairs until they get what they're looking for. It's a bit like returning sq-1 back to a square ... even if you have no idea what you're doing, if you fiddle for long enough you'll get it. Doing this over time you'll eventually build up a system and recognise all the sub-cases.

I think what sets apart an intuitive method from an algorithmic method is that the details don't really matter ... as long as the solver is aware of the sub-goals to be reached, they should be able to kludge together the bits in between, even if the exact moves have been forgotten.
 
Last edited:

Athefre

Member
Joined
Jul 25, 2006
Messages
1,248
Still considering making a topic. But I worry that it will be a repeat of my NMCMLL topic. I'm still not sure if anyone besides Gilles, Thom, and Mini truly understand the recognition.

I'll definitely need pictures to make sure people really see the idea behind Step 2. I found you can not only use F' and R moves during the step, but also F and R' - gotta pay attention to the RUF or FUR sticker though so you don't form a face on D. I can also combine Option 1 and Option 2 in Step 4 and give a better description of those.
 

ThatGuy

Member
Joined
Apr 6, 2009
Messages
625
WCA
2010GUSH01
i had a thought about color neutral training. I've been doing 2gen solves alternating colors so who knows if it works.
 
Top