TRIM a.k.a. the rice method

Discussion in 'General Speedcubing Discussion' started by rice, Mar 8, 2012.

Welcome to the Speedsolving.com. You are currently viewing our boards as a guest which gives you limited access to join discussions and access our other features. By joining our free community of over 30,000 people, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is fast, simple and absolutely free so please, join our community today!

If you have any problems with the registration process or your account login, please contact us and we'll help you get started. We look forward to seeing you on the forums!

Already a member? Login to stop seeing this message.
  1. rice

    rice Member

    25
    0
    Mar 8, 2012
    Background to the method. See [this post] for the latest developments.
    -----------------------------------------------------

    CFL stands for

    Step 1: Corners
    Step 2: First Six Edges
    Step 3: Last Six Edges

    We are already familiar with Corners, just use your favorite 2x2 method. LSE should be familiar to most of us also, refer to any rouxtorial for a refresher if you need it.

    FSE is divided into 3 sub-steps: solving 2 adjacent E layer edges, solving the other E layer edges, then DL/DR or DF/DB while solving for the L/R or F/B centers. To solve for these edges, we will pair them up and then insert them into the correct position. While FSE is meant to be executed intuitively, some algorithms are provided below for illustrative purposes.

    For the entirety of FSE, the following applies. U and D turns, as well as any slice moves, can be freely performed without affecting the corners. This may lead to some confusion at first as the U and D layers won't necessarily be aligned with each other and the centers will be all over the place. All other turns X that precede an algorithm must conclude with X'. Since there is no EO in FSE, some pairs will have skips or shorter solutions than others.

    Step 2a
    There are four possibilities for the first pair: FL/FR, RF/RB, BR/BL, and LB/LF.

    Both edges will be in the U or D layer 42.42% of the time, so our job will be easy here. All we need to do is execute U/D/slice moves to pair up the edges. Let's say we opted to pair up the FL and FR edges and the pair is on the U layer. AUF to put the pair in the UF/UB position, then execute M' F M2 F' or M' F' M2 F.

    24.24% of the time, one edge will be in the U or D layer and the other edge will be the E layer. To solve for the pair, move the first edge into a position that would allow you to R' E' R or L E L'. Hopefully, our pair will be in the correct position, as indicated by corner alignment. If not, move the pair to the FL/FR position, then F M2 F2 M2 F or F' M2 F2 M2 F'.

    If both edges are in the E layer unoriented, it would be best to choose another pair. There is a 6.06% chance of this occurring.

    Step 2b
    Assuming our first pair is in the FL/FR position, perform an E2 or y2 to store it in the back. Now, E moves are forbidden and R/L moves only should be used in conjugates, such as R U' M2 U R', L F' M2 F L', and their many varients.

    Both edges will be in the U or D layer 62.22% of the time, so Step 2a can be repeated.

    17.77% of the time, one edge will be in the U or D layer and the other edge will be the E layer. Use a conjugate, then execute M F M2 F' or M F' M2 F to store our second pair in the front.

    If both edges are in the E layer unoriented, that really sucks. There is a 2.22% chance of this occurring.

    Step 2c
    We are now confined to U/D/M/S moves. There are two possibilities for the last pair: DF/DB and DL/DR.

    Pair up two edges of whatever the D color is and move it to the U layer. Align the D-layer corners with the E-layer edges. Suppose that we paired up edges that belong in the DF/DB position. The center of the F layer must be the opposite color of the corner/edge pairs in the F layer. If not, move the center into position without breaking up the third edge pair.

    To insert our third pair, AUF the FU sticker to match with the center and then perform M2. An E slice can be done if the center is already positioned and the pair is in the BU/BD position to preserve the center. Do a y/y' turn to move on to LSE.

    Optimization
    With Steps 1 and 3, our work is already cut out for us. Step 2 can be improved by positioning pairs without needing to align corners, reducing cube rotations, pairing edges from different angles, and finding shortcuts. The example FSE solve shown below is not particularly optimized, but we can assume the average move count of CFL to be around 45 to 55 STM.

    tl;dr FSE example solve
    Another solution for the example scramble

    Video example of edge pairing: http://www.speedsolving.com/forum/showthread.php?35847-CFL-aka-the-rice-method&p=727926&viewfull=1#post727926

    List of insightful solves
    1. Cool Frog - 43 moves
    2. Kirjava - 35 moves
    3. Ranzha V. Emodrach - 37 moves
    4. Cool Frog - 33 moves
    5. Ranzha V. Emodrach - 32 moves
    6. y235 - 38 moves
    7.
    8.
    9.
    10.
     
    Last edited: Mar 23, 2014
  2. Kirjava

    Kirjava Colourful

    6,130
    3
    Mar 26, 2006
    WCA:
    2006BARL01
    YouTube:
    snkenjoi
    waterman/roux hybrid

    it works, nice to see a proposed method actually be something new
     
  3. benskoning

    benskoning Member

    285
    0
    Aug 13, 2011
    Wisconsin
    WCA:
    2011SKON01
    YouTube:
    benskoning
  4. TheAwesomeAlex

    TheAwesomeAlex Member

    140
    0
    Nov 1, 2011
    Earth
    WCA:
    2011MAAS01
    I actually use this method
     
  5. rice

    rice Member

    25
    0
    Mar 8, 2012
    What did you use before and what prompted the change to CFL?
     
    Last edited: Mar 9, 2012
  6. oll+phase+sync

    oll+phase+sync Member

    361
    0
    Oct 1, 2010
    I tried to use something similar long time, but inserted any L- with any R-edge instead of strictly using L- R-edges from the same face, so CFL also solves the trouble of doing mixed color LSE.

    When do you solve centers (L/R)?

    Do you have a strategy to benefit from a single, already solved edge?


    Regarding move count
    corners: 16
    first 6edges: 21 (maybe to optimistic)
    LSE: 15

    =~ 52 (wich I think is reasonable for a simple/intermediate Corners Fist method - I should really sit down and measure that for som 50 solves)
     
    Last edited: Mar 12, 2012
  7. rice

    rice Member

    25
    0
    Mar 8, 2012
    If you have a guide/tutorial for that, I would be willing to look it over.

    L/R centers are solved simultaneously with the insertion of the third pair.

    A solved edge in Step 2a is easy: Just have the two edges in the FL/UR position, then do R' E' R.
    For Step 2b: You could do L' U M2 U' L M F M2 F' or R' F' M2 F R F' M2 F for a FL/UR pair.
    One way you could solve for Step 2c: If the pair is in the DB/RU position, M' U will suffice if the centers are already solved.

    While I am happy to provide algs for certain situations, FSE is meant to be intuitive. Depending on the case, complex conjugates with multiple set-up moves can be used, though care has to be taken to preserve the edge pair when undoing the set-up moves.
     
  8. TheAwesomeAlex

    TheAwesomeAlex Member

    140
    0
    Nov 1, 2011
    Earth
    WCA:
    2011MAAS01
    I was trying to learn the beginners method but I was too lazy to learn the algs for the last layer so i thought this would be easier to learn since i knew how to do a 2x2
     
  9. rice

    rice Member

    25
    0
    Mar 8, 2012
    Some editing done to the OP and most interestingly, I found a 13 move solution to the example scramble.
     
  10. oll+phase+sync

    oll+phase+sync Member

    361
    0
    Oct 1, 2010
    I didn't have time to do lots of solves, but my current experiences are:

    - I'm most of the time I'm turning the whole cube and do the insertion in the U layer

    - 2a - this is the hardest so far (even is somteimes just 4 moves), I don't feel like having the time to look for the optimal pair, instead if I spot any L/R-edge it take that and search the corresponding - sometimes it's obvious it a ugly pair but searching a different pair would be to much loss of time.

    - 2b if much more focused since there is no freedom in choosing an edge.

    - 2c I regulary use DL/UR- , UR/DL-pairs to optimize this step
     
  11. rice

    rice Member

    25
    0
    Mar 8, 2012
    I'm thinking about generating EG algs that don't affect the edges and that should help with pairing them up since hypothetically, a cuber could use inspection and step 1 solve time to hunt for edges, thus ensuring a steady, flowing solve. As a nice side-effect, it could also be used for bld. Is there any interest for this?
     
  12. http://tinyurl.com/algR2F2RF-R-xy

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

    R2 F2 R F' R' xy' RUR'U'R'FRF' u/ corners lol
    M2 xy'z' / R face edges
    E' B' L E2 L' B x' / L EDGES
    D' M U2 M' U2 D S x2/ first two block lol

    U M2 R U' r' U' M' U r U R' M' U' u2 M' u2 M'
    43 moves.

    Linear BTW
     
  13. JonnyWhoopes

    JonnyWhoopes Premium Member

    1,003
    6
    Jun 25, 2009
    Massachusetts, USA
    WCA:
    2011GRAY02
    YouTube:
    JonnyWhoopes
    Yes, and you should name the variant Fried Rice.
     
  14. oll+phase+sync

    oll+phase+sync Member

    361
    0
    Oct 1, 2010
    I actually started to plan ahead 4 Corners and 1 edge, then continue with an CLL that don't touch this 1 edge (this is no real restriction). Step 2a is much easier this way ( just one edge to solve and you already know the colors of this edge during inspection.

    A variation of this would be to pair up two edges during inspection and easily place them after the corners step. (I think that's the same idea as yours)

    But while I may use/learn edge-save-PBL I would not go for any EG ideas.
     
  15. rice

    rice Member

    25
    0
    Mar 8, 2012
    Good point, it's better to know all of the effects of an alg than to learn an edge-saving version, which btw has a very high move count.

    Cool Frog, would you mind if I added your post to a list of insightful solves in the OP?
     
  16. rice

    rice Member

    25
    0
    Mar 8, 2012
    If anyone finds new shortcuts or had an amazing solve, please post it and I'll add it to the list.
    btw, has anyone tried a CN solve?
     
  17. Kirjava

    Kirjava Colourful

    6,130
    3
    Mar 26, 2006
    WCA:
    2006BARL01
    YouTube:
    snkenjoi
    "a CN solve"

    >_>
     
  18. Ickenicke

    Ickenicke Member

    908
    0
    Jul 19, 2011
    Sweden
    YouTube:
    NickeCubing
    Is this method having potential?

    Corners is possible to do in 2-3 seconds (2x2)

    LSE is also possible to do very fast.

    How fast do you guys think FSE can be?
     
  19. Kirjava

    Kirjava Colourful

    6,130
    3
    Mar 26, 2006
    WCA:
    2006BARL01
    YouTube:
    snkenjoi
    D2 B2 U' F' R' L2 B F2 R2 L D' U' B U' F' R2 L D2 L' U2 D' R' F2 R' L2

    F' R x U2 F L2 U' y R2 D // Step 1
    r2 F M2 F' R2 // Step 2a
    E L' b' M2 F M U // Step 2b
    L2 U' M2 U R2 // Step 2c
    U R2 U' y M2 U F2 M' U M' U2 // Step 3

    35 moves
     

Share This Page