On Speedcubing

Video

Watch me solve the cube in 12.79 seconds

History


Matt Wilder introduced me to the cube.
(Looking Glass Rock, 2006)

My first cube

Although I wrote my Virtual Cube in the spring of 1998, I never actually knew how to solve it on my own until a year later. When I got my first cube in March of 1999, my roommate Matt Wilder already knew how to solve it in under a minute, so I had high standards for speed-cubing from the beginning. I wanted to be able to average 30 seconds by the time I graduated in 2.5 years. I immediately borrowed "Conquer that Cube" from the library and got to work memorizing the solution. In a few days I was already doing time trials with the book in my lap. I averaged 3 minutes. After a month or two of non-stop cubing I started to get down around 1 minute, so Matt had to get faster to keep ahead of me. He got down to 50 seconds, and then I started to catch up again. Then he dropped into the 40's, and I caught up again, and we continued leap-frogging all the way into the low 20's.

Mirek Goljan, Jessica Fridrich, and I.
(2003 World Championship in Toronto)

Meeting Jessica Fridrich

The most significant break-throughs came from learning Jessica Fridrich's solution method. Her web page is what inspired me to keep pushing past 30 seconds. During the summer of 1999, after I had been cubing for about 4 months and was averaging about 26 seconds, I met Jessica and her friend Mirek Goljan. This visit was indispensible in my development as a speedcuber, because they demonstrated in person some of their "finger tricks," (i.e., clever ways to flick the sides of the cube without letting go). Jessica hadn't cubed in months (she had a hard time even finding a cube around the house), but she was still averaging 17-18 seconds consistently.

Erno Rubik and I.
(London, 1999)

Meeting Erno Rubik

I began to post my progress on my web page, and several months later, in the fall of 1999, I received an email from a TV producer for a British morning show called, "The Big Breakfast." She told me that Erno Rubik was coming to their talk show, and that they wanted a few speedcubers to be on the show with him. When I agreed to come, she replied, "Okay, great. Are you in London, and can you be at the studio Tuesday morning?" It was Friday, and I was in Vermont. After some deliberation, I eventually replied, "yes, I'm in London, and I can definitely be there." I then found a cheap flight to London, and four days later I was on the show with Rubik.

During the filming, I noticed two women offstage who where pointing at me. When the show ended, they introduced themselves as Rubik's public relations team. It turned out that Rubik was at the start of a week-long publicity tour around England, and they liked the idea of having a young speed-cuber beside him for the TV appearances. They invited me to come with them for the week, and we all headed off to Rubik's hotel. Over the next few days, we did a number of talk shows and news interviews (ITV news, Granada Tonight Show, etc.), several radio shows, and a few newspaper interviews. Rubik was friendly, though generally quiet and pensive. He solved my cube.

Solving the Cube during Freefall.

Pre-Championship

By the spring of 2000, I was averaging around 18 seconds. Over the next couple of years, I dropped down to 16.8. I made a few other TV appearances, including a reunion show of, "That's Incredible." I also made two somewhat embarassing solution videos for Rubiks.com, and solved a cube during freefall for a pilot show on VH1.


Dan at the WC 2003 Championship.

The 2003 World Championship

The first World Championship since 1982 was held in 2003, in Toronto. I had the second fastest single time (16.71), and the fastest average during the finals, which put me in first place. My final average was terrible, 20.00 seconds, but everyone else choked under the pressure just as much as I did. After this, I had quite a few more TV appearances, including The Ellen DeGeneres Show and The Today Show. I also wrote the new solution booklet that is included with Rubik's Cube.

Dan Knights on the "set" of Cubers.
(Asheville, 2007)

2003 to Present

In 2004 I became a high school math teacher, and put the cube aside for about 2 years. I went to the 2005 World Championship, but I was completely "out of shape" and bombed all three trials in the qualifying round. Despite my poor performance, I enjoyed seeing friends from the 2003 championship again, meeting new cubers from around the world, and learning all sorts of new finger tricks and algorithms. I also found it extremely frustrating to watch people solve the cube in an average of 13-14 seconds, when such times would have been unthinkable just a few years earlier. The experience invigorated me. In 2006, I began speedsolving again, and learning the new "state of art" solving techniques. I have gotten my best average down to 13.82 seconds (as of June, 2007).

Also, stay tuned for the release of, "Cubers," a documentary by Canadian director Richard LeBlanc that follows me and several other speedcubers from the 2003 World Championship through the present.

My Method

I use the standard Fridrich method. This method is for expert speedcubers, and requires one to memorize around 100 special algorithms. If you are looking for a good beginner's solution, I suggest you use this one.

Fridrich solves the cube as follows:
  1. Solve the first side cross (I do this on top, and then turn the cube over). (about 2 seconds, done by intuition)
  2. Insert each of the four first-layer corners with its neighboring middle-layer edge piece. (1.75 seconds each = 7 seconds, requires 20 algorithms and some intuition).
  3. Orient all corners and edges in the top layer simulaneously (i.e., get the top all one color). (2 seconds, requires 57 algorithms)
  4. Permute (i.e. position) all corners and edges in the top layer simultaneously (2.5 seconds, requires 21 algorithms)

This method yields a total average solution length of about 56 turns. Learning all of the algorithms required for this method, however, takes months of practice and hard work. For those who wish to use the "layer-by-layer" approach, but have too much of a life to learn these 100 algorithms, I suggest a fast but much simpler top-layer method: do steps 1 and 2 as above, but use steps 4-7 of my beginner method to solve the top layer. With this method I can average under 30 seconds, and around 70 turns for the whole cube.

If you are planning to learn the Fridrich method, then I suggest first learning Fridrich steps 1, 2 and 4 from above, and then using steps 4 and 6 of my beginner method to replace Fridrich's step 3. This will allow you to average under 20 seconds, and around 65 turns for the whole cube, but requires you to memorize only 4 algorithms in place of Fridrich's 57 for step 3.

Here is a comparison of 3 speed methods for solving the cube.

Intermediate - 4 Looks, Min. 25 seconds Advanced - 3 Looks, Min. 18 seconds Expert - 2 Looks, Min. 12 seconds
Solve the Cross 7 turns
Insert Bottom Corners with Middle Edges 27 turns
Orient Edges 6 turns
Position Corners 8 turns
Orient Corners 11 turns
Position Edges 8 turns
Solve the Cross 7 turns
Insert Bottom Corners with Middle Edges 27 turns
Orient Edges 6 turns
Orient Corners 11 turns
Permute All (Fridrich step 4) 12 turns
Solve the Cross 7 turns
Insert Bottom Corners with Middle Edges 27 turns
Orient All (Fridrich step 3) 9 turns
Permute All (Fridrich step 4) 12 turns
Total: 67 turns
Total: 63 turns
Total: 56 turns

Getting a Good Speedcube

  1. Buy a DIY (Do-It-Yourself) kit from Rubiks.com. Assemble it with 0.8-1.2 mm of clearance (when pulled against the spring) between the base of each center piece and the central mechanism. These cubes have been excellent speed-cubes recently (as of March 2007).
  2. Buy stickers from cubesmith.com
  3. Lube your cube. Use pure silicone spray. Make certain that there are no petroleum-based ingredients. Pop an edge piece out and spray one or two small drops of lubricant inside. Play with it for a few minutes to spread the lube around.
  4. Clean your cube frequently. Dust builds up inside and will slow you down. I clean my cubes about every 300 solves. Disassemble the cube and wipe the dust completely off of all interior surfaces using a damp cloth or paper towel.

Speedcubing Tips

Finger Tricks

Speed cubers turn the cube at an average of 3-4 turns per second, with bursts of over 9 turns per second. For example, I can perform the 17-turn "Y Permutation" (F R Ui Ri Ui R U Ri Fi R U Ri Ui Ri F R Fi) in 1.72 seconds, which is 9.9 turns per second.

Finger tricks have come a long way since the 1980's. I started cubing in 1999, before webcams and digital video cameras were commonplace, so the only finger tricks I knew where those I picked up in person from Jessica and Mirek, and those that I discovered myself. Unfortunately, my cubes were relatively stiff, so my tricks were limited to very stable and forceful ones using only the right hand. It wasn't until after the 2005 championship that I actually began to "flick" with my left index finger and do "F" and "Fi" turns as triggers without regripping. I also learned to turn two layers at a time in certain situations (i.e. turn "U" and the middle layer instead of turning "D"). These new tricks are extremely important for anyone who wishes to be a world-class speedcuber.

I used to have a set of video clips on my page demonstrating certain finger tricks, but these were mostly "old school" tricks. There are now several speedcubers who have posted effective videos and/or explanations of more "new wave" tricks:

Intead of re-posting a list of videos of specific finger tricks, I have instead posted videos of some of my favorite top- layer algorithms. If you are interested in learning about finger tricks, these videos should be sufficient. They demonstrate nearly every trick I use.

Faster Algorithms

Computer solvers have generated excellent new algorithms for most situations in Jessica Fridrich's method. Cubers around the world share their favorites on the web, and as a result everyone can have access to the best algorithms. I mostly use those found on Lars Vandenberg's page. For the "H" permutation (swap opposite edge pairs), I use Bob Burton's excellent trick. By picking and choosing my favorites, I can now average under 2 seconds for nearly all of the PLL (last layer permutations--Fridrich's step 4) cases.

Because there are such great speedcubing resources on the web these days, I have not posted my own full algorithms page. I do, however, have a few favorites that I would like to share with the community:

Fridrich step 3: Orient Last Layer
  • Double-Pi case

  • standard: B L U Li Ui Bi F U R Ui Ri Fi [12 turns]
    performed: f U R Ui Ri F fi U R Ui Ri Fi
    best time: 1.36 seconds
  • Y with bar

  • standard: R2 Ui R F Ri U R2 Ui Ri Fi R [11 turns]
    performed: (R2 Ui) (R F) (Ri U) (R2 Ui) (Ri Fi R)
    best time: 1.37 seconds

    Fridrich step 4: Permute Last Layer

  • "F Permutation"

  • standard: (Ui) Ri U R Ui R2 Fi Ui F U R F Ri Fi R2 [13 turns]
    performed: (Ri U R Ui) (R2 Fi) x y (Ri U R) yi (R U Ri Ui R2)
    best time: 1.67 seconds
  • "V Permutation"

  • standard: Ri U Ri Ui Bi D Bi Di B2 Ri Bi R B R [14 turns]
    performed: (Ri U li fi) x (Ri U Ri Ui) xi (R2 Ui) (Ri U R U)
    best time: 1.69 seconds
  • "Y Permutation"

  • standard: F R Ui Ri Ui R U Ri Fi R U Ri Ui Ri F R Fi [17 turns]
    performed: F (R Ui Ri) Ui (R U Ri) Fi (R U Ri Ui) (Ri F R Fi)
    best time: 1.72 seconds
  • "T Permutation"

  • standard: R U Ri Ui Ri F R2 Ui Ri Ui R U Ri Fi [14 turns]
    performed: (R U Ri Ui) (Ri F) (R2 Ui) (Ri Ui R) (U Ri Fi)
    best time: 1.51 seconds
  • "R Permutation"

  • standard: Ri U2 R U2 Ri F R U Ri Ui Ri Fi R2 (U) [13 turns]
    performed: (Ri U2) (R U2) (Ri F) (R U Ri Ui) (Ri Fi R2)
    best time: 1.52 seconds
  • "Z Permutation"

  • standard: R Bi Ri B F Ri F Bi Ri B R F2 (U)[12 turns]
    performed: (R Bi Ri B) (F Ri F) Bi (Ri B R) F2
    best time: 1.46 seconds
  • "H Permutation"

  • standard: M2 U M2 U2 M2 U M2 [7 slice turns, 11 face turns]
    performed: M2 U M2 U2 M2 U M2
    best time: 1.10 seconds

    Other Tips

    1. Spend some time minimizing the number of times that you turn the whole cube during your solution, especially during the first two layers. As an exercize, try to solve the first two layers without ever changing the orientation of the cube. Then take these ideas and try to optimize your speed-cubing so that you use as little turning of the whole cube as possible, while still allowing some turning when it makes the moves faster for your hands. The less you have to turn the cube around, the easier it is for you to look ahead.

      Practicing this will often force you to insert corner-edge pairs in the back slots of the cube. I recommend inserting in the back as much as possible during actual speedsolving, because it allows you to watch the remaining empty slots more easily.

    2. It is also very important to minimize the number of turns during your first two layers. There are often shortcuts available that get rid of uneccessary turns while you are inserting a corner-edge pair, especially during the first few corner-edge pairs when you have still have some "empty" slots left. You should be able to average under 27 turns for the four corner-edge pairs with some practice.

      Macky's page has a great introduction to using empty slots. There are also a number of specialized algorithms available for solving F2L cases with empty slots or using double-layer turns. A great reference by Dave Orser is here, at speedcubing.com. I have also found several other useful algorithms, which I present here. I especially like case 5c.4.

    3. Between 2005 and 2007, I cut my personal best average from 16.8 seconds to 13.82. One of the key break-throughs that has allowed me to break through the 15-second barrier is a change in the quality of my mind while speedcubing. There are two ways in which I have changed the quality of my mind:

      1. The first was attending a Vipassana meditation camp. I have experimented with several different kinds of meditation, and they all have some benefits. I believe this particular technique, however, to be unique and profoundly effective. Vipassana is sanskrit for, approximately, "seeing clearly." The technique is surprisingly simple: one trains the conscious mind to be aware of all sensory input, and then one simply relaxes while observing said input. In practice, this is not as easy as it sounds. The amount of raw data streaming into the brain at any given time is enormous--every single nerve in the body is constantly sending sensory signals to the brain. The subconscious or "reptile" portion of the brain monitors these inputs 24 hours per day, but filters out the vast majority before passing only a few signals on to the conscious or "intellectual" portion of the brain. The problem, according to teachers of Vipassana, is that the subconscious is constantly reacting to these signals using very primal and irrational instincts. The subconscious tends to over-react, causing stress and anxiety, but the conscious brain is helpless to intervene. With Vipassana, one brings the conscious mind down to the same level of perception as the subconscious, and then coaxes the subconscious into a calmer behavior pattern. The changes can be significant and permanent.

        The best way to learn Vipassana is by attending a 10-day course at one of many centers around the world. The courses are free, and are supported by donations made by previous students. If you are considering going to a course, please feel free to contact me for more information.

        After attending one 10-day course, my cube times dropped immediately. Vipassana helped me re-train my subconscious to be more calm and relaxed. This has helped my speedcubing in that when I get nervous, the physiological effects (shaky hands, fast heartbeat, impaired cognition) are much less severe. I am also able to concentrate better while speedcubing, because my mind is less apt to wander.
      2. The second is a subtle but highly effective mental exercise I call, "suppressing motor attention." Once you have learned the algorithms for the Fridrich method, there are two main areas in which to get faster: 1) faster execution of the algorithms (finger tricks), and 2) being able to "look ahead" and discover the next step while you are still finishing execution of the current step. The technique of "suppressing motor attention" helps me in part 2, "looking ahead."

        Every speedcuber understands that, with practice, algorithms become encoded in "muscle memory" so that the conscious mind does not need to think about the execution of the algorithm--the hands seems to do the work on their own. In fact, most speedcubers will tell you that they cannot execute certain algorithms slowly, because they no longer remember the turns in conscious memory. The execution of a cube algorithm can be likened to touch-typing one's own name or email address: the fingers do all of the work, leaving the conscious mind to think about other things.

        I noticed that my conscious mind has a strong natural tendency to focus on what my fingers are doing, even though the fingers are being controlled by muscle memory. Perhaps this is a natural function of the brain. Muscle memory does most of the work, but the conscious mind still pays attention in case something goes wrong. Ideally I would let the muscle memory perform all of the physical work of solving the cube, while 100% of the conscious attention would be devoted to looking ahead to determine the next step. Unfortunately, I find that my natural tendency is to waste almost all of my conscious attention on watching my fingers, even though the muscle memory is perfectly capable of doing it alone.

        When I realized this, I began training my conscious mind to suppress or ignore the work being done by muscle memory. Overcoming this strong natural tendency has required constant vigilance, and I still often catch my conscious mind watching the pieces I'm solving instead of searching for others to solve. When I do manage to disconnect the two processes, however, the effect is dramatic--I can often solve the cube in 12 or 13 seconds with ease. In contrast, when I am failing to suppress motor attention, and I am allowing the conscious mind to "indulge" in watching the fingers, I struggle to solve the cube in under 15 seconds.

        I don't have a perfect system of training for suppressing motor attention. It is more an exercise of mindfulness. The following two training exercises help, though:

        1. While solving the first two layers, pause halfway through inserting the first pair. Find the second pair, and then force yourself to watch the second pair as you allow your hands to complete the first pair. Repeat with the remaining pairs. The most important part is that your eyes are physically tracking other pieces on the cube, instead of the pieces you are solving.
        2. Solve the cube with a metronome, first at 1 turn per second, then 2 turns per second, then 3 turns per second. Not being allowed to "miss a beat" forces you to search for other pieces while your fingers are finishing an algorithm.

        Good luck, and Happy Cubing!

    Visitor Counter by Digits

    Contact me: danknights "at" gmail "dot" com.