Hints for speed cubing

Customizing algorithms

It is very important to customize each algorithm for your hands. Some of us are right handed, some left handed, some may prefer algorithms which use only 2 or 3 faces so that alternate twisting from left hand to right hand is avoided. Sometimes, it may be wise to perform an algorithm with the cube turned upside down, or twisted by 180 degrees. This adjustment must be done by each individual separately because everyone may have different views of which algorithms are user friendly and which are not. This takes a lot of time, but it may cut an important chunk from your total time.

Multiple algorithms

As you may notice, some positions in the last layer have several algorithms associated with them. I alternate between them to minimize turning the cube as a whole, thus cutting on time.

Finger shortcuts

Most speed cubists have also developed special sequences ("shortcuts or macros") of two to four moves which can be performed astonishingly fast by pushing the faces with your fingers. Yes, it does require some dexterity. On my video page you can watch me solve the cube a few times. I also perform some finger tricks.

Move algorithms to your subconsciousness

It is also important that your brain automatizes the algorithms into inseparable units - elementary actions, because then you will not have to think about individual moves. The individual moves will be performed "by your hands" rather than making your brain busy. At this stage, one can afford to think more about the next step rather than about the algorithm which is being performed. It is done for you automatically by your subconsciousness! I noticed that this automatization goes that far that if I am interrupted while performing some longer algorithm, I will not be able to finish it! In a sense, I do not know the sequence of moves and perceive the algorithm as one unit. This may sometimes create comical situatioins when somebody asks you about a specific move, and you will not able to show it slowly - and will get stuck after several moves having to start over again to see the remainder of the algorithm.

No delays between algorithms

Another thing which is very important is to cut on delays between consecutive algorithms. One should minimize the decision time to almost zero. This issue is strongly connected with another one - the question of twisting speed.

Faster twisting does not have to mean shorter times

Dogma: One needs to be especially dextered to be able to solve the cube that fast (in 17 seconds). I would be lying to say that some dexterity is not important, but I insist that an average person possesses the necessary dexterity to solve the cube in really short times. I believe that almost everybody can achieve the twisting speed of 3 twists per second. Remember, all you are required to do is to learn a finite set of algorithms perform quickly. This relates to the important issue of adjusting algorithms for your hands. So why is it possible that faster twisting speed may bring you longer times? By performing the moves really fast, one deprives him/herself of the [important] knowledge of what is actually happening to the cube. After performing an algorithm, one is then suddenly thrown into a new position and needs some time to decide which move to choose next. If you had turned the cube just little slower, you could actually see what is happening to the cube, and choose the best next move during the last couple of moves of the previous algorithm. If you compare the times: fast turns + delay between moves and slow turns + shorter delays, you will find out that the second summation may be shorter! Another argument for the second alternative is that it is very hard to turn the cube really fast, and one often encounters "stuck" cubicles, or breaks the cube to its atoms. This can slow you down as well as frustrate.

Preparing the cube for record times.

I have heard people recommend a variety of different lubricants for the cube. Among others, sillicon oil, graphite, and soap were mentioned. From my experience, sillicon oil worked best. Be careful before using other lubricants because some of them may be pretty aggressive and speed up the aging process of your cube. Intense twisting causes a fine dust to develope inside the cube. Some cubists say that this kind of natural lubricant is the best one. I recommend to grease the cube because a lubricated cube will turn easier and you will be able to "cut corners" while speed cubing. But be aware of the fact that putting lubricant into a cube will make the cube more vulnerable to an accidental dismemberment.

Hard work

I would like to end with a couple of more remarks on the cube. First, the secret of achieving amazingly short times is not just the algorithms themselves. After all, a system will never solve the cube. Humans do! Probably the most important factor is dedication and a lot of practicing. As you may notice, some positions in the last layer have several algorithms associated with them. I alternate between them to minimize turning the cube as a whole, thus cutting on time.

So, what is the best system for speed cubing? I do not think that there is such thing as the best system. One system may better fit one person, other system may be more natural for somebody else. I believe that any system which is worked out into sufficient perfection is good. We should not be comparing systems but cubists. Those certainly are comparable.

What are the limits of speed cubing?

Any algorithmic set which can be performed by a human must be limited to a couple of hundreds at most thousands of algorithms. These algorithms need to be performed in a fast manner without too much thinking. This puts limits on the amount of time needed to solve the cube. If there was a hypothetical person who could see the shortest or the almost shortest algorithm right away in the beginning (which is quite improbable), he or she would need about 2 seconds, provided the farthest position is around 20 face moves at the twist rate of 10 moves per second. Since the assumption for this estimate will probably be unrealistic for many years to come, I estimate the limit for speed cubing at 5 seconds (the average time). One should totally abandon the concept of a record time since it has very little informational value. If somebody messes up the cube carelessly, one can take advantage of it and solve the cube in a few seconds. Therefore, for comparing purposes, I suggest to use an average of 10 consecutive times. For my system, I defined the concept of a modified record: I discarded record times whenever more than one stage was skipped during the cube solving. By skipping a stage, I mean: placing the four edges using less than 3 moves, too much luck for the four blocks (in the second layer), skipping the orientation of 8 cubicles from the last layer, skipping the permutation part in the last layer. For the first two layers, it is hard to estimate the probabilities, but the last layer can be calculated exactly. The probability that after solving the second layer, the last layer will have the correct color is 1/216, and the probablity that after orienting the cubes in the last layer one will not need to permute them, is 1/72. So, for example, if the last layer got assembled by chance right after the second layer, I discarded the time since the probability of that happening is too small: 1/(216*72). So, what is my modified record? It is 11 seconds. My best average out of ten was often 17 in 1983. I kept myself in a good shape for many years, and I can still get to an average of about 18 after all those years. Going back to 17 or lower would require a lot of effort, good cube, and a complete devotion that only a rookie can possess. So, good luck everybody and do not give up!

Home page Cube page