Rubik's Cube Record Broken

martinX writes "The Courier Mail has a story about a San Fran software developer who spins the cube faster than anyone else on the planet: 20 seconds to solve Rubik's Cube. I didn't think anyone still played with them." The winner, Dan Knights, said "It's sort of a dream come true", and credited Jessica Fridrich's cube-solving method as the key to winning, leaving the originator of the method in second place at the World Rubik's Cube Championships in Toronto at the weekend.

Comments

Obligitory Lego URL

[Kris_J]

http://jpbrown.i8.com/cubesolver.html

20 seconds???????

[Eager Newbie]

That's way faster than I could solve the cube...

By peeling and re-applying the stickers.

Re:Video?

[jnana]

You can see the creator of the Fridrich method solve the cube here: http://www.ws.binghamton.edu/fridrich/video.html

I think I saw him...

[whorfin]

I think I saw him doing a little sidewalk show at Fisherman's Wharf a couple weeks ago. I thought "this is odd, there's a guy showing people he can solve the cube in under 30 seconds...", and just walked past.

Little did I know I was passing by my chance to meet a world champion. I just hope he doesn't let it go to his head and end up screwed up like Mike Tyson or Todd Marinovich.

Solve times

[Kethinov]

I know a guy who can solve them in under 3 minutes. He memorized all the algorithms and so on. Fascinating concept. I never knew they were actually solvable until I met him. A lot of people joke that "oh gimme a screwdriver and I'll solve it" but when you break it down step by step, they actually do have a simple set of steps that can be taken to solve them, like any other logic problem. Although he did tell me that they can be taken apart and rearranged in such a way that they CAN be made unsolvable. He told me "never try to solve a cube until you've seen it solved. You never know what kinds of assholes will stack the deck on you to make you look stupid."

Re:Solve times

[mcmonkey]

Sub-3 minute times aren't too hard with a little practice. I have "The Simple Solution to Rubik's Cube" (not my auction, just the first link to the book I found).

As you can see from the picture, it's barely a book--more of a pamphlet. There aren't more than a couple dozen algorithms in the whole thing, and I don't think more than 8 or 10 possible positions you have to look for at any one stage.

The plan here is to get you solving the puzzle by minimizing the number of cubes you're positioning at once and so minimizing the number of possible combinations to memorize.

I've topped out at a little under 2 minutes, and I'd bet even with a good, well-oiled cube this method tops out at over a minute. The short-comings of this method really show when you get to the last 4 cubes to place. There's basically one configuration that leads to the solution. And you repeat a short series of moves until the last four cubes are in that configuration.

It wastes time, but you only need to memorize 1 configuration--the solvable one--and 2 series of moves--1 to solve and 1 to rearrange.

I thought sub-minute times were world-class, but I had no idea how much had been done sub-30 seconds. I think a question better than vi or emacs is top-down or bottom-up.

initial state of cube?

[croddy]

how is it possible to have a rubik's cube competition? is there a standardized initial state for the cube?

Re:initial state of cube?

[zero_offset]

From the rules:

Scrambling:
In all portions of the championship, puzzles will be scrambled using random moves generated by computer. The same sequences of random moves will be applied to the puzzle of each competitor to ensure each competitor will be starting with the same random puzzle state. This same method of scrambling will be used during the averages or best of 3 ?- type portions of competition. These random moves will be applied by independent jury-members.

The number of random moves depends on the puzzle. See the below example for explanation.

Rubiks 3x3x3: 25 moves
Rubiks 4x4x4, 40 moves
Rubiks 5x5x5, other: 60 moves etc

But what I REALLY want to know is, how the hell do you solve a cube while blindfolded? Seriously, there are three categories of blindfolded competitions.

Re:initial state of cube?

[Soul-Burn666]

In some competitions you are allowed to examine the cube before actually moving parts... So this might be the case.

If i'm not mistaken, the fastest for a pre-examined cube is something like 9 secs.

Re:initial state of cube?

[Vryl]

I am pretty sure that there is a proof that any configuration of a cube is at most 21 moves away from any other. That is also to say that you can solve any position in 21 moves or less. It is known, I believe, as 'Gods Algorithm'.

So, 25 random moves is at least 4 that are redundant, ie, don't add to the complexity of the puzzle.

I am forshamed

[wizarddc]

I can do the cube, and have been able to since like 8th grade. When I first solved it, it would take me like a half hour. Then in 10th grade, I found a solution online that allowed me to do it in 10 minutes. And then my senior year, I found a 3 minute solution. After a few more years of using that method, I can regularly get a cube solved in a minute and a half. I know there are quicker algorithms out there, but I have dedicated the time for them. Most people have the patience and intrigue to stick around for a couple of minutes while I solve it. They always say "You can do that so fast!". I'ts like someone telling you your Ford Taurus is fast when they've never seen a Corvette. I'm always like "I don't do this fast, trust me". I always thought the record was less than 20 seconds. I remember seeing a guy do it in 16 seconds before. Maybe this is the average of a few cubes, I guess.

I think this is a real problem

[JMZero]

I'm no expert on the cube (it takes me a couple minutes), but it seems to me the competition would be very much influenced by "free stuff" (ie. steps in solving that you don't have to do). From a solving perspective, there's all sorts of situations you can run into and they are not equally favorable.

It's also pointless to standardize the number of turns used in randomization. Beyond a trivial number of turns (say, about 10 - maybe 15 for a pro), the cube is randomized enough that you'll be "starting from scratch" (ie, it's likely your first turns will be moving it further from solved, but towards a situation you can deal with).

I've seen someone solve a cube in 13 seconds. They were moving fast, but they also admitted they got lucky.

Just a question of time and monkeys

[janaagaard]

Given enough time and enough monkeys they will eventually solve the cube in 19 seconds.