Solving a Rubik's Cube: Tips & Tricks

[hover]

Just an interesting question but has anybody here ever solved a Rubik's cube before? I need some serious help with solving one and it is driving me nuts! Anyone have some good tips?

Thanks!

 

[Cyrus]

Snap appart the cubes, then re-assemble them by snapping them back together.

 

[Moo Of Doom]

I just learned a few weeks ago. Stumbled upon the last transformation needed pretty much by accident, too, while trying to apply group theory to the thing.

Tips are hard to give out online. I would try to find some move sequences that change some of the stones without affecting most of the others.

 

[gabee]

I've been solving one and I've actually accomplished a bit--first two layers done. Now I have the last layer to solve (which will probably be the most difficult, since I don't want to mess the others up!).

I saw a guide online and picked up a few tips. The key to solving the cube is not to get "colors" in the right place...you can have one single-colored face "solved" but the pieces can be in completely wrong places. So, you have to get the pieces in the right places. (Example: The edge piece with red and yellow on it must be on the edge between the red and yellow faces.)

Then you solve in layers. Pick a side to solve--say you choose yellow. First, you will solve the edge pieces on the chosen face (the pieces with only two colors). When you are done with this, your chosen face will look like a cross of yellow, and above each side's center piece will be that face's color (except for the bottom of course). Then solve the corner pieces for your chosen face. After that you'll have the whole first layer solved. Then, solve the edge pieces for the middle layer (this will be a little harder). I don't know how to solve the bottom, but it seems the "moves" get more complex as you progress to each subsequent layer.

I learned that while you are solving the first layer, it is not necessary to keep the other layers in order, so you could use the bottom layer to "transport" pieces from one place to another to get them in the correct places. The side pieces on the second layer are a little harder but you can figure out the required moves with a little thought. The third layer I haven't done yet. ;) Good luck!

 

[hover]

When you are done with this, your chosen face will look like a cross of yellow, and above each side's center piece will be that face's color (except for the bottom of course). Then solve the corner pieces for your chosen face. After that you'll have the whole first layer solved. Then, solve the edge pieces for the middle layer (this will be a little harder).

Thats really where I'm having trouble. I can get the top side done and the first layer but i can't seem to position the edge pieces in the middle.

And just think, once I'm done with this i can do one of http://www.geocities.com/jaapsch/puzzles/cube5.htm" ...

 

[gabee]

I can't remember exactly, but I think the way I solved the middle layer was by doing things like L2, U2, R, U2, L2. I think this would work if you can position the piece on the right on the bottom face so that if you were do R the piece would be in the correct position on the front face (of course this would mess up the top face). So, you would do L2, U2, R, U2, L2 and it wouldn't mess up any face. I also think I was able to solve the middle edge pieces by doing opposites one after another...that is, looking down on the middle layer from above, I solved the bottom right edge piece then top left, then bottom left then top right. The last piece was harder to get in the right spot without messing the others up and to tell you the truth, I got it by accident :X!

I hope this helps!

 

[dontdisturbmycircles]

Someone here has a guide to solving these in their signature... I forget who though. :(

 

[hover]

I can't remember exactly, but I think the way I solved the middle layer was by doing things like L2, U2, R, U2, L2. I think this would work if you can position the piece on the right on the bottom face so that if you were do R the piece would be in the correct position on the front face (of course this would mess up the top face). So, you would do L2, U2, R, U2, L2 and it wouldn't mess up any face. I also think I was able to solve the middle edge pieces by doing opposites one after another...that is, looking down on the middle layer from above, I solved the bottom right edge piece then top left, then bottom left then top right. The last piece was harder to get in the right spot without messing the others up and to tell you the truth, I got it by accident :X!

I hope this helps!

Actually before even reading your post i managed to figure out how to place the middle pieces just by playing with it. I just have the last layer to deal with. By the way the bottom face to the original cross doesn't have to be a cross, right??

 

[grant9076]

The way I taught myself (which I think is the easiest), is to solve the 8 corner pieces first. This will allow you to solve the 12 edges without messing up the corners. If try to solve the edges first, then solving the corners will wind up messing up the edges. I hope this helps.

 

[DyslexicHobo]

I can solve the rubix cube in ~45 seconds. If anyone wants help, send me a PM. I can try and give you some hints.

My junior year of high school (last year) was very boring, so I decided to learn how to solve a rubix cube! I succeeded. :) Obviously, my faster times are from memorization, then applying the methods I learned from previous solves.

By the way, my method is the layer method. Solve the bottom layer, then middle layer, then top layer.

As someone already mentioned, it's pointless to go by sides. Instead of thinking of the cube having 6 sides, you really need to think of it as 26 blocks that need to be arranged in the correct order and orientation. Even if one side is the same color, the blocks might not be arranged correctly to allow the other sides to be solved.

As for the 8 corners -> sides method that the person above me mentioned, I haven't thought about it like that. I'm going to go mess around with that now. ;)

 

[Integral]

Many years ago I was shown the 2 moves needed to solve the Rubik's cube. One "move" rotated a single corner leaving the rest unchanged. The other, much simpler move, rotated a edge piece.

However, the corner rotation "move" was quite involved requiring between 20 and 30 individual rotations. I have long ago forgotten the sequence.

 

[MeJennifer]

I can solve the rubix cube in ~45 seconds.

That was my average as well when I used to play with that.

By the way, my method is the layer method. Solve the bottom layer, then middle layer, then top layer.

I do it a bit different:

First I position the bottom layer middle blocks, then I position three middle layer middle blocks. I leave one open to use as a "passing path" for bottom layer corner blocks to use faster formulas for positioning the bottom layer corner blocks. Then I position the bottom layer corner blocks. Then, with a single formula, I make the top one color without caring for positioning, with solving the last middle layer at the same time, if possible. Then I reposition the top layer blocks, usually with one or two formulas.
Also, I always start with the same colors, that is theoretically not the best approach, but that way I know where everything must go and I hardly have to look.

 

[||spoon||]

They are really easy once you get the hang of it.

I solve a cross of one of the colours first, making sure taht the edge pieces are the correct ones by checking against the middle square of the adjacent sides. Then fill in the corners of teh side with the cross accept for one so that you can move back and forth without wrecking the rest of your side when trying to complete the rest of the cube.

THen you solve the edge pieces accept for one aswell... line up the unsolved edge and the unsolved corner to get to make a cross on the top... it is too hard to explain fully, but if someone else knows this methiod they should be able to recognise it from that...

 

[t!m]

A third on the ~45 sec average. My best was a bit above 30 and I haven't been practicing lately, so it's gotten closer to about 1 minute. I first solve a 2x2x2, then 2x2x3, then what's called 'fixing edges', followed by completion of the first two layers (i.e. 2x3x3). LL is solved by first appropriately placing the corners, then orienting the corners, then correcting the edges. The F2L are mostly intuition at this point, but the LL is always memorized algorithms for me, depending on the state of cube at that point. I probably have only around 6 algorithms memorized + reflections.

 

[Integral]

Snap appart the cubes, then re-assemble them by snapping them back together.

Take a randomized cube, snap out a single piece, rotate it and replace it. Now give it to one of these 45sec whizes and watch them go crazy!

 

[ziad1985]

while trying to apply group theory to the thing

Mathematicians!

 

[Mk]

Ah, always check the wiki. They even have a textbook on it!

http://en.wikibooks.org/wiki/How_to_solve_the_Rubik's_Cube

 

[DyslexicHobo]

Take a randomized cube, snap out a single piece, rotate it and replace it. Now give it to one of these 45sec whizes and watch them go crazy!

Hah, I did that once just to mess around with it; later on when I tried to solve it and had forgotten that I changed a single cube, I was really confused! >.<

 

[BobG]

I haven't played with one in years so I don't remember any of the moves, but...

Instead of solving the whole cube, learn how to swap a side piece with the opposite side piece and how to swap a side piece with an adjacent side piece.

The corners can usually be set up just by playing with the cube. I think I learned at least one formula for swapping corner pieces (that was the easiest swap I learned), but you can struggle by without one.

If you don't know how to swap two side pieces, the last layer is just frustration.

 

[Danger]

I can solve the rubix cube in ~45 seconds.

I have a really good booklet that shows how to do it, and it still takes me over 45 minutes following that. My sense of spatial orientations absolutely sucks.

 

[setAI]

green first (^__-)

 

[Moonbear]

Let's see, when I was a kid, I first solved it by popping out all the pieces and then popping them back in in the right patterns. Then I got a book and followed the instructions in that, then memorized the book.

 

[grant9076]

Take a randomized cube, snap out a single piece, rotate it and replace it. Now give it to one of these 45sec whizes and watch them go crazy!

Hah! That only works on those 45 sec fools. Us true 80 sec people will take less than 2 minutes to figure out that it is an impossible combination.

 

[larkspur]

My first cube was one from the first year they hit the US. It turned and twisted so smoothly and effortlessly. It was probably thrown away when I went off to college. These new ones don't turn easily like that one did. My sister and I used to race each other to see who could solve it the fastest.

 

[turbo]

I got to the point that I could solve any cube in less than 90 seconds, then used to bet for beers at bars that I could solve it in under 2 minutes. I would take plenty of time (5-10 minutes) to solve the cube initially to make sure someone hadn't screwed with a segment, then when some bar-fly saw that I had solved the cube and wanted to challenge me, I was ready for it. I'd let him set the time limit, and it was usually well over my two-minute drop-dead time. I'd let him twist and turn away to his heart's content, then hand it back finished in plenty of time. I have to agree with Larkspur that the original puzzles were fast and smooth - the inevitable knock-offs would sometimes bind and catch and slow you down.

 

[light_bulb]

each middle cube of each side will always be one color, changing stickers on two middles wrecks it, give someone that then place the bet. lol

 

[Donski]

I got bored with that 3x3x3, so I bought a 4x4x4 called a Rubik's revenge. After that I bought a Rubik's Professor that's a 5x5x5. I was never able to do the regular one in 45 secs, it takes me around 3 min. It takes me around 15 min to do the 4x4x4 and 30 min to do the 5x5x5. No one ever taught me how to do them, I figured them out on my own. I guess that's why it takes me so long to do it, I put one or two pieces in at a time without algorithms. It maybe hard to believe, but with the proper algorithms you are never more than 10 moves from solving the cube.

When I do the regular cube I start with getting one layer complete, usually the white. Then I do the corners giving me an X on all sides. To swap two corner pieces is an 11 step move. Then I do the sides by getting them on the proper axis, so it doesn't matter if it's red or orange just as long as they're opposite each other. The main move for that is crank the center & spin the top. You need to orient the pieces you want to swap so they're across from each other before you do the move, then remember how to get back. After that it's easy to swap them straight across.

When doing the 4 & 5 cubes, you start by getting the center pieces. After that it's pretty much the same as the 3x3x3. Again you can mix your red & orange, blue & green or yellow & white. There are also a few different variations on crank the center and spin the top to get the different side and center pieces. But there's one move where you may have to swap just two side pieces which involves cranking one of the centers and spinning one of the sides, it messes it up a little and you have to solve part of it over again.

Square 1 is also a good one. You normally have to get it out of square to get the last couple pieces.

 

[MeJennifer]

It maybe hard to believe, but with the proper algorithms you are never more than 10 moves from solving the cube.

Do you have a reference for that? As far as I know the minimum number of moves from any initial position has never been mathematically proven.
According to Wolfram MathWorld the current best proven algoritm is 29 turns (or 42 "quarter-turns") by Michael Reid (1995).

 

[apples]

Take the stickers off, and put them back on again correctly.

 

[Donski]

Do you have a reference for that? As far as I know the minimum number of moves from any initial position has never been mathematically proven.
According to Wolfram MathWorld the current best proven algoritm is 29 turns (or 42 "quarter-turns") by Michael Reid (1995).

I may have been mistaken and it was 20 and not 10, but that's not measured in quarter turns.
http://www.seas.ucla.edu/hsseas/press/1997/korfcube.html

I remember finding a site that let you input the positions of all the colors and it would show you move by move how to solve it. I can't seem to find that site, I wish I had bookmarked it. But I know it was less than 29.

 

[BobG]

I got bored with that 3x3x3, so I bought a 4x4x4 called a Rubik's revenge. After that I bought a Rubik's Professor that's a 5x5x5. I was never able to do the regular one in 45 secs, it takes me around 3 min. It takes me around 15 min to do the 4x4x4 and 30 min to do the 5x5x5. No one ever taught me how to do them, I figured them out on my own. I guess that's why it takes me so long to do it, I put one or two pieces in at a time without algorithms. It maybe hard to believe, but with the proper algorithms you are never more than 10 moves from solving the cube.

Why keep going bigger? You really need to try more dimensions, instead.

Here's a four-dimensional cube created in Java: Magic Cube 4D. Only 58 people have ever solved it.

After that, you can go for the 5-dimensional cube: Magic Cube 5D Seven people have solved this one.

Of course, the most challenging is trying to set the record for most cubes solved underwater with a single breath. Underwater record The record is 6 cubes. Second place is 4 cubes.

This quote from someone who solve 2 cubes, plus a 2x2 cube "My best friend and I would see who could swim the farthest and hold their breath the longest on the swim team. I never thought it would come in handy." He thinks he could solve 3 cubes, but could only procure 2 - don't they have telethons to help people like this? (help them obtain another cube, I mean).

Another quote from another underwater solver: "This was done in a jacuzzi...it got pretty hot down there. The air bubbles were hard to see past too. I had to stop after a few solves because I got dizzy, I wouldn't suggest jacuzzi solving."

Edit: I'm having second thoughts about having posted this. Traffic on PF will go way down, sending the value of the PF website down into the six-figure range for sure. Greg's going to have me banned.

 

[edward]

Here is a simpler version.

http://www.wrongway.org/cube/solve.html

 

[hover]

Hey everyone! So happy for the help you gave me. I can do a rubiks cube in about 2 to 3 minutes now(not very fast compared to other but anyway). I have a new question now. How can someone solve a rubiks cube blindfolded and could i do it?

Thanks

 

[light_bulb]

"Edit: I'm having second thoughts about having posted this. Traffic on PF will go way down, sending the value of the PF website down into the six-figure range for sure. Greg's going to have me banned."

 

[tiny-tim]

free 3D program

Ah, always check the wiki. They even have a textbook on it!

http://en.wikibooks.org/wiki/How_to_solve_the_Rubik's_Cube

There's a free 3D program (runs under Mac OS 10.2.4+ and Windows) downloadable from Tom Davis's http://www.geometer.org/rubik/index.html" which allows you to rotate a Rubik's cube yourself, or which solves any configuration you set up.

 

[hover]

Ah yes I remember this thread. Interesting to see how many views its gotten and how much I have improved solving since.

 

[Monocerotis]

http://www.teamliquid.net/forum/viewmessage.php?topic_id=95473

if you have the time, read through it and you'll be able to solve a rubik's cube.

Then all you need to do is get down on your luck and have your wife leave you and be all alone with your 2 year old son and pursue a career in banking/investment and end up in the car of a man who's hiring and solve the cube by the time the cab takes him to his home thereby impressing him so much he'll offer you an opportunity to be hired.

Also, you need to be will smith.

 

[ideasrule]

A few months ago, I decided to try solving a rubik's cube without reading about any algorithms from the net. I spent 30 minutes doing what Euclid did: identifying axioms & building on them. I got some helpful conclusions, but unfortunately got bored and didn't bother continuing.

If anybody here can honestly say they've no idea how to solve a cube but manages to come up with a way, tell me and I'll acknowledge you as a genius. :)

 

[hover]

A few months ago, I decided to try solving a rubik's cube without reading about any algorithms from the net. I spent 30 minutes doing what Euclid did: identifying axioms & building on them. I got some helpful conclusions, but unfortunately got bored and didn't bother continuing.

If anybody here can honestly say they've no idea how to solve a cube but manages to come up with a way, tell me and I'll acknowledge you as a genius. :)

Naw it really doesn't take any smarts to learn how to solve a rubik's cube. Somebody just needs to teach you how to do it and if your patient and listen you will learn. You can always try youtube. Some great videos there.

Leave genius for the people who have solved fundamental problems in physics.

*cough* Einstein *cough* :P

 

[bat Dancho]

I have a solution for 2x2 and 3x3 cubes, which I figured out myself. Memorizing somebody else solution is no use, no fun. Here is how it works. I am not a mathematician physicist, just el enginer, so my explanation may sound moronic , but it is unic in its way because it take smallest amount of memorization ( but the most time to solve). I will not get in details , but in generally it is like this. The top layers always is just common sense. If you get 2x2 the moves which really twist the gut of the cube and change something is the the move to get down one of the top corners one way and get it back up the other way. in formulas this mean L-, D-, L+ , back is F+,D+,F- . This moves again gets a corner form top puts it down and gets it back up, so the top side is same but it has efect on bottom. Efect is 2 corners in from swap places one stays and one is rotated 120 degree. With this knowlegde there is no configutation that you cannot put bottom corners in positions. If this procedure is made 2 times it will keep the bottom corners in position but turn 3 of them 120 degree, made 4 times will do 240 degree. Basicly there is no situation that with apropriate orientation execution this procedure needed times to solve 2x2. The 3x3 is done with same principle. when the "corner up-down" is made ip places mid level edges from 3 level to second ( actually what I am talking is a very popular formula for solving second layer, but use it over and over...)when this formula is used get down egde from second layer to 3-layer it has effect on swapping bottom edges , positioning bottom corners and finally rotating corners 120/240 gedree... It is bit time consuming but I believe it is the solution with least memorization . Any comments I will appreciate.

 

[╔(σ_σ)╝]

I average just over a minute (my cube sucks)

Best time : 47secs ( on a really good day. )

I use a modified beginners method. I developed a few of my own algorithms which took my time from 2mins 30secs to 1min 10 secs.

I am trying to learn the CFOP (Cross - F2L - OLL - PLL) system.

 

[bat Dancho]

Just for the clarity of my last message I feel I need to spell it out. The only formula which one have to know for 2x2x2 cube is R-D+R+D+F+D-F- This formula executed once swaps 2 front corners on bottom layer ( down side) and keeps back corners in same position , one of the back corners ( left back) is rotated 120 degree. When executed twice in sequence it rotates 3 corners 120 degree ( because it swaps back the 2 front corners). The right back corner is intact again. This is all you need for 2x2x2. For the 3x3x3 cube this formula is just illustration of the idea. If D+ ( so the D-) are doubbled it modifies the effect of the formula over the bottom layer , but no effect on 2 layer and 1 layer . All you need for 3x3x3 cube is 3 formulas for the bottom layer ( the doubbling modifications give you all 4 formulas) . This set of formulas is complete. It takes around 2 minutes to solve a 3x3x3 cube, but this method has the smallest count of formulas. As I know for speed solving you have set of 50+ formulas which gets your bottom side from any possible combination directly to solved state. This method is the complete opposite slowest but less memory involved ( as in PCs computational power is achieved trough speed or memory...)
---------------------------------------------------------------------
Cube sides Up, Down, Right, Left, Front, Back, where R- is Right side counterclockwise , D+ is Down side clockwise,

 

[bat Dancho]

One more. This same method works for what I call pentacube or megaminx. This is figure which have 12 facets , which are pentagons. It is quite more complicated with around 7-8 formulas involved, but in principle is the same. It is solved in layers again. The layesr besides the very bottom are just common sense ... for the last the above method works. Time it takes ? Time to develop the formulas 1 week . Time to solve 1 hour. ha-ha-ha! Good luck!

This is the pentacube / megaminx

https://www.amazon.com/dp/B003631BOM/?tag=pfamazon01-20

 

[mugaliens]

My best time was back when I was a freshman in college. About 1m 50s. Never did figure out how to solve it myself, so I bought a solutions book and memorized the steps.