So I was talking to a guy at church about the Rubik’s. He does it pretty fast but I’m faster, so he was interested in just learning how I solve it. So I wrote him an email but then I thought man I sure do blabber a lot, so I thought I’d turn it into a blog post that completely nobody will ever care about it.

So, if you ever want to know how I solve the Rubik’s cube, here you are. Some of the descriptions (especially at the beginning) are kind of lengthy, and in some (at the end after I realized how long this was going to be) aren’t

==========

I think that the 2 best sites / methods are

Lars Petrus – http://www.lar5.com/cube/

Jessica Fridrich – http://www.ws.binghamton.edu/fridrich/cube.html

Also – for cube syntax, see http://www.ws.binghamton.edu/fridrich/Mike/notation.html

That talks about how to describe various cubes and faces – like if there is a move that goes R U2 R’ that means turn the right face 90 degrees clockwise, then the top (up) face 180 degrees, and then the right face 90 degrees counter-clockwise

Both of them are original speed cubers from back in the 80s and have some good systems. The particular system that I use is sort of a combination of the 2 of them, hampered by the fact that I don’t really have the desire or the time (or at least I have not yet spent the time) to memorize all of the algorithms that they use. I have memorized a few of them, so sometimes I have to do 2 algorithms (taking twice as long for that part) because I don’t know the moves to just do it all at once. So, here’s a breakdown of what I do.

1. First I try to get a 2x2x2 block (so 3 middles, 3 edges and the corner between them). This is step 1 on Lars’ site. In “official” tournaments you get a 15 second inspection time before the timer starts, to look at the cube and plan what you’re going to do. I don’t really give myself a time limit but when I am looking at the cube I look for a corner / edge pair that are already together. I find that usually there exists such a pair around 80% of the time. If there isn’t one, then I look for a corner / edge pair that can be matched up with just one move, and there almost always is one of those. So once I get the c/e pair, then I try to match it up with another edge. You have 2 options here, so pick the best one. Example – if you’ve got the Red / Yellow edge next to the Red / Yellow / Green corner, then you are looking for either the R/G edge or the Y/G edge. Usually one of them will be in a better position than the other. So I find say the R/G edge, and put that next to the Red middle piece, and then slide the c/e pair next to it to match. At this point you’ll have a 2x2x1 block. So then you’re looking for the other edge (in this case Y/G), to put in place (next to the Yellow and Green centers). You’ll notice here that you actually DON’T want your 2x2x1 block lined up with the other middles just yet. In this example, now that you have the 4 red pieces in a 2x2x1 block, the ideal place to have it is not lined up with the Yellow and Green middles, but on their OPPOSITE faces (above the red / white / blue corner). The reason for that is that you want to be able to have the Y and G faces free to turn to get the Y/G edge in place. So get the Y/G edge in place, and then you turn the Red face to line it up and there is your 2x2x2 block. Lars’ site has a lot of good info and examples on this. As you become more familiar with this, you can actually find all the pieces you need as you are inspecting the cube, and start to plan out what moves you need to make to get your 2x2x2 block. If you’re *really* good (which I’m not usually) you can figure out the 5 or 6 moves you need to get the 2x2x2 block right from the start. You can usually get this block done in 7 moves or fewer. Alternatively, an easier way to start is to just get a cross (say the red middle and all 4 edges) – that’s the way that Fridrich does it.

2. So now that I have a 2x2x2 block, the next thing I do is pick a “bottom” color and get my other 2 edges on. Which reminds me – you should learn to solve either the cross or the block on the bottom – that way you don’t have to flip your cube around to solve the last layer. So since I have 3 colors in my block, I can look to see which 2 edges will go into place the easiest. Let’s say red – so I get the red / white and red / blue edges in place on the bottom. So my cube has red on the bottom, with all 4 edges in place, and one additional corner / edge pair (red / yellow / green in this example) also in place.

3. So now I switch over to the Fridrich method to get the other 3 c/e pairs (http://www.ws.binghamton.edu/fridrich/Mike/middle.html). Here’s the way I think of things. For now, ignore cases where a corner or an edge is incorrectly placed in one of the bottom 2 layers, or if they are together (next to each other) in the top layer but flipped incorrectly. If your c/e pairs are like that, you just need to get them to the position where they are both in the top layer but not next to each other. There are various ways to do that, but for now let me just share what to do if they are in that case

Before I talk about that, let me step back to talk about which sides you can turn. That’s the major downside to solving the entire first layer first. Once you get to that point, every move that you make messes up what you’ve already done, and so you are limited in what moves you can make, since you have to “fix” the first layer. With this you have several ways you can turn the cube. I usually keep my block in the Back left down (BLD) corner. At that point, you obviously have the entire U face you can turn without messing things up. Additionally you can turn the front and right faces 90 degrees either way and you’re still okay, since the c/e pairs that would be moving up to the top aren’t solved yet. Obviously as you solve c/e pairs you limit your available moves. For instance, if you solve the FL c/e pair (on my cube in this example that would be the R/W/G corner and the W/G edge) now you can only turn the F face counter-clockwise – if you turn it clockwise, then you’ll bring that RWG corner up to the top and mess things up.

Case 1 – The bottom color (red in this case) is on top of the corner (on the top layer). On my cube in this example, I have the R/W/B corner with red facing up. Move that corner to front / right. You don’t have to actually PHYSICALLY move it there if you can imagine what it would look like if it WAS moved there. Look at what color is facing front (blue). Find the blue/white edge. If blue is on top then you’ll need to turn a face counter-clockwise. If white is on top, then you’ll need to turn a face clockwise. What face to turn depends on which c/e pairs you’ve already solved – rotate the U face till you get to a face that won’t be disturbed by such a turn. In this example, I can turn the L face counterclockwise, since that will bring the FL c/e pair to the top, and they haven’t been solved yet. So turning L’ puts the B/W edge in the back-left-middle edge. Then I can bring the r/w/b corner over to that side, so it’s right above the edge and then turn L. Now I have the r/w/b c/e pair in place on the top. I’ll get to what to do with that in a sec.

Case 2 – The bottom color (red) is on the side on the top layer

Case 2a – The top color of the corner is the same as the top color of the edge

In this case, you’ll want to turn the face that has the red color counter-clockwise. Rotate the U face until you can make such a move without disturbing the c/e pairs that are already solved. In my example here, I have the r/y/b corner, with yellow on top. The y/b edge is also on the top layer (but not next to its corner) with the yellow edge also on top. So rotate the U face till r/y/b is ini the front right, which would bring the back right c/e pair (unsolved) to the top layer. Do a R’, then rotate the U face until the y/b edge is in the UF (top front) position, and bring back R clockwise to match up the y/b edges.

Case 2b – The top color of the corner is different from the top color of the edge

Again, you’ll want to turn the face that has the red color counter-clockwise. But here, you need to make sure that the edge is in the correct place. Look at the red face of the corner. On my cube, it’s the r/w/g face, in the U-F-R position. Red is facing front, and green is on top, and on the green / white edge, white is on top. If the 2 pieces are both on the top layer, and not adjacent, the w/g edge can only be in UL (up-left) or UB (up-back). When red is facing front, the edge needs to be in UL. If red is facing right, then the edge needs to be in UB. The way I think about it is that it has to be in the “next” possible edge position (besides the one RIGHT next to the corner). Once the corner and the edge are in that position, rotate the U face till the corner is directly above where it needs to go. In my example, r/w/g needs to be in the front left position. So I rotate U until r/w/g is in UFL, and since red is facing left, the “next” edge position is UB (up-back). Then move L’ U’ L and the c/e pair is not only moved together but in the appropriate position.

So in cases 1 and 2a you get the c/e pairs together on the top level. Also sometimes you get lucky and they’re already together on the top layer. In case 2b then part of the process of putting them together automatically moves them to where they need to be, but for the other cases, you have to get them from together in the top layer to where they need to go (in the first and 2nd layers). Once they’re together, take a look at the corner they need to go to. Let’s say you have the white/blue/red corner with the w/b edge. Depending on how they were put together, you’ll either need to turn the white face counterclockwise or the blue face clockwise (both times bringing the pieces that are where the r/w/b corner and b/w edge NEED to go up to the top layer. So look at your c/e pair. In my example here, the blue pieces are on top and the white pieces are on the side. Whichever color is on the SIDE (not the top) – that’s the face you need to turn. So here white’s on the side, so move your c/e pair out of the way (say to the UBR and UR position), and turn F’ U F and the c/e pair slides into place.

This part is definitely the trickiest part and the most difficult to really get the hang of. Just take it slow and try to figure out what you’re doing and WHY the pieces are moving where they are. The rest of it is just how much memorization of various algorithms do you want to do.

At this point you have the whole bottom 2 layers solved and just have the top layer remaining. First thing to do is orient the edges (i.e. put all the orange faces on top, without regard to what position they are in) A real speed solver would do this all in one step (there are 48 algorithms to memorize). I don’t know even close to that many, so what I do is I orient the edges first and then the corners.

To orient the edges there are 2 moves I use. If you have 2 edges already orange and they’re at UL and UR, then I do B L U L’ U’ B’. If the 2 edges are at UF and UR, then I do B U L U’ L’ B’ (which you’ll notice is the first move done backwards and opposite). If there are 4 edges already orange then obviously skip this step, and if there are 0 then you have to do both of those moves. Actually one of the things I’ve been working on lately is that depending on how you put in the c/e pairs in the previous step, you can usually make it so that you already have the 4 top edges oriented.

To orient the corners, you only need 2 moves, and they work if you already have one oriented corner. Lars talks about them in Step 6, and they’re called Sune and Anti-Sune. If you have 0 or 2 oriented corners then you just need to do Sune and/or anti-Sune twice. http://www.lar5.com/cube/fas6.html

At this point, the bottom 2 layers are done, and the top face is all oriented correctly, so you just have to move those top layer cubes around. There are 21 cases for this. One site I like for this is http://cubefreak.net/PLL.html

In looking at that page, I only know 8 of them, and that’s enough that you only ever need to do 2 of them to solve this last step. Sometimes I get lucky and get one that I know, other times I have to do one algorithm that puts it into a state where I know the algorithm to solve it.