So recently I was at a garage sale, and I decided to get a Rubik's cube... I've always been bad at it, but hey it'd be neat if I could solve it in like 20 seconds. So I bought one there, brand new, and so far I've managed to get it in 35 seconds, using a mix fairly mostly intermediate strategy.
There are 43 quintillion permutations of the rubik's cube, so having an algorithm to solve one from every starting position is not really similar. Most solutions to the Rubik's cube are rather similar, from the beginner level to the speedcubing WR of 4.9 seconds, which is the done with the method discussed last. Just to touch on the methods done by 99% of people who cube:
Beginner E-C 1. White cross - intuitive process, can always be done within 8 moves, usually 6. Assuming your top starting face is white, you need to get the edge pieces of that face in correct order CCW (blue-orange-green-red). 2. White corners - now simply slot the white corners into the correct spot, ie. if corner piece has a white-blue-red, and then you want to slot it between the blue and red edge. It is a 3-move intuitive process, the only trick is if you have a piece on the bottom of the cube you need to do a little special trick to get it to the bottom row, and if you have it in the top row, you must do the procedure once, to get it back to the bottom row. 3. Second layer - Now up to step two, must people can get through with intuition and some playing with the cube, here is where it becomes difficult. To complete the 2nd layer, you need to turn the cube such that the white face is turning the bottom, and rotate the cube until you line up a column of 3 of the same color in any center row. Then you do 1 of 2 8-move algorithms to get it to move to the left or right. This is done four times to get all the 2nd layer edge pieces into their right places. Only trick here is that sometimes your pieces will be where you're trying to get them, just in wrong orientation, and in that case, you need to perform the algorithm twice, once to prop the piece out, and then to put it back in. 4. Last layer cross - Now we want to get a cross on the face opposing the white face (yellow face). In this scenario, you'll see one of 4 things, one of them means it's completed, 2 of them you need to perform one of two 6-move algorithms, and one you need to do both algorithms together. 5. Last layer edges - The goal here is to orient the yellow edges you now have facing up with the correct side faces. This involves 1 8-move algorithm, depends on how your edge pieces need to move, the algorithm moves two pieces that are next to each other, but occasionally, you need to move the two opposing sides, hence apply twice. 6. Last layer corner position - 8-move algorithm that you perform once or twice to switch positions of all the corners depending on their orientation. The 3rd possible permutation is that they are correctly oriented, and then you skip the step. 7. Last layer corner orientation - It's a 4 move algorithm that you perform 2x if the corner cube needs to be rotated CCW, and 4x if the corner cube needs to be rotated CW. Kind of a confusing step since it looks like your whole cube is being destroyed. Repeat for all corners.
Beginner C-E (I used this one) 1. White cross (same as E-C) 2. White corners (same as E-C) 3. Second layer (same as E-C) 4. Last layer cross (same as E-C) 5. Solid top layer - Here is where it gets different. Compared to the E-C method, instead of correctly orienting the edges, you get all the yellow corner pieces to point upwards instead, and not worry about the edges being in correct locations just yet. There is one of 4 permutations you will see here (depending on how many yellow corners are pointed up [0,1,2,4]). The algorithm you perform here is a 8-step algorithm you perform for every case, the only difference is you need to hold the cube differently for the different cases, and perform the algorithm 1-3 times depending on what you have. 6. Last layer corner position - You can now hopefully see why this is called the C-E, because we do the corners before the edges. Here you have to perform a beastly 13-move algorithm, which swaps two corners which are next to each other. You'll always have at least two corners in the correct position, and hence if the two correct corners are opposite from each other, apply this algorithm twice. 7. Last layer edge position - Now all that is left is the edges. Either you'll have 0,1,4 edges in the right spot. If you have 4, voila you're done. If you have 1 edge, you look whether the other 3 need to be moved one over in CW or CCW orientation, based on that you perform one of 2 12-move algorithms, fortunately they're almost identical, and the only difference between the CW and CCW one is whether your rotate the top face CW or CCW on two occasions. If you have zero correct edges, just perform either one, and you will get to the one correct edge case, and them to the appropriate algorithm there.
Intermediate 2-look (my current one) 1. White cross - only difference here is that you hold the white cube white face down, so you can identify everything quicker. The nice thing about the 2-look system is it uses algorithms in the full CFOP, the method used by the WR holder, and by most speedcubers in the world. 2. F2L - First two layers. Do it with intuition they say, but man, this is tough. The idea here is that combine the white corners and second layer steps together into one step. This means that you need to create doubles of the corner and 2nd layer piece and slide them in without destroying the cube. For the intermediate method, the recommendation is to learn it by intuition and memorize 5 algorithms for the more difficult cases. 3. EOLL (orient last layer)- Essentially the last layer cross from the beginner guides, but it adds one more algorithm to remove the one case where you need to perform two separate algorithms. So 3 algorithms for this step. 4. OCLL - Simply step 5 from the C-E guide, except instead of having to do that 1 algorithm 1-3 times, you learn 7 algorithms based on the orientation of the top face to do it in 1 move every time. 5. CPLL (permute last layer) - Same as step 6 of C-E guide, except adds a 2nd algorithm for the case where you'd need to perform two as in the beginner guide. Algorithms chosen are quicker (based on hand positioning) but more difficult to learn. 6. EPLL - Same idea as step 7 of C-E guide, except now instead of only having 2 algorithms for CW and CCW rotation, we also add 2 more algorithms that can deal with the two different cases of when no edges match-up.
Advanced Full CFOP 1. Cross (yes even this started with the white cross) 2. F2L - memorize even more of the 41 in best orientation, since intuition usually doesn't lead to the best algorithm for longer series. 3. OLL - Here we combine the EOLL and OCLL into one move. Naturally there is a lot of different permutations, so instead of having 3 algorithms and then 7 algorithms, we have a grand total of 57. This is the most difficult part of speedcubing to learn. 4. PLL - Once again, we combine CPLL and EPLL together, instead of doing them separately. From 2 algorithms followed by 4 algorithms, we're up to 21 algorithms.
So, what's your method, and how quick are you? Also no shame in having to use a guide to solve a cube, more than 99.9% who are able to solve the Rubik's cube could not have done it unaided.
edit: My gooood, I mispelled Rubik's cube in the title. Mods please fix
On September 17 2016 06:55 YokoKano wrote: there are videos of lego robots on youtube capable of solving a cube in less than 10 seconds, competing with most world record times.
Yes, Rubik's cubes are very algorithm based, so that is natural.
Then again, any Chess player with standard time controls would be beaten by Stockfish 7 running on a modern i7 with 1 second per move time limit.
It's a nice exercise in memory, dexterity, and intuition.
Solving the cube using instructions someone gave you seems really boring and I don't understand why someone would do it, it´s almost like work
I once got hold of a cube someone had a didn't know how to solve and I started trying to solve it, then I bought my own and ofter some time (I don't really remember) I created all the moves needed - I have no idea if those were anyhow efficient, but they worked and I was able to solve a given cube under 3 minutes. That's about where I stopped caring about it, because it was a solved problem for me
What a really admire is people who just look at the cube, make ten moves and it is suddenly miraculously done - all my techniques are just step-by-step because I just do not see any global structure in the cube positions.
I've spent a lot of time cubing a couple of years ago thanks to this thread. Great guide with pictures, in depth explanation and different methods to solving.
Back then I was working a callcenter job doing pointless surveys for minimum wage. I mostly used a cheat sheet for the two algorithms but I managed to do it without once of twice.
On September 18 2016 16:24 opisska wrote: Solving the cube using instructions someone gave you seems really boring and I don't understand why someone would do it, it´s almost like work
I once got hold of a cube someone had a didn't know how to solve and I started trying to solve it, then I bought my own and ofter some time (I don't really remember) I created all the moves needed - I have no idea if those were anyhow efficient, but they worked and I was able to solve a given cube under 3 minutes. That's about where I stopped caring about it, because it was a solved problem for me
What a really admire is people who just look at the cube, make ten moves and it is suddenly miraculously done - all my techniques are just step-by-step because I just do not see any global structure in the cube positions.
That's very impressive Opisska, figuring it out on your own is definitely quite the feat.
The fewest moves you can use to solve any Rubik's cube is 20, and that was achieved through simplifying the problem a lot through symmetry and some other mathematical tricks, and lots of computational time. Computer algorithms can find a solution in 21 moves or less rather quickly with the code here:
I don't think anyone comes close to doing that by hand though, the CFOP is the fastest method, and that requires:
6-8 moves for cross 4x8~ for F2L 8-12 for OLL 10-12 for PLL
Petrus is a tad more move efficient, but I don't know much about it. So yes, they still use all the algorithms, they just have some very good foresight, which too is also impressive, similar in vain to playing chess blindfolded.
On September 18 2016 16:24 opisska wrote: Solving the cube using instructions someone gave you seems really boring and I don't understand why someone would do it, it´s almost like work
I once got hold of a cube someone had a didn't know how to solve and I started trying to solve it, then I bought my own and ofter some time (I don't really remember) I created all the moves needed - I have no idea if those were anyhow efficient, but they worked and I was able to solve a given cube under 3 minutes. That's about where I stopped caring about it, because it was a solved problem for me
What a really admire is people who just look at the cube, make ten moves and it is suddenly miraculously done - all my techniques are just step-by-step because I just do not see any global structure in the cube positions.
That's very impressive Opisska, figuring it out on your own is definitely quite the feat.
The fewest moves you can use to solve any Rubik's cube is 20, and that was achieved through simplifying the problem a lot through symmetry and some other mathematical tricks, and lots of computational time. Computer algorithms can find a solution in 21 moves or less rather quickly with the code here:
I don't think anyone comes close to doing that by hand though, the CFOP is the fastest method, and that requires:
6-8 moves for cross 4x8~ for F2L 8-12 for OLL 10-12 for PLL
Petrus is a tad more move efficient, but I don't know much about it. So yes, they still use all the algorithms, they just have some very good foresight, which too is also impressive, similar in vain to playing chess blindfolded.
Thanks, I wasn't trying to brag It never really felt that difficult and I think the reason is that my moves are probably very inefficient but very easy to learn. Essentially every move after the cross is "remove this already settled piece, change something not touching the piece and put the piece back" and the moves for the last 4 pieces are recursive variations of those ... it may seem daunting, but it's actually pretty easy to remember just a couple of operations and then pile the upon each other, the only problem was when I got distracted in the middle of a sequence and lost what I was doing, it ment starting from zero.
So even the fastest human solvers use step-by-step algorithm? To me it always seemed like they are doing the whole thing at once, but that's probably just my complete lack of understanding outside of my box of my system.
On September 18 2016 16:24 opisska wrote: Solving the cube using instructions someone gave you seems really boring and I don't understand why someone would do it, it´s almost like work
I once got hold of a cube someone had a didn't know how to solve and I started trying to solve it, then I bought my own and ofter some time (I don't really remember) I created all the moves needed - I have no idea if those were anyhow efficient, but they worked and I was able to solve a given cube under 3 minutes. That's about where I stopped caring about it, because it was a solved problem for me
What a really admire is people who just look at the cube, make ten moves and it is suddenly miraculously done - all my techniques are just step-by-step because I just do not see any global structure in the cube positions.
That's very impressive Opisska, figuring it out on your own is definitely quite the feat.
The fewest moves you can use to solve any Rubik's cube is 20, and that was achieved through simplifying the problem a lot through symmetry and some other mathematical tricks, and lots of computational time. Computer algorithms can find a solution in 21 moves or less rather quickly with the code here:
I don't think anyone comes close to doing that by hand though, the CFOP is the fastest method, and that requires:
6-8 moves for cross 4x8~ for F2L 8-12 for OLL 10-12 for PLL
Petrus is a tad more move efficient, but I don't know much about it. So yes, they still use all the algorithms, they just have some very good foresight, which too is also impressive, similar in vain to playing chess blindfolded.
Thanks, I wasn't trying to brag It never really felt that difficult and I think the reason is that my moves are probably very inefficient but very easy to learn. Essentially every move after the cross is "remove this already settled piece, change something not touching the piece and put the piece back" and the moves for the last 4 pieces are recursive variations of those ... it may seem daunting, but it's actually pretty easy to remember just a couple of operations and then pile the upon each other, the only problem was when I got distracted in the middle of a sequence and lost what I was doing, it ment starting from zero.
So even the fastest human solvers use step-by-step algorithm? To me it always seemed like they are doing the whole thing at once, but that's probably just my complete lack of understanding outside of my box of my system.
Yeah, maybe now that I've used the cube some, and I've been practicing the F2L, which has a lot of moves based on intuition, I could do it... But it was very hard for me to figure out how to move something without completely breaking everything else right when I started. A lot of the algorithms for the OLL/PLL are not intuitive at all. Right as I got the cube, I tried to get as far as I could without intuition, so the first face is as far as most people can get, and then I played with it for a bit, and figured out how to complete the second layer, but afterwards you mess the cube up so much in the process of moving something, so you need to have the foresight which I didn't have (and still don't have lol).
And yep, the method briefly described in my last post is easily used by 95%+ of top cubers. It's very algorithm heavy, and hence it's very quick if you put in the effort to memorize, since you don't have to do much thinking or inspecting, you just look for a couple key spots the blocks are in, and boom, it's all instant. The cross they solve simply by inspecting the cube, and F2L uses some intuition, but most of the time, they are just inspecting the cube and not thinking per say.
CFOP: 4.9 second 1 trial, 7~ second average Petrus: 12 seconds 1 trial, 13-14~ second average ZZ: 9 seconds 1 trial, 10-11 second average Roux: 3rd fastest cuber in the world uses Roux right now, pretty quick.
Though it's not necessarily a fair comparison, since so many people haven't tried the other methods since they just jump on the CFOP is best, so no reason to bother learning anything else, or at least dedicate a lot of time to get quick.
On September 18 2016 19:12 Branch.AUT wrote: I've spent a lot of time cubing a couple of years ago thanks to this thread. Great guide with pictures, in depth explanation and different methods to solving.
Back then I was working a callcenter job doing pointless surveys for minimum wage. I mostly used a cheat sheet for the two algorithms but I managed to do it without once of twice.
Ah, neat that someone did a nice article for it on TL.
Yeah, that's the Beginner C-E method I described but he flipped the order of step 5 and step 6, not really sure why since it makes inspection time longer, but it's more or less the same algorithms.
Petrus is supposedly a fun way to solve a cube because it's so move efficient, but also very thinking and spatial awareness heavy.
I have a fairly worn down five Row Rubiks cube that i do sometimes to occupy my hands. I also have 2, 3 and 4 row ones i barely ever do. Then i have a couple crazy ones like this http://www.elllisblog.de/wp-content/uploads/2012/11/DSC06495-Erläuterung1.jpg , but i rarely do these because it involves a tad bit more thinking ^^
yea but as you say that's the point of the puzzle. it's a 3x3 cube that takes quite a long time to solve mechanistically speaking 0.9s is quite a long time. the world is roughly not a 3x3 cube and doesn't mechanistically resolve to any kind of intentional point despite appearances that it might in fact be acting like this.
On September 20 2016 16:12 Cascade wrote: I tried solving a 2x2x2 cube for a good 15 minutes. I completely failed, which was pretty frustrating.
The pain, lol.
Don't worry, 15 minutes isn't that much time to put in, I've never done a 2x2x2, but looking at the algorithms for it, it doesn't look that much easier in terms of that last few steps and algorithm complexity. The steps seem long enough that the average person would struggle finding them without sufficient time put in.
On September 20 2016 17:19 riotjune wrote: Solved one of these when I was 3 by peeling off the stickers and sticking them back on, takes a while tho.
These things come with step-by-step instructions on how to solve it now, don't even need the internet.
The 2v2v2 is just a 3v3v3 without middle parts, so as a matter of fact you just need like half the steps you would need for a 3v3v3. It's still challenging if you have no idea in which direction you are going though, i agree with that. Once you know how to solve a 5v5v5 you are effectively able to solve cubes of whichever size, you have all the needed steps at hand.
EPT
Also no shame in having to use a guide to solve a cube, more than 99.9% who are able to solve the Rubik's cube could not have done it unaided.
