Sexy method and improvements

[renyhp]

Hi everyone,

first of all let me say I'm a beginner and I've been cubing for only one week or so. I'm writing this post to tell you my impressions on some beginner methods and to ask other cubers' opinion.

Last week, while on YouTube, I stumbled upon some tutorial on how to solve a Rubik's cube (it was based on a classical beginner's method, similar to the famous badmephisto's one). I was curious so I watched the video then I got myself a Rubik's cube to try. I solved it a couple times then I realized there were too many algorithms to memorize and I would never have memorized them all, also because I don't have a good memory.

I was about to throw away my cube when I realized there were other tutorials with other methods. In particular, I was very attracted by a video that said you could solve the cube with only one algorithm (berthamilton's guide). After watching it, I was very satisfied by the fact that I wouldn't have to memorize a lot of algorithms and the only algorithm was a simple one (the sexy move R U R' U' and its mirror L' U' L U). Also the guide is very well done because it lets you understand what you are doing to the cube.

After some research I found it is some variant of the 8355 method, or Sexy method. Basically, after creating the cross, you:
1. Solve three bottom-layer corners:
• Place a white corner directly above their slot
• Use the sexy move (right-handed or left-handed depending on whether the corner is on the right or on the left) multiple times to insert it
• Leave always the unsolved corner in the front face, placing it on the right when you have to perform a right-handed sexy move, or on the left if the sexy move is left-handed
2. Solve three middle-layer edges (the three above the solved corners):
• Place an edge in the upper layer, so that it matches the face where it belongs (use the sexy move once if it's inside its slot)
• Take the cube so that that face is lateral and you face the edge's slot
• Use the sexy move twice to move the edge to its slot. Use a right-handed sexy move if the slot (and the edge piece) is on the right, or left-handed if it's on the left
3. Solve the upper layer edges using appropriate sexy moves (intuitively, based on the cycle that a sexy move makes edges do). The remaining edge will be automatically solved.
4. Solve the last 5 corners:
• Reverse the cube so that your bottom face becomes the top one and realign the top layer; take the cube so that you have the keyhole in your front right
• Perform multiple sexy moves using the keyhole as a "buffer": move only the bottom layer to place the buffer's slot directly below the keyhole, then use sexy moves to place the keyhole into its slot, and repeat this multiple times
• If you end up with a solved bottom layer but a scrambled cube, then keep performing sexy moves until you end up with only two adjacent corners unsolved; take the cube so that they go to the bottom layer and one of them is on front right, then repeat the previous step (6 sexy moves total will be required)

This, however, requires a lot of moves. In my research on how to reduce the number of moves, I've searched for other methods then I have realized there are some improvements I could do to this method.

1. In the first step, you will have to perform 3 sexy moves if the white corner is facing up, and 5 if it's facing in front of you.
• In the first case, you can use a shorter version (right handed is R U2 R' U') which makes the corner face right so that you can use one sexy move to solve it
• In the second case, a left-handed sexy move or an inverted sexy move (U R U' R') will do the trick
• If the white corner faces right, R U R' will be enough, sparing the U' move
2. In the second step, an Y-move (F R' F' R) or its inverse (R F' R' F) can be used, though I find it less immediate. Anyway, you can do better than the two required sexy moves by doing U' (R U R') (right-handed case)
3. This is the step with which I had the most problems. I've found it can be solved with two substeps:
• Create a cross in the upper layer (not correctly permuted): this can be achieved by F (R U R' U') F' or f (R U R' U') f' depending on whether the already oriented edges form a "---" shape or "r" shape, respectively (do both if none is oriented)
• Now permute the edges. Moving the upper layer, there is always a position where the front edge is solved and both the lateral edges aren't. If the back edge is not solved, then the unsolved edges have to be permuted in a clockwise or anticlockwise way. To permute them anticlockwise, place the keyhole on the right then do (R U R' U) R U2 R'; to permute them clockwise place the keyhole on the left and do the mirror version, ie (L' U' L U') L' U2 L. If the back edge is solved, do either (anticlockwise permutation) u' (anticlockwise permutation), or its mirror (clockwise permutation) u (clockwise permutation)

I'm pretty happy with this method because it requires only one algorithm and few variations of it, and I guess it's not much less optimized (in terms of number of moves) than the beginner's method, while as I said it requires a lot less algorithms to remember. In particular I find the keyhole very useful to solve the edges, though the 4th step is longer than the beginner's PLL: that's, I think, the price to pay for having less algorithms.

As for reducing the number of algorithms I've seen the Marshall method and some others but it seems to me that this method is the one that balances the most between optimizing the number of moves and reducing the number of algorithms.

Are my considerations right? What do you think about the improvements I propose to the sexy method? Can they be improved any further? Let me know your opinion, and happy cubing to everyone!

 

[cuber314159]

I would suggest that this method may be very easy to learn but not good for speedsolving so I would recommend you learn mats valks method in his tutorial, it is based mostly off the sexy move as well but is faster and easier to transition to cfop from I assume

 

[renyhp]

Well I don't know about transition to cfop (which I will probably do in a bit...), but to be honest it seems to me that my method is way faster. Using a keyhole lets my method be faster in solving the middle layer edges and in permuting the top layer edges; probably the only part where mine is slower is my last step, where I need to solve the keyhole together with the last layer corners, while he only has to permute the corners. Besides, some improvements I did to the basic sexy method can be done to his method: for example, you can solve the bottom layer corners faster than he does (by using R U2 R' U' or a left / inverted sexy move), or you can orient the last layer edges with f (sexy move) f' instead of two sexy moves...

 

[cuber314159]

Well I don't know about transition to cfop (which I will probably do in a bit...), but to be honest it seems to me that my method is way faster. Using a keyhole lets my method be faster in solving the middle layer edges and in permuting the top layer edges; probably the only part where mine is slower is my last step, where I need to solve the keyhole together with the last layer corners, while he only has to permute the corners. Besides, some improvements I did to the basic sexy method can be done to his method: for example, you can solve the bottom layer corners faster than he does (by using R U2 R' U' or a left / inverted sexy move), or you can orient the last layer edges with f (sexy move) f' instead of two sexy moves...

Yes but recognition would be harder solving five corners after a cross that is probably why people use CFOP, hasnt every world record since 2003 used CFOP. well done on just learning how to solve it and the knowing enough cube lingo to make such a long post of improvements to the method though. But I think it's the case of yes, you could be sub15 with this but with the same effort you could be sub10 with that

 

[renyhp]

Sure, I'm sorry I wasn't clear enough about this but I was just speaking about it as a beginner method, ie as a method where you solve one piece at a time, and comparing this method and the sexy method to Mats Valk's beginner's method or similar LBL beginner methods where you don't have the keyhole and you do OLL and PLL. It is of no doubt that CFOP is faster, that's why I wasn't comparing this method with it

EDIT: As for times... I'm nowhere near sub20, at the moment I'm averaging about 2:15 xD but it is also true I've bought a decent cube only since three days

 

[cuber314159]

What I'm saying is that eventually you will probably want to switch to CFOP or CFLOOPP and mats valks tutorial or youcandothecube is almost definitely faster than this and they are both beginner methods

 

[renyhp]

So you do say the "classic" beginner's method, with full LBL and (long) OLL and (long) PLL, is faster than mine? Why so?
My impression is that, to place middle layer edges, in that case you do way more moves while in this case you only do one sexy move. Same thing about orienting the top layer cross... while, yes, solving the last four corners in that case is faster than solving the last five corners in mine. So my impression is that those things kind of balance together, and basically the two methods are equally fast, while mine uses less algorithms. Am I wrong?

 

[xyzzy]

My impression is that, to place middle layer edges, in that case you do way more moves while in this case you only do one sexy move.

Yes, beginner's LBL is trash and using keyhole drastically improves it. In terms of move count, keyhole gets you something like 50% of the benefit of F2L over LBL, and you don't even need to learn any new algs.

Same thing about orienting the top layer cross... while, yes, solving the last four corners in that case is faster than solving the last five corners in mine. So my impression is that those things kind of balance together, and basically the two methods are equally fast, while mine uses less algorithms.

I don't entirely get what you're trying to say here. I'm not sure if 8355-style last five corners (L5C) has comparable speed to other beginner's LL methods (e.g. 4-look LL, with 2-look OLL and 2-look PLL), but it's clear that there's no easy way to incorporate algs into how you solve L5C to make it even faster (case recognition is awful, algs are awful, and there's a ton of cases), whereas you can upgrade 4LLL just by learning more OLL/PLL algs.

If you're looking for long-term improvement in speed, you should give standard 4LLL another chance, instead of dismissing it because it uses too many algs. 16 algs might sound like a lot, but there are common patterns among them, and I guess the important part is that you don't have to learn all of the algs in order to start using 4LLL. If you get a case you don't know, just do something to turn it into a case you know (try a Sune, or a U perm, or whatever), and maybe when you're done with the solve, look up how that case can be solved.

 

[renyhp]

I see what you mean. In few words what you guys are trying to say is that "standard" beginner's methods, although being less optimized than that, are an easier way to transition to CFOP... exactly because they use more algs. And that's the reason why, if you want to become a speedcuber, you should better try beginner's method instead of some sexy / 8355 method - simply because these won't let you experience what an advanced method is.

 

[Rafael Paulino]

If you're worried about the sheer number of algs to learn, you won't actually remember the moves in your brain. Iso helps to do a few solves with theach Notation of the algs in front of you. As you do them over and over, your muscles will adapt to it. From a sub-24 solver's perspective, the sexy method is VERY inefficient.