Older cubers discussions

[JanW]

In my view, OP is a good foundation for learning comms. Comms is just about solving two pieces at a time. You solve exactly the same sequences of pieces. So first learn a method that lets you solve the corners one piece at a time, and then you can add comms at your own pace and pick off pairs within the sequence that you recognise. OP doesn't have to slow down your progress on comms at all.

Sorry, I'm not arguing for the sake of arguing, just want to understand the reasoning behind this. At the moment I find comms fairly easy to recognize and execute intuitively. I had several more successful blind corner solves yesterday using comms. Slow, obviously, about 1 minute memo and 1:30-2 minutes execution, but from what I've read about people's first blind successes, this should be quite acceptable for corners in first attempts. I am fairly sure that I couldn't solve the corners blind with OP any faster with this little practice. So I'm at pretty much the same starting point for both comms and OP (except that I don't know the Y-perm yet). From here I could continue practicing with comms to make them faster, or switch to OP and make that faster to eventually add in some comms. I just can't see how the second option would yield better long term results.

However, I can see one advantage for comms without any OP involved. I could use UFR as the buffer, which makes execution of several insertion cases a lot easier. It also makes visualization easier, since I'm quite used to inserting the UFR corner into different positions already from F2L. I used UBL yesterday, did a few sighted solves with UFR today and it immediately cut 10-20 seconds of my times. With UFR as buffer, I could handle parity for example by solving UR edge to FR and vice versa, then there are several different 11 move RUD algs to choose from for shooting UFR to either FDR or RDF and solving the edges. All corner stickers but DFR can be brought to either of those positions in max 2 setup moves. Some of the possible algs would in reality be 10 movers as the initial D-face move mostly would be a cancellation to the setup. The rare case of parity with FDR as last target would require either 3 move setup, or learning a separate alg.

 

[mafergut]

Sorry, I'm not arguing for the sake of arguing, just want to understand the reasoning behind this. At the moment I find comms fairly easy to recognize [... stuff removed...] The rare case of parity with FDR as last target would require either 3 move setup, or learning a separate alg.

You know what? I don't know much about commutators. I could barely solve successfully blind once or twice like 1 year ago with OP for corners and edges. But here's what I know:
- Solving with OP took ages and the move count is horrible, like 18-20 moves per target on average counting setup + Y/T/J/L-perm + undo setup.
- Reading your post it's clear to me that you already know a lot about the cube, how it works, how comms work and how to solve the corners.

Taking these two facts into account my advice is, maybe you should take your own decision instead of following any advice (including mine). It's not like you can barely understand the cube and will get confused or anything, as I said, you seem to know very well what you're doing already. Wanna skip OP and go straight to comms? Then do it. And tell me (us) how it goes because I want to go back to BLD at some point in the future and, for sure, I won't want to start all over again with OP..., unless I try another method and fail miserably.

 

[mark49152]

Sorry, I'm not arguing for the sake of arguing .... (stuff)

There's no argument, and I'm not offering advice either, only opinion. Discussing cubing is the fun of the forum. Whether you agree or disagree, that's equally good, but if nobody wants to talk about it, well that makes me disappointed .

The advantage of starting with a 2-cycle method is that you don't have to know or deduce a comm for every pair. You can add them as you go and transition gradually. It sounds like you are making great progress with comms alone, which is impressive, so maybe that transition is of no benefit to you. I don't find comms that easy, and could not figure them out for every pair. Learning hundreds of them alg-style isn't appealing to me either. So starting with M2/OP and adding some easy comms as and when I feel ready suits me personally.

As for speed, ultimately move count will become important at around 1:00-1:30, but you also have to factor in thinking time and TPS. Same trade-offs as in regular speedsolving. Yes OP has a lot of moves, but it's just simple, repetitive, no-brainer spamming of the same easy alg and easy setups. Usually my corners execution has pretty much zero pauses. I just tried an ao12 of corners exec only, with pure OP, and got 23.56. That's ~6 tps which is faster than my 3x3 last layer. So a comm is no use to me unless I can recall, regrip and execute it in ~5 seconds or less. Sure I will have to transition to comms and reduce move count eventually, to get faster, but I'll do that by adding them at a pace at which I can integrate them without hurting speed, just like when learning full OLL.

You're right that if you want a different buffer, OP wouldn't be helpful.

Mark's advice was just an analogy. He was saying how he finds Hoya easier than Yau in the same way that he finds cross + F2L easier than block building.

What I meant by that comparison was that Hoya cross is more structured at the expense of extra moves, whereas Yau cross is more free-style and requires hunting round the cube a bit more, perhaps using more intuitive solving. Same things you could say about F2L and blockbuilding respectively. I gave up on Yau before I was any good at it though, so maybe that's a misrepresentation.

 

[Jason Green]

Here's my cheat sheet in case it is helpful:

http://cubing.mikeg.me.uk/Ortega.pdf



Mark's advice was just an analogy. He was saying how he finds Hoya easier than Yau in the same way that he finds cross + F2L easier than block building.

Yau does not utilise block building and both methods are forms of reduction. The four cross pieces are solved prior to completion of the centres, leaving just eight edges to be paired before 3x3 stage. Watch videos for both and have a play with both to see if you have a preference.

It's very helpful thanks, I could not find a cheat sheet I liked yet and this looks great! My 2x2 comes today so great timing.

I understood it was just an analogy, but the fact that he felt it was an appropriate analogy still made it compelling. I'll check them both out... Some day.

 

[JanW]

The advantage of starting with a 2-cycle method is that you don't have to know or deduce a comm for every pair. You can add them as you go and transition gradually. It sounds like you are making great progress with comms alone, which is impressive, so maybe that transition is of no benefit to you. I don't find comms that easy, and could not figure them out for every pair. Learning hundreds of them alg-style isn't appealing to me either. So starting with M2/OP and adding some easy comms as and when I feel ready suits me personally.

Yes, I can see that point. OP is certainly more straight forward. When it comes to comms, I think they're not at all as complicated as they first seem. I honestly think there are F2L cases that are harder to figure out intuitively than the most difficult commutator cases. Not saying comms are easier than F2L though. They are harder, because there are so many different cases and you have to deal with all corners and think from several different directions, instead of just dealing with U layer.

The multitude of cases isn't that bad, because they basically fall into a few groups of very similar cases. For example UBL to LDB to FDL (L U L', interchange then reverse) is basically the same execution as UBL to any combination on LDB, FDL, RFD or BRD. Only difference is the rotation of interchange layer. With one setup move it's also the same as UBL to LDB to any corner on the cube except BUR, UFL, LDF, DBR, and a whole lot of other cases involving FDL or BRD. Or any of those in reverse if you do interchange first. In reality the amount of truly different cases is fairly small. Some of them are a bit more tricky, when those come up, I have a solved cube next to me that I can use to work out a good solution and try to remember the principles of that solution for later cases. Because similar tricky cases will come up in different shapes later. The amount of fundamentally different tricky cases is very small.

The problem of having to deal with the cases from all different directions became a lot easier once I started using UFR as buffer. Maybe if you are more used to inserting pairs at the back during F2L, then this doesn't make much of a difference.

 

[mafergut]

[...] When it comes to comms, I think they're not at all as complicated as they first seem. I honestly think [...]

Seeing how you find so easy something I gather as quite complex on your first try makes me feel kinda dumb

 

[JanW]

Seeing how you find so easy something I gather as quite complex on your first try makes me feel kinda dumb

Wasn't immediately easy at all. After watching Noah's videos on corner commutators I tried for myself, got the first attempts horribly wrong and didn't understand at all why. I think Noah's videos do a really good job explaining the principles of what a commutator is, but he doesn't really explain at all how to approach them when working out new cases. I mean the basic thought process that should go through your head to figure out the steps. Took an hour or two of really thinking hard about the simplest example cases from the videos and the underlying mechanics to get what it's all about. But once I had it broken down to simple "if A then B, else C" kind of thinking, then it suddenly became all clear. And since that basic principle is the same for every single case, it's all a matter of setting up the pieces for the switch after that.

Basically my thinking goes like this:

1. Locate the stickers on the cube
   
2. Check if they are already located in such a way that I can do a commutator
   • A possible case requires that 2 stickers are interchangeable by rotating one layer only and the third corner piece must not be on the same layer. If this is the case, move on to step 4. If not, go to step 3.
3. Set up corners for possible commutator insertion
   • If none of the three stickers are interchangeable, find the one move that makes 2 of them interchangeable in a different layer than the third piece.
   • If 2 are interchangeable but the third piece is in the same layer, move it out (or sometimes turn another layer into an interchange layer).
     
     This is the hard part, but as I said in the last post, the amount of different possible cases is not very big at all. The trickiest cases I find to be those with all 3 pieces in the interchange layer, but I've worked out good solutions for most of those cases already, I think. One example that was tricky at first, but in reality is very simple, was cases like UBL->RUB->UFR. At first glance UBL and UFR are interchangeable and should be the interchange, but RUB is on the corner between them and cannot be moved out without disturbing any of the other two pieces. The solution is simple, setup with F and now the R layer is your interchange layer. On top of that you get very easy insertion with U' L' U.
4. Insert and interchange
   • At this stage the interchange layer is basically like your D layer in F2L. To insert, you bring one corner out, replace it with the desired piece and put it back. Just like in F2L, no other piece than the corner you are inserting may be affected by the insert. Then rotate interchange layer to bring the second target to the right spot and repeat insertion moves backwards.
   • There are 3 possible ways to insert a piece, depending on the orientation you want. For example, inserting UBL to D layer can be done as L U L', B' U' B or L2 U L2 U' L2. Which alg to choose depends on what sticker you are aiming at.
   • To know whether you should do interchange first or not, the rule is that the sticker you insert should always shoot to the sticker from the location where it is supposed to end up.
     
     For example if we have the cycle UBL->FLU->RFD, UBL and FLU are interchangeable with an L move and third piece is not on L layer. Great start. On top of that, you can insert RFD to FLU with a simple U' R U. To figure out if you should do interchange first, you have to think about where the RFD sticker is supposed to end up when everything is done. In reality the cycle is UBL->FLU->RFD->UBL, because the third sticker is always brought back to the buffer, so you know the first insertion should aim at the buffer. Hence, you have to do the interchange move (L) first to bring your buffer to FLU. So the full cycle becomes L - U' R U - L' - U' R' U. If the cycle was the other way around, UBL->RFD->FLU, you can tell that RFD is supposed to end up at FLU, so you do insertion first, then interchange.
5. Undo setup
   • Easy forget how you set it up. At least I still do it too often...

There's still a lot that can easily go wrong. It's not always easy to see which insertion method to use when inserting into awkward corners. It can very easily end up in the wrong orientation if you choose wrong. Sometimes I still mess up and do insertions in wrong order, forget to do interchange first. And then there's the worst mess up of all, performing the insertion alg in wrong direction... for example if I want to insert FRU to RFU with U' R' U, I instead insert RFU to FRU with U L U', which will royally mess up the whole cube if I then do interchange with L layer, reverse and undo possible setup moves.

 

[h2f]

It looks like you got it.

patataj patataj patataj

 

[Jason Green]

Here's my cheat sheet in case it is helpful:

http://cubing.mikeg.me.uk/Ortega.pdf

Mike, hope you don't mind if I share a link to you cheat sheet in the comments of a YouTube video. If you do mind I'll take it out. I'll share the video after I make it.

 

[Jason Green]

Here's the video I made. It's pretty boring, I really just made it because I thought my friend's son would enjoy it being "for him."

https://www.youtube.com/watch?v=bHWvrcuI76s

I finally had an Ao12 under 30, woohoo! It wasn't for my race though. I also had a 21.20 single that could have been my first sub 20 possibly but I glanced at the time and got nervous!

Spoiler

Generated By csTimer on 2015-11-5
avg of 12: 29.43


Time List:
1. 29.15 R' U2 F' R2 D2 B2 U2 R2 F' D2 U2 B D B' U B' F L2 R' B F2
2. 31.86 U2 F2 L2 D B2 U L2 F2 D B L B F L' R B' L' U' R2
3. 27.25 R2 B2 D2 U' B2 D2 B2 R2 D2 R2 B' R U2 L2 U2 L2 U' B' F L
4. 31.60 L' U F' D2 R F U2 D2 R' U2 F D2 B L2 F2 B L2 F2 R2
5. 30.91 L F2 L2 F2 D2 L2 F L2 B R2 F2 D' B U B' U' F R' D2
6. (26.40) B' F2 L' B2 F2 D2 L' R' D2 F2 L' F2 U B2 U L' B D B2 L
7. 28.07 R F2 D2 R2 D2 F2 R2 B' F2 L2 D2 U' L2 R U' F2 L F D B'
8. (33.93) D2 U2 B2 F U2 B2 L2 U2 F' R2 D2 L' D' B2 L2 B' L D2 R'
9. 31.68 D' R F U' R2 L B' D2 F' L' B U2 B2 D2 B' L2 F' D2 F L2 D2
10. 27.47 B2 D2 F2 U L2 D U2 B2 R' B2 D2 U' F2 D2 U' F' L U
11. 26.48 R U2 B' U2 L2 F D2 U2 L2 F R2 F' R U2 F2 L' D' L' R2 B F
12. 29.89 D' B2 F2 U' R2 B2 D R2 B2 U' L2 F' L D' B L' U' L2 B' D2 R

 

[muchacho]

It's very helpful thanks, I could not find a cheat sheet I liked yet and this looks great! My 2x2 comes today so great timing.

Yeah, that cheat sheet is very well done, gj!

Here's the video I made. It's pretty boring, I really just made it because I thought my friend's son would enjoy it being "for him."

https://www.youtube.com/watch?v=bHWvrcuI76s

It's good, what a newbie needs.

The PBLs from the cheat sheet are better than those from crazybadcuber's video (btw, that's the one I learn 2x2 from).

You should learn as soon as possible to make the first face on bottom, and being CN.

I finally had an Ao12 under 30, woohoo! It wasn't for my race though. I also had a 21.20 single that could have been my first sub 20 possibly but I glanced at the time and got nervous!

Spoiler

Generated By csTimer on 2015-11-5
avg of 12: 29.43


Time List:
1. 29.15 R' U2 F' R2 D2 B2 U2 R2 F' D2 U2 B D B' U B' F L2 R' B F2
2. 31.86 U2 F2 L2 D B2 U L2 F2 D B L B F L' R B' L' U' R2
3. 27.25 R2 B2 D2 U' B2 D2 B2 R2 D2 R2 B' R U2 L2 U2 L2 U' B' F L
4. 31.60 L' U F' D2 R F U2 D2 R' U2 F D2 B L2 F2 B L2 F2 R2
5. 30.91 L F2 L2 F2 D2 L2 F L2 B R2 F2 D' B U B' U' F R' D2
6. (26.40) B' F2 L' B2 F2 D2 L' R' D2 F2 L' F2 U B2 U L' B D B2 L
7. 28.07 R F2 D2 R2 D2 F2 R2 B' F2 L2 D2 U' L2 R U' F2 L F D B'
8. (33.93) D2 U2 B2 F U2 B2 L2 U2 F' R2 D2 L' D' B2 L2 B' L D2 R'
9. 31.68 D' R F U' R2 L B' D2 F' L' B U2 B2 D2 B' L2 F' D2 F L2 D2
10. 27.47 B2 D2 F2 U L2 D U2 B2 R' B2 D2 U' F2 D2 U' F' L U
11. 26.48 R U2 B' U2 L2 F D2 U2 L2 F R2 F' R U2 F2 L' D' L' R2 B F
12. 29.89 D' B2 F2 U' R2 B2 D R2 B2 U' L2 F' L D' B L' U' L2 B' D2 R

That was your first sub-30 Ao12? I had the impression I was behind you in the race to sub-10... my best Ao12 is sooo much better, 29.38

 

[Jason Green]

That was your first sub-30 Ao12? I had the impression I was behind you in the race to sub-10... my best Ao12 is sooo much better, 29.38

Yeah I think so. I had some ao5 of like 27.x but not ao12.

At this point I just want to solve to 2x2 fairly fast (judged by a non cuber). I try to limit what I get obsessed about because once I do, I can't stop! Like the 3x3 now!

 

[mark49152]

@Mike: are you going to both days of UKC?

 

[Logiqx]

Mike, hope you don't mind if I share a link to you cheat sheet in the comments of a YouTube video. If you do mind I'll take it out. I'll share the video after I make it.

No problem Jason. Nice video.

Two tips...

First face: Do it on the bottom and plan it during inspection, aiming for 5 moves or less. Once you can plan the first face you should also learn to predict the bottom permutation. This means that you never have to look at the bottom layer after inspection and only need to look at the top layer permutation during PBL. This takes practice but planning and predicting is one of the fun things about 2x2.

Recognition: Learn to recognise corner permutation by looking at just two sides. The corner colours on a side can either be matching, adjacent or opposite. When you see two sides it is possible to infer the other two sides. The side opposite to "headlights" will always have opposite colours... I call these "tail lights".

@Mike: are you going to both days of UKC?

Looking at the weather forecast, I'll only be doing the Sunday of UKC.

 

[mark49152]

Looking at the weather forecast, I'll only be doing the Sunday of UKC.

No 2x2 or 5x5? 2x2 is early as well.

 

[Logiqx]

No 2x2 or 5x5? 2x2 is early as well.

I definitely won't do 5x5 since that's when I'll be on the water and I haven't picked it up in 3 weeks. I could potentially do 2x2 but it would be at the expense of any 3x3 / OH practice tomorrow. I'd leave immediately after 2x2 and not have the opportunity for any cubing until the competition. My mistake at ABHC was neglecting 3x3 / OH practice in the lead up to the competition.

Speaking of which, I wish 3x3 wasn't so early in the morning. Jo is going to hate me for being woken up early on a Sunday and I'm likely to be half asleep!

 

[mark49152]

Speaking of which, I wish 3x3 wasn't so early in the morning. Jo is going to hate me for being woken up early on a Sunday and I'm likely to be half asleep!

Yeah it doesn't leave me much time for a leisurely fry-up in the hotel . Not complaining though because I can see why they need to do it. Have fun surfing and see you on Sunday!

 

[MarcelP]

Good luck guys! Get some nice PB's

 

[muchacho]

Good luck and better PBs!


For BLD I'm going to start adding letters pairs to a list. I want to make it mostly from Spanish words but there are some letters that are not very useful (K, W, X, Y)... I'll need to use one of those (for English words), which one would be better? K or W?

My other 23 letters:
A, B, C, Ch, D, E, F, G, H, I, J, L, M, N, O, P, Q, R, S, T, U, V, Z

 

[mafergut]

Good luck and better PBs!


For BLD I'm going to start adding letters pairs to a list. I want to make it mostly from Spanish words but there are some letters that are not very useful (K, W, X, Y)... I'll need to use one of those (for English words), which one would be better? K or W?

My other 23 letters:
A, B, C, Ch, D, E, F, G, H, I, J, L, M, N, O, P, Q, R, S, T, U, V, Z

When I practiced 3BLD a bit I would just use an unaltered Speffz letter scheme (ABCD, EFGH, IJKL, MNOP, QRST, UVWX) but I don't know the reason. Probably because that's what Noah recommended in his video on memorization, I guess.

What I did was:
- for C I would use Z-sound Spanish words: e.g. for CN I'd use something like "cena", "cisne".... That's why I wouldn't recommend to have both C and Z in your letter scheme
- for K I would use K-sound Spanish words, even if written with C: e.g. for KN I'd use: "cana", "cuna"
- problem is that Q also sounds like K in Spanish (and in English ) so, for Q I'd use if possible words that are written with Q, but there are not that many of those in Spanish, so for example for QN I I'd probably use an English word like Queen (I like chess, so I'm sure I'd use Queen).
- for W, which is very difficult to associate to a phonem and also there are almost no words in Spanish, I don't know but maybe I'd use words with V but I would imagine two of them, e.g.: for WN I'd use "vino" but imagine two bottles instead of one.
- for G I used soft G sound, e.g. GN = "gana"
- and for J I used strong G sound, e.g. JN = "Jaen" or "Jonas", even if the word is written with G (like JM = gema) but I'd prefer a word written with J if possible (JM = jamon).

I don't think there are other problematic letters in Spanish for image/word memo but if you memo your corners with sounds in a scheme where you memo edges-corners and execute corners-edges then I don't really have experience with that but I would just encourage more the association to the phonem and not the spelling. Then I most surely would rely on short English words for W, like water, wine, ... but, really, I don't know what I'm talking about here. My short term memory is quite bad so I always used long-term image memo for both corners and edges.