[Help Thread] Megaminx/Master Kilominx/Gigaminx BLD discussion

[bobthegiraffemonkey]

Hi guys, thought I would start a thread for megaminx and bigger BLD discussion (what little there is), starting with my tutorial. As of starting this I only have part 1, but I'll post the rest here when I make it. I'll do my best to answer any questions here about the subject, both about the tutorials and other general questions.

My tutorials:
Part 1: memo
Part 2: execution
Part 3: tracking
Part 4: other stuff (coming soon)
Part 5: wings
Parts 6+: other MK/giga (coming soon)

Hope to spark some interest in this as I think it's pretty fun .

Matt

 

[Username]

I've always wanted to do megaBLD, and I will start practicing now

 

[etshy]

This is the same lettering schema as mine except I use arabic instead of greek and I start numbering from 1 not 0
I can't wait to get a success :rolleyes:
Great tutorial Matt

@username : you should try it , it's very challenging and so much fun ( although I still didn't get a success )

 

[emolover]

I've always wanted to do megaBLD, and I will start practicing now

I also have a memo tutorial.

 

[A Leman]

I still don't understand how you make images out of those letters anymore than I did a while ago, but I guess that is ok. could you tell me what you would memo

6B H9 (alpha)3 (beta)w 45 (gamma)(delta)? That is not to much but it covers combinations and I would find it insightful. Also, I really like this thread. The challenge is difficult but I think many more people are capable of doing it than there are now. Also, I have been changing and refining my execution so I will hopefully have a faster success sometime soon.

 

[bobthegiraffemonkey]

Nice to see some responses, and I've not even posted a full tutorial yet!

I still don't understand how you make images out of those letters anymore than I did a while ago, but I guess that is ok. could you tell me what you would memo

6B H9 (alpha)3 (beta)w 45 (gamma)(delta)?

I've posted this before, but last time I tried to find it I couldn't so here it is again. It's almost certainly inferior to standard letter pairs in terms of speed, but I like it and it gives me the flexibility to go do megaBLD with minimal fuss, and it's also less to learn.

I make images with 4 characters at a time, with english letters (26), greek letters (24), and digits (10). I have a 4*60 'items' in my memo scheme, one for each character in each position. The first character gives me the main object (mostly Pokemon), the second gives some action for it to do, the third is for weather or a similar effect, and the fourth is another object in the image. I take these components and combine them.

After using it for a while I've had to relax the criteria for each to allow for better images, and a lot of components have some degree of flexibility in how I apply them.

As an example from above, 6BH9 is scizor [6] (a Pokemon) bowling in hail [H], and the ball hits a bully [9] (not in the usual sense, it's an enemy in Super Mario 64 ...). Note that while most components started off being related to the letter in some way, it's easy to not follow this rule when changing small amounts of the system and just learning them since you can just think of good things to use and you don't have to worry about making them fit. I like using images and routes from computer games as it allows for exaggerated slapstick images which I find comfortable.

I'll also take a note of this description in case anyone asks again.

 

[bobthegiraffemonkey]

OP updated with execution video.

 

[A Leman]

Nice to see some responses, and I've not even posted a full tutorial yet!


I've posted this before, but last time I tried to find it I couldn't so here it is again. It's almost certainly inferior to standard letter pairs in terms of speed, but I like it and it gives me the flexibility to go do megaBLD with minimal fuss, and it's also less to learn.

I make images with 4 characters at a time, with english letters (26), greek letters (24), and digits (10). I have a 4*60 'items' in my memo scheme, one for each character in each position. The first character gives me the main object (mostly Pokemon), the second gives some action for it to do, the third is for weather or a similar effect, and the fourth is another object in the image. I take these components and combine them.

This is a very interesting method and like the PAO a lot of people use for cards with weather added to make the system even in number like letter pairs. I can see how with a lot of practice it would be almost/about as fast as letter pairs for cubes and It would definitely be good for Megaminx as well to expand a single letter system like that. I was under the false impression that you were using 3600 unique pair images which you would make up on the fly.(hence why I put the different combinations of letter type and was confused). Thank you for sharing.

 

[etshy]

OP updated with execution video.

Your 2 cycle method is easier than the one I use , using the 3 Cycle Alg and its inverse makes setup moves easier than mine , because I have to deal with 4 edges that change position with each target , so I don't have that much freedom , although X-PLL is a very easy alg to execute but the fact that I have 2 slave buffers is really not that good , but nevertheless I think that my setup moves are almost very good at this stage .

I'm actually very surprised that you can have a 12 mins execution with a 2 cycle method , I can barely get a 25 mins execution
but I have a question concerning the other 2 methods , you said that one of them could be used on Gigaminx , I never tried or thought about how to do wings or centers on MK or gigaminx , but does a 2 cycle method can be used on those too ?

Once I get a success I will be making a video showing my setup moves on the method that I used , may be it will help the people who are interested in different methods

 

[bobthegiraffemonkey]

I'm actually very surprised that you can have a 12 mins execution with a 2 cycle method , I can barely get a 25 mins execution
but I have a question concerning the other 2 methods , you said that one of them could be used on Gigaminx , I never tried or thought about how to do wings or centers on MK or gigaminx , but does a 2 cycle method can be used on those too ?

My execution is still slow, sub-10 execution is easily possibly with my method. With a faster execution method in the hands of a fast BLDer, I'd say sub-10 solves are possible. I'd stick with 2-cycles anyway, 3-cycles are harder on megaminx.
As for giga, it's just the algs I was talking about, the <R,U> 3-cycle doesn't work on giga, the other two do, and it's not too difficult to apply them. The idea of the methods are identical, just the algs and setups are different. I would use the second of the three edge cycles I showed if I tried gigaBLD.

 

[TheNextFeliks]

Does mega compare better to 4x4 or 5x5 bld?

 

[etshy]

My execution is still slow, sub-10 execution is easily possibly with my method. With a faster execution method in the hands of a fast BLDer, I'd say sub-10 solves are possible. I'd stick with 2-cycles anyway, 3-cycles are harder on megaminx.
As for giga, it's just the algs I was talking about, the <R,U> 3-cycle doesn't work on giga, the other two do, and it's not too difficult to apply them. The idea of the methods are identical, just the algs and setups are different. I would use the second of the three edge cycles I showed if I tried gigaBLD.

I will stick to 2 cycles too on Mega , while ago I thought that no one can get decent times with it but you proved me wrong so I will try work on my execution

I won't attempt MK BLD any time soon but I will try these cycles on it ( when it arrives ) to just have an idea on how things works on it

and I'm thinking of doing a tutorial on how I do 2 cycles , if anyone is interested

Does mega compare better to 4x4 or 5x5 bld?

It actually depends on how you do it

for memo , if you use 2 letter per target ( 1 image per target ) you will need approximately 50 images , which is almost as much as a 5BLD
but if you use 1 letter per target , you will need only half of these images , which is something like a 4BLD
but the main issue here is how to trace pieces while memoing , because you have 30 edges and 20 corners , so tracing for me is a lot harder than 5BLD

For execution , you really can't compare it to any of them because it's a totally different way of dealing with setup moves ,Faces,,etc , but the main problem with Megaminx is that it's VERY EASY to lose your orientation while solving , that's probably the reason of half of my DNFs

 

[etshy]

does anyone have a better alg for X-PLL R' L' U2 L U R y R L U2' R' U' L' y' that doesn't have rotations ? I had a lot f DNFs because of forgetting to do the y' at the end
I thought about replacing the y and y' with U and U' but still I might forget them

 

[bobthegiraffemonkey]

does anyone have a better alg for X-PLL R' L' U2 L U R y R L U2' R' U' L' y' that doesn't have rotations ? I had a lot f DNFs because of forgetting to do the y' at the end
I thought about replacing the y and y' with U and U' but still I might forget them

Odd timing, I'm just uploading part 3 of my tutorial.

I'd use R U' R' U2 R U' R' U2' R' D' R U R' D U2' R U R' U R, but that's weird and there is probably a more common one. The one you have already is fine, you just need to be more careful if you're using it, and also doing U/U' instead of y/y' doesn't work.

 

[etshy]

Odd timing, I'm just uploading part 3 of my tutorial.

I'd use R U' R' U2 R U' R' U2' R' D' R U R' D U2' R U R' U R, but that's weird and there is probably a more common one. The one you have already is fine, you just need to be more careful if you're using it, and also doing U/U' instead of y/y' doesn't work.

that's a nice coincidence

yeah I think I need to be more careful
U/U' won't work , I just thought that the first part of the alg only damages the top layer , and obviously I was wrong

 

[bobthegiraffemonkey]

Part 3 of my tutorial can be found here, OP also updated. Part 3 is tracking pieces during memo.

 

[bobthegiraffemonkey]

Wings: memo + exec. Because I felt like it. (OP updated).

 

[Jacck]

There seems to be a bit of interest in Megaminx bld (maybe Gigaminx bld, too) again, so I revive this thread.
I will repeat my short discription of the Gigaminx bld and a discussion with Roman and bobthegiraffemonkey first and than I will add a new post with a new, easy(?) concept for the Megaminx-edges. Since now I was solving the Megaminx-edges with X-perms, but I think, the new one is easier to learn.

But first the posts:

Thanks to all!
It's a slowcuber method: not optimized for speed or movecount (therefore not many algs, not thinking over the entire cube, no extraordinary intuition), but optimized for success - even if it takes longer.

The most important thing is, I guess, that I solve (normally) only one cubie at the time and bring this one to the front with D+ and D- first, so I have quite easy setups. You could solve two cubies at the time, but you would have to think more while setting up (which is upsetting a Slowcuber).

So, the memo is less color-orientated but position-orientated. With the number of the cubie I can say, how many D+ it is away from the front and which setup I have to take.

Spoiler: Memo

White is on top, green in front (it's a Shengshou).

Wings are numbered like the edges on Megaminx (my method), starting Top-Front with 1 (white-green) and counterclockwise to 5. Next "layer" starts with 6 (purple-green), then 11 (brightblue-green), 16 brightyellow-green, 21 (brightblue-brightyellow) and 26 (grey-brigthyellow).
I add a 30, if the wing has to be flipped. But when is it flipped? Not flipped is a wing, that I can bring to the target with my "easy" setup. Look at the 6: it is the lower purple-green and can be brought to the target (1) with a simple F. The setup for 36 (the up-purplegreen) would be D+ F'.
The setups for 6-10 e.g. have only a different number of D+.

T-Centers (between center and midge) with two digits: Face and position in face. White is -, green is 1, red 2 ... brightyellow is 6, brightblue is A and grey is B. Position goes from 1 to 5, so I get something like 1 (the white one in front), 23 (third on the second side, red) or B5 (fifth on bottomside).
I have a major-list, which I extended with A and B. The reason: I am bad with orientation, so I solve the 3x3x3 with orientation first, which leaves 12 edges and so I have one digit for them only.
Edit (2015-07-18): A Minx can be done with a normal major-list, too: instead of A1 to A5 you could take 01 to 05, instead of B1 to B5 you could take 6 to 10.

X-Centers same as the T-Centers.

Corners from 1 to 5 on Top, next "layer" 6 to 10, then 11 to 15 and 16-20 on bottomside. If it has to be twisted clockwise, I add 20, for counterclockwise 40. With new cycles I try to bring in the Top-Corners with the correct orientation, the other Top-corners are orientated at the end. "Orientated" corners are with white side up or grey side down or on front of the upper-face (there could be better systems, but I started that way for myself on the megaminx).

Midges: like the wings.

So there are only two-digit numbers (if you take A and B as numbers), which I memorize in pairs on routes.

Spoiler: execution

wings: buffer is 34 (yellow-white), target 1 (white-green), co-buffer 35 (purple-white)
for the first wing: L' U' L U l' U' L' U' Lw
and for the second wing the inverse
(this alg I found in the early 80s for a LBL-method for the 4x4x4 - spended all my money for the cube and had nothing left for a solution book, and the internet wasn't invented yet.)

T-Centers: buffer is 5 (near purple face), target 61 (brigthyellow near green), co-buffer is 1 (near green)
and this is an alg from this forum:
Lw' Rw f2' Lw Rw' U' Lw' Rw f2 Lw Rw' U
and the inverse.
Sometimes you would not need the U at the end, but all my setups in the upper half would mess the top-face. E.g. a front-T I would bring to 61 by Fs, till it is near the brightblue face and then DLw DRw (and the Fs mess the top-face).

X-Centers: buffer is 1 (near purple-green), target is 2 (near green-red) and I get the cubies there over the 13 (green near red-brightyellow). I shoot to the target from A2 with dr, from 24 with dr' or from B1 with dr2. E.g. A2 (brightblue near green and brightyellow): r' dr r U' r' dr' r U and the second X would be the inverse, that means starting with U'. The main thing is the U' and U. Not shure, whether I explained it right.

Corners: similar as the X-centers, e.g. setup for 7 (red-brightyellow-green): R' DR' R then the U' and back with R' DR R.
You will need a few setups there, but they are pretty intuitiv.
Orientation at the end with: (R F' R' F)*2 and I'm shure, you will guess the rest.

Midges: like T-centers, buffer is 5 (white-purple), target 16 (brightyellow-green), co-buffer is 1 (white-green)
and again an alg from this forum:
Lw' Rw F2' Lw Rw' U' Lw' Rw F2 Lw Rw' U
and the inverse. Just the same as the T-Centers, only an F instead of an f.
E.g. for 41 (green-brightblue, the left one): DL DR alg DR' DL'
or for 11 (green-brightblue, the right one): D+ DL' alg DL D-
Flipped-midges on top at the end: I bring one to 16, take the other as buffer, do the alg, flip the 16 to 46 by D- DR2 DL, do the alg' and then all back.

For all cubies (except wings): if there is a Top-cubie involved, this one will be used as target, co-target or buffer and you will save one alg.

OK, that's it: really not a lot of different algs and the Midges and T-centers are very similar.

Spoiler: Hmmmm,

perhaps it is possible to find a more coherent memo- and exe-system with my ideas. But my goal was not to get a smart method, i tried to find a method with the things I already knew, and which is working. Good luck for you all and I hope, I did not make too many mistakes.

And now the discussion:

What is your method and why is it so hard - memo, exe, setups?
I do corners the same as with the Gigaminx:
https://www.speedsolving.com/forum/...omplishment-thread.3582/page-779#post-1170748
and the edges mostly with X-Perms. I use 4 as buffer and 1 as target (and 2 and 5 as the "changing-pair") and:
X-Perm (1<>4, 2<>5): L' R' U2 L U R y L R U2' R' U' L' y'
X-Perm with flip (1/<>4/, 2<>5): R U' R' U' F U F' R' F R U' F' R U2 R'
U-Perm for 2 and 5 (4>2>5): (R' U' R) R' U' F' U F R L F U F' U' L' (R' U R)
Edge-flip (1/2/): L F R' F' L' U2' R U R U' R2' U2 R
they are all from https://sites.google.com/site/permuteram.../old-p-bld
Having a X-Perm with flip gives you easier setups, but the midges-method of the Gigaminx works too - and you don't need a U-Perm.

Please, don't give up!

I use 3cycles as I already got used to them. Moreover, I think that "old Pochmann" on megaminx isn't much easier than 3cycles. I was thinking of a TuRBo method with learning a bunch of 3ELL algorithms, because I couldn't think of a way to perform pure edges comms on mega (there's no M moves possible). But then I asked Lilia for help, and she showed me edge-3cycle basics: ((L' R) F' (R' L)) changes only one element on the top layer. Instead of F' it could be F, F2, F'2. So a typical 3cycle looks like [L' R F R' L, U2] or [L' R2 F' R2' L, U]. And with corners, it's more or less obvious how to do commutators, they're the same as in 3x3. I will probably do a video with more detailed explanation of my method when I have my first success.
Now the most difficult part is tracking pieces during setups, and understanding what layer I have to turn, because when I close my eyes I instantly forget what does the megaminx look like and what faces does it have!
I will not give up, but I can't say how long will it take me to have my first successful solve.

I think this is the way to go. Before, I used pseudo-2-cycles, but when I try megaBLD again (eventually, might be a year or two before I do it) I'll do it with 3-cycles. Pseudo 2-cycles make it much easier, but 3-cycles are practical enough that there's no good excuse to not do them to get really fast. The only problem is that 3-cycle setups are hard and I've not worked out the details yet (I'd rather figure it out myself in my own style rather than watch a tutorial for it).

Good luck Roman, I want to see what you can do with megaminx BLD.

Sorry, I don't really know the meaning of a 3-cycle or a pseudo-2-cycle. But the ((L' R) F' (R' L)) is the same I use for Gigaminx-midges only with a F2 in the middle (and it could be done as well for Megaminx-edges). I think, for a real 3-cyle I would have to put a second piece on the top-face, to solve two at the time, right? I just solve one at the time an my second top-edge is only a helper (is that a pseudo-2-cycle?).
But the setups for a real second top-edge must be hard and I don't know, if you save much time and if you make it easy (by first bringing that one to the front with D+, then put it on Top, then D-, then setup for the other edge), you save many moves.

The advantage of solving only one edge at the time is, that you don’t have to think about difficult setups and you can solve really fluently. Okay: it is a slowcuber-method to get a success at all - and getting really fast needs more difficult ways. But is my method that slow? I memorize a 5bld in about 12 min and exe is about 10 min. That’s 300 % of a real good blindcuber, but with the Megaminx I am less than 150 % of the UWR.

Clear: if you could solve the Megaminx with 3-cycles as fluent as with the one-after-each-other-slowcuber-method, then you would be very fast. But you'll have to learn a very lot before.

@Roman: have you tried to memorize positions or colors?

@Jacck 3-cycle is moving 3 pieces with an alg, usually solving 2 of them, it's what 3-style does which is what most top BLDers use on cubes. What I'm calling 'pseudo 2-cycles' is something like M2 or Old Pochmann, where you solve with 2-cycles (a single swap of pieces to solve one piece at a time, this seems to be what you use) except you can't do pure 2-cycles so there are side effects. I guess a lot of people just call this 2-cycles, but for some reason I prefer saying pseudo 2-cycles for megaminx, I guess since I use 3-cycles for it and it feels weird.

For gigaminx, one piece at a time for middle edges might be better, since the presence of wings means you have less freedom. I used <R,U> U-perms on megaminx for BLD, which is fine on megaminx but it doesn't work for gigaminx. The comms that you and Roman both described are probably the best option for gigaminx.

P.S. Someone should probably move this to the megaminx BLD thread.

And the Gigaminx:

 

[Jacck]

Slowcubers-Blind for Megaminx edges

Only one alg for all your problems! And even better: it can be used for Gigaminx-midges, too!

The basic concept is to solve one edge at the time and bring it to the front first with D+ or D++ and so on. The first edge is solved with <alg>, the second with <alg'>. It is therefore very useful to memorize in pairs.

Shengshou color-sheme: white on top, upper half green in front, to the right red-blue-yellow-purple, lower half starting with brightyellow (under green and red) and to the right pink-brightgreen-orange-brightblue, grey bottom.

Memo is layer by layer counterclockwise (looked from above) starting top-front with 1 (white-green) and counterclockwise to 5 (white-purple). Next "layer" starts with 6 (purple-green), then 11 (brightblue-green), 16 (brightyellow-green), 21 (brightblue-brightyellow) and 26 (grey-brightyellow).

11 and 16 are in the same „horizontal layer“, but they need different setups move, so I take them as different „memo-layers“. Look at the basic concept: 16 (brightyellow-green) and 17 (pink-red) are the „same“, because I can bring the 17 with a D- to the front (at the place of the 16) - after that, they have the same setup.

Flipped edges: I add a 30, if the wing has to be flipped. But when is it flipped? This rule you can make by yourself. I have 1, 6, 11, 16 as not flipped, if green is in front (be careful: the green-red is number 7 and belongs to the red side!), 21 with brightblue on the left and 26 with the grey side down. The highest number is the 60: blue-grey.
Another, maybe better explanation: „the first color of a color-pairs (like I called them in the memo above) is the side, where you have to look at while memorizing“.
(Sorry, I have huge problems with the orientation and solve 3x3x3 bld with orientation first. I haven’t found the trick yet and so I don’t know, whether someone can understand my explanations - just find it out by yourself!)

The alg (found it in this forum):

L' R F2' L R' U' L' R F2 L R' U

and its inverse:

U' L' R F2' L R' U L' R F2 L R'

The buffer is 5 (white-purple), the target 16 (brightyellow-green) and the helper 1 (white-green).

If you have the pieces-pair brightblue-brightyellow,blue-red (21,38), it goes like this:
setup21 alg setup21' setup38 alg' setup38'

Exception: if a top-edge is involved, you will take this as the helper and will only need one alg and it is used as a normal commutator (see below).

Here are my setups:
6: U2' F2' U2
36: D+ U2' F U2
11: D+ DL'
41: DR DL
16: F5
46: D' DL2 DR
21: D+ DR2'
51: DR
26: D- DL' DR
56: DL2
All the others need only some D+ before and some D- afterwards.

Example for a top-edge: pair (3,53):
D-- (bring 53 to front, to 51)
DR (setup for 51)
U' (bring buffer to 1)
L' R F2' L R' (first half of alg)
U2' (bring the 3 to 1 - it is now in position 4!)
L' R F2 L R' (second half of the alg)
U2' (bring 5 to 5 again)
DR' (setup' for 51)
D++ (bring 53 back to 53)

Example for a top-edge: pair (51,3):
DR (setup for 51)
U2 (bring 3 to 1)
L' R F2' L R' (first half of alg)
U2 (bring the buffer 5 to 1 - it is now in position 3!)
L' R F2 L R' (second half of the alg)
U (bring 5 to 5 again)
DR' (setup' for 51)

Ok, if both edges belong to the top (e.g. 2,3), you can give up this attempt … or bring the 2 to 16 with U F2 U', then solve it like a 16,3 and do a U F2 U' to bring the 2 back.

A top-edge flipped?
Setup the other edge of the pair, then flip the top-edge without affecting the buffer 5 and the target 16. The easiest way for the 33 will be BR' R' and you will solve it then like a 2 (and don't forget to bring it back afterwards).

If you have a edge in place but flipped, e.g. 51, just solve 51,21 at the end.

Now look again at the alg: there is a U at the end and alg' begins with an U' (that is clear, because it is the inverse). You will only need this U and U', if one or both edges of the pair have setups that affect the top-face (only 6 and 36). Otherwise they cancel each other.

(Another idea for flipped top-edges: solve them without flipping and flip them at the end in pairs. The alg would be e.g.:
L F R' F' L' U2' R U R U' R2' U2 R
from https://sites.google.com/site/permuteram.../old-p-bld
But this doesn’t work for the Gigaminx. When I made that I hadn’t thought of the easy way I described above. I solved it at the end by bringing one edge to 16, executed the alg, flipped it, executed alg' and brought it back to the top).

 

[Jacck]

OK, next would be a discription of the corners, but I think, most of you should get it with the explanation for the Gigaminx - if you were able to understand, what I have written