• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 30,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Full PLL for Megaminx!

Joined
Aug 10, 2007
Messages
1,693
Likes
4
Location
Kauparve, Hejde, Gotland, Sweden
WCA
2005GUST01
YouTube
Kubbuth
Thread starter #1
Amoung crazy cube method projects we are proud to present full PLL for the Megaminx :)

http://sites.google.com/site/permuteramera/

There are still algs to add, some just needs to be mirrored and or inversed from one that is already there, some we have just not gotten to add yet but for some we lack a short solution...

But there are solutions, algorithms for all cases in 2 gen (RU and mirrors) but some are awfully long, 20+ turns. We are still looking for solutions for the rest and if you got a 3 gen or more alg for some of those situations that is not longer than 16 turns, then palease share.

The images we made from the cases may look a little silly but it is much easier to remember the case if you have a picture of it's shape rather than positions for induvidial pieces, then I played games inventing names for them, some are really clever but mabye not all of them :p


We don't seriously expect anyone to learn all cases, we did it most because it is a fun project and as far as we know nobody did it before us. But still, it is a great resourece for those who like to improve their LL times.
 
Last edited:
Joined
Apr 6, 2009
Messages
1,753
Likes
11
Location
Sweden
WCA
2010JERN01
YouTube
jernqvist
#2
So how many cases are they? I guess the fact that both edges and corners always are in a even permutation makes thing a bit easier, both in respect to the number of cases and recognition (you only need to know where 3 corners and 3 edges are). What's the average move count compared to other methods?
 
Joined
Dec 18, 2007
Messages
7,830
Likes
27
Location
a <script> tag near you
WCA
2006GOTT01
YouTube
qqwref2
#3
Heh, very cool to see that this is actually up :p A long time ago I had the idea of making full OLL and PLL, and I ended up calculating the number of positions and making about 20 algs (didn't use ksolve, just did it by hand :p) so it's cool to see someone is going to go all the way with this. I strongly suggest working on megaminx OLL too if you can, it's about the same number of algorithms and many of them can be adapted easily from 3x3 OLLs.
 
Joined
Aug 10, 2007
Messages
1,693
Likes
4
Location
Kauparve, Hejde, Gotland, Sweden
WCA
2005GUST01
YouTube
Kubbuth
Thread starter #4
Stini calculated the number of OLL's but I cant remember the exact number, but some 19x it was. Mabye we will see that too, it has been discussed :)

No, we did not use Ksolve but Johannes Laries Meganinx 2g solver that Stini modifyed a bit and he is considering expansion to 3g (RUL or RUD = EO perserved that is easiest). The rest of the algs we got was found by hand or we reused 3x3 algos.

Irontwig, 152 cases and of those 10 unusual. Average move count is not calculated yet, we like to add more algs first... we still find more of them.
 
Joined
Jul 21, 2008
Messages
150
Likes
0
Location
Finland
WCA
2007TIIN01
YouTube
teemutiinanen
#5
There are 260 OLL-cases in total (if inverses and mirrors are considered separate cases). Only the solved case has the probability 1/1296 and the rest have probability of 5/1296.
 
Joined
Jun 29, 2008
Messages
1,802
Likes
0
WCA
2008COUR01
#6
If it can help, I know the following CPLLs (both 2-gen and 15 moves):
Q case: y L2 U2 L'2 U L2 U' L'2 U L2 U' L'2 U L2 U2 L'2
Q' case: y L2 U'2 L'2 U' L2 U L'2 U' L2 U L'2 U' L2 U'2 L'2
Other than that, I don't do OLL-PLL on the megaminx so I haven't searched very far in that direction...
 
Joined
Mar 28, 2006
Messages
1,341
Likes
12
#8
If anyone else wants my 2-gen solver, the C source code is here: http://laire.dy.fi/t/minx_UR.c.txt. If it makes no sense or you can't figure out how to use it, don't ask. :p I wrote it a while ago for this thread: http://www.speedsolving.com/forum/showthread.php?t=6305.

With symmetries, I counted 148 OLL cases. How much using inverses saves depends on the exact algs used.

And for 1-look LL, in case anyone cares, I got 186,632 cases, 93,316 with mirrors, and 47,148 with inverses as well. (Wouldn't surprise me if these are wrong.)
 

jazzthief81

Premium Member
Joined
Jul 17, 2007
Messages
301
Likes
0
WCA
2003VAND01
YouTube
jazzthief81
#9
And there are programs for Megaminx LL. that go up to 5g (URFLB)
Yes, I developed a 5-gen LL solver for megaminx about 2 years ago but I never polished it off and published it because I lost interest. It works really well and a few people have been using it successfully.

This is a nice project that you guys are doing! Maybe it will inspire me to put my solver online soon.

But at the moment I have some other things going on that I can't tell you about but that will be revealed soon. It's also cubing/programming related. :D
 
Joined
Aug 15, 2007
Messages
493
Likes
2
Location
Fyn, Denmark
WCA
2006BUUS01
YouTube
henrikthedane
#10
And there are programs for Megaminx LL. that go up to 5g (URFLB)
Yes, I developed a 5-gen LL solver for megaminx about 2 years ago but I never polished it off and published it because I lost interest. It works really well and a few people have been using it successfully.

This is a nice project that you guys are doing! Maybe it will inspire me to put my solver online soon.

But at the moment I have some other things going on that I can't tell you about but that will be revealed soon. It's also cubing/programming related. :D
I would think you should, its a great program. :D
 
Joined
Aug 10, 2007
Messages
1,693
Likes
4
Location
Kauparve, Hejde, Gotland, Sweden
WCA
2005GUST01
YouTube
Kubbuth
Thread starter #11
Thank's for the links Henrik.

Lars I really like to see your solver, go for it! :)

Possible to change B to D(FR)? It is more usefull

Why I think so? take this case and try it in your solver to see if it finds a shorter one (possibly but mabye not) : R2' D U2' R2 U' R2' U R2' D' R' U' R2' U2' R2 [U2]


TMOY, thank's for the algs, I actually found the same ones, I just have not listed them yet :p
 
Last edited:

jazzthief81

Premium Member
Joined
Jul 17, 2007
Messages
301
Likes
0
WCA
2003VAND01
YouTube
jazzthief81
#12
And there are programs for Megaminx LL. that go up to 5g (URFLB)
This is a nice project that you guys are doing! Maybe it will inspire me to put my solver online soon.
I would think you should, its a great program. :D
Lars I really like to see your solver, go for it! :)
Here it is!

http://qblog.be/2010/02/last-layer-megaminx-solver/

I'm glad I finally published it. I've been putting this off for such a long time. Thanks for giving me the kick in the back, guys! :D
 
Joined
Aug 10, 2007
Messages
1,693
Likes
4
Location
Kauparve, Hejde, Gotland, Sweden
WCA
2005GUST01
YouTube
Kubbuth
Thread starter #14
And there are programs for Megaminx LL. that go up to 5g (URFLB)
This is a nice project that you guys are doing! Maybe it will inspire me to put my solver online soon.
I would think you should, its a great program. :D
Lars I really like to see your solver, go for it! :)
Here it is!

http://qblog.be/2010/02/last-layer-megaminx-solver/

I'm glad I finally published it. I've been putting this off for such a long time. Thanks for giving me the kick in the back, guys! :D
Gee man, thank's, it will come in wery handy for us :)
 
Joined
Apr 6, 2009
Messages
1,753
Likes
11
Location
Sweden
WCA
2010JERN01
YouTube
jernqvist
#15
Kenneth, you need to learn to spell "very" and "orca". I have been thinking about megaminx LL some, and I think for fewest moves OE-PE-CLL would be best (CLL mostly solved with two 3-cycles or something like that) and for speed OE-PEOC-PC might be the best because all steps have pretty easy recognition, OE is really short and PEOC is of course 2-gen-able. I'm not sure about the number cases, probably a bit too many for anyone who's not completely obsessed at getting faster at megaminx LL, since you you would only gain a few seconds with good recall and execution of hundreds of algs.
 
Joined
Aug 10, 2007
Messages
1,693
Likes
4
Location
Kauparve, Hejde, Gotland, Sweden
WCA
2005GUST01
YouTube
Kubbuth
Thread starter #16
Ya ya, I newer spell 100% because I newer use a spell checker, I humanize it =) Well, wery I know but I keep confusing w and v all the time, but only if I don't proof read, enough of that...

PE-OC are more cases than PLL but yes, for recognition it is a fine method.

You will have a horrible move count if you solve edges and then CLL in two steps, only that will have a nice 25+ moves on average and the number of cases is great, only for 3x3 it is some 80 (the worst LL case of them all for 3x3 is within this group, 16 HTM, H no permut, and nobody uses this method exept for some ZBLL solvers).

Now when there is a solver I'm planning a 3LLL that is more adapted to Megaminx than the normal 3x3 approches we use now. I have no idea how it will look, it is a later question, first I will try to compleate the algs for the PLL pages.
 
Last edited:
Joined
Apr 6, 2009
Messages
1,753
Likes
11
Location
Sweden
WCA
2010JERN01
YouTube
jernqvist
#17
Then why don't you get a spell checker? Unless you're using some super obscure browser there should be one for it.

Sure the cases where all corners permuted, but not oriented sucks, but you only get those 1/(5*4*3)=1/60 of the time, while all 5-cycles and double swaps can be done with 2 3-cycles for 14-24 moves (I just guessed that the worst 3-cycles requires 12 moves and that there's a case which requires two of them, correct me if I'm wrong). And when you go for fewest moves you have the time to be cautious and avoid nasty cases.

I don't see how you would be able to do 3LLL with a reasonable amount of algs (about the same number for 2LLL for 3x3) without some sort of LL preparation. If you make an oriented 2x2 block then you've greatly reduced the number of OLL cases and a 3LLL is possible (potentially 2LLL with much too much work). I would say that worst thing about megaminx LL seems to be that you recognition is hard if you don't completely know your colour scheme and that you don't have any slices to quickly solve edges. Oh, and the greater number of pieces doesn't really work too well. Well, I've rambled on enough now.
 
Joined
Dec 18, 2007
Messages
7,830
Likes
27
Location
a <script> tag near you
WCA
2006GOTT01
YouTube
qqwref2
#18
I think you could do pseudo-3LLL:
- when on last slot, first create a 2x2 block on the last layer.
- expand that to a 2x3 block.
- insert a corner/edge pair, similar to the Tripod method.
- solve last 5 pieces at once (# of total cases: 2*3*9 = 54).
 
Joined
Feb 11, 2008
Messages
657
Likes
1
Location
Malaysia
WCA
2009AZRA01
YouTube
amplify943
#19
I think if you remove the possibility of getting a 5-cycle of corners, the number of cases will be reduced and length of algs would be shorter. But I wouldn't know because I'm sort of clueless on these stuff. Just my 2 cents.

Perhaps during CO(assuming you would do EO->CO->PLL) you could cycle the corners to avoid 5-cycles?

EDIT
ninja'd by qq. He explained it better.
 
Joined
Aug 10, 2007
Messages
1,693
Likes
4
Location
Kauparve, Hejde, Gotland, Sweden
WCA
2005GUST01
YouTube
Kubbuth
Thread starter #20
The way I solve LL as now is VH-F2L for EO and partial MGLS (mostly sexy moves) for at least two corners, the ones next to the pair. Then I put the pair down and I get 6 diffrent OLL's, then PLL in as many steps I need. Working and pretty fast but turn intensive, I get like 40-50 for the whole LL if I don't know the PLL, else 30 something.
 
Last edited:
Top