Old bloke needing help with LL Algs.

[pete]

I don't know why I subscribed to "speedsolving.com" forum but I guess there wouldn't be any "slowsolving" forums ?

Anyway, here is the deal - my wife decided to buy me the Rubik's cube for Xmas and that worked out to be almost exactly 25 years since I've seen the thing last time.

Back in the old days I remember we used to do layer by layer with LL being a 4-step solution and it was possible to solve the cube on average in about 30-40 secs if I recall correctly.
For obvious reasons I don't remember any of the algorithms we used to have
so at the moment I'm solving the cube by intuition only :
Step 1.
First I do all 6 crosses, meaning solving all 12 edges with correct permutation and orientation (I start with white cross, then 3 edges in the middle layer and I use the last unsolved edge of middle layer as a buffer to solve my LL cross which on my cube is yellow). This step takes me about 30 moves.
Step 2.
Then I move on to the corners and I solve those with what I think is these days called 3-cycle commutators.

It depends on how lucky I get in step 2 but on average it takes me +/- 55 secs to solve the cube with this method (with zero algorithms learnt).

I'm now trying to modify this approach so that first part of of Step 2 solves 4 corners of any face (this takes additional 20 moves but with carefull planning in step 1 it often is as few as 10 moves), this would leave me with the typical F2L completed including completed LL cross (cross correctly permuted) after a total of 40 moves (with carefull planning in Step 1) or 50 moves (with no planning ahead) and up to this point everything is solved by intuition.

From here onwards one can continue with commutators to finish the cube but I find last 4 corners taking too long in some bad cases.

And here is where I decided to change the method and learn some new algorithms to solve those last 4 corners in the most efficient way without having to rely on some lucky cases.

So I want to introduce Step 3
and I would really appreciate if someone can point me in right direction because after visiting dozens of other rubik related websites I'm unable to find algorithms that can solve orientation of my LL corners without affecting my solved edges. My way of thinking is to break Step 3 in to :
Step 3a) orient all LL corners correctly
then
Step 3b) permute all LL corners correctly

This is surely possible to do in 20-30 moves ??

I believe there are only 7 possible scenarios (2 of which are symetric) to solve orientation of all LL corners in this situation, followed by 7 possible scenarios (with 1 symetric case) to permute LL corners (after orientation was already completed in previous step).

I can use 7 of the Fridrich PLL algorithms to permute those corners but I'm lost at how to approach my Step 3a and I just can't find algorithms that don't mess up my solved LL edges.
For this reason I can't used selected Fridrich OLL algorithms because LL edges do get misplaced.

OK, I know many would say what a waste of time trying to solve the cube with this crazy method but I'm happy to use this method for what seems to be a 60-70 move solution if I can find and learn just 14 algorithms (or in fact only 11 if I don't count symetric cases).

at my age I don't have a time nor motivation to learn 100 Fridrich algorithms and I'm quite happy to aim for under 40 secs average which I'm sure those 11 required algorithms may help me to achieve.


Btw. is there a system somewhere that works with edges first (all 12 edges) ?
perhaps somebody already figured out better and more efficient way to solve the cube with all edges completed first.

Cheers
Pete

ps: is there a software that could figure out my required algorithms ?
I've heard of Cube Explorer but I'm not sure whether that would be the right tool.

 

[Lofty]

sorry I don't actually have there algs but myself... but there are a few options you can use to generate them.
The easiest may be Ron's CubeSolver on speedcubing.com. It has a picture of the cube and you just set the stickers how you want it to solve and it does the work for you.
ACube is also available but is a bit more difficult to use and I don't know how to use it for the algs you want but it is a much faster program.
Also I am pretty sure there are far fewer Permutation algs. the H, A (and a symmetric) and an E. The A I know is just a commutator with one move that cancels making it 9 moves instead of 10. And the E is just a combination of two orientation algs.
I can generate the orientation algs for you if you would like. I'll edit my post with them. Do you have a preference for which face turns to use or just the shortest algorithm or of course you can do it yourself if you want.
"Headlights" two oriented corners two facing the same way. U case
R2 D' R U2' R' D R U2' R
R' U L' U R U' L U2 R' U R
two oriented corners on the side near you and a sticker facing on each side on the back two corners. T case
B'R'FRBR'F'R
L case:
Inverse of T case
Pi Case:
Sune case (three corners not oriented): known as a niklas cycles three corners but not edges, also a commutator.
L' U R U' L U R'
H case: (racecar)
F R U R' U' R U R' U' R U R' U' F'

 

[pete]

thanks for the help, I appreciate that.

unfortunately I didn't have a time yet to figure out ACube command line usage.

I've also downloaded Cube Explorer but unless I missed something it appears it's unable to find any algorithms if I specify only 1 sticker on corner cubies.

looking at the Cubesolver now to see whether it accepts 1-sticker cubies...

typically I prefer R-U-F-L-D-B in that order, i.e. if can get away with turning just 2 faces then R + U, if the algorithm requires 4 faces turns than R+U+F+L and so on.

 

[Lofty]

I'm giving you most of the algs I use myself in certain cases. Since you are calling symmetric cases as the same I'll assume you are good at moving the algs around the cube so you can try to find an angle or fingertricks yourself. Most of them are very short 8 move commutators so even if the turns aren't just the ones you like they the shortness of it makes up for it.
as for permuting your corners you can find those algs on any PLL list.

 

[pete]

thanks Lofty,
these are perfect for me, it's only the T case algorithm that feels
bit awkward in my hands so I turn the cube clockwise around
the "y" axis before I start executing it (so that B becomes R,
R becoms F & F becomes L). hmmm, now that I think of it I should've
perhaps turned the cube around "x" axis and work with L , U & R - there are just too many options

I wouldn't worry about fingertricks, my fingers "don't support" that feature

 

[TimMc]

it's only the T case algorithm that feels
bit awkward in my hands

Try:
(R U R' U') (R' F R2 U') (R' U' R U) (R' F')

See:
http://www.randelshofer.ch/test/ShowCubePlayer/RubiksCube.html?RUR-U-R-FR2U-R-U-RUR-F-

Personally I prefer using LRUF faces and try to avoid DB at all costs

Tim.

 

[pete]

it's only the T case algorithm that feels
bit awkward in my hands

Try:
(R U R' U') (R' F R2 U') (R' U' R U) (R' F')

Tim.

hey Tim, different type of T case (corners only).

but the one you posted I like too, i was using it last week a lot when
experimenting with blind solving.


btw. isn't it sad Australia is the only continent with no official competitions ?

 

[guusrs]

Hi Pete,

I'm an old bloke myself, started in 1981, but I use my own method with 100+ formula's. 50 - 55 moves while speedsolving. Fridrich for F2L (actually Fridrich probably learned this system from me, not sure) then LL-corners then LL-edges.

For your system check http://www.ai.univ-paris8.fr/~bh/cube/solutions_c2.html for all LL corner algorithms and unaffecting edges, average ±12 moves. If you don't want to learn them all just pick the shortest for orientation and learn the 3 permutation algs. I Also advice you to learn al 3-corner algs, 6 algs of 7 - 10 moves, you can use it in ±30% of all LL cases. Even better: try to solve LL corners as a combination of two (sometimes 3) of these 3-cycle algs. Average of about 16 moves. So only 6 algs.

However if you don't like learning algs at all try Ryan Heise's method (www.ryanheise.com), nice block/pairing mechanism, almost no algs and <50 moves. Ok for 20sec+ solvers, not for 15sec- solvers.

Good luck

Gus

 

[TimMc]

btw. isn't it sad Australia is the only continent with no official competitions ?

Yeah x.x

Although, it's still popular among people when you get out there and are speed solving in public

Not many people can successfully act apathetic when you're spinning the faces at 3-4 turns per second and solving it Sub 30 :rolleyes:

There must be a few hundred people on the east coast of Australia alone that can solve a Rubik's Cube in less than a minute... It's just that nobody wants to risk a few hundred$$$ to try and setup an event nationally (at least no individual that I know).

*pokes Crown and Andrew for sponsorship*

Tim.

 

[Swordsman Kirby]

btw. isn't it sad Australia is the only continent with no official competitions ?

Yeah x.x

When Antarctica has a comp, then you can start whining.

 

[Leviticus]

well hey at least if there was a comp it would be on the east coast, im on the west, but if it was big enough i would most likley go, i just remebered where going on a holiday to melbourne this year, so hopefully

 

[pete]

Hi Pete,

I'm an old bloke myself, started in 1981, but I use my own method with 100+ formula's.

For your system check http://www.ai.univ-paris8.fr/~bh/cube/solutions_c2.html for all LL corner algorithms and unaffecting edges, average ±12 moves. If you don't want to learn them all just pick the shortest for orientation and learn the 3 permutation algs. I Also advice you to learn al 3-corner algs, 6 algs of 7 - 10 moves, you can use it in ±30% of all LL cases. Even better: try to solve LL corners as a combination of two (sometimes 3) of these 3-cycle algs. Average of about 16 moves. So only 6 algs.

However if you don't like learning algs at all try Ryan Heise's method (www.ryanheise.com), nice block/pairing mechanism, almost no algs and <50 moves. Ok for 20sec+ solvers, not for 15sec- solvers.

Good luck

Gus

Did you keep in shape for that long ? Or have you started again after a long break and had to re-learn all those 100 algs ?

thx for the link, I had a look - most of it was already covered by Lofty in previous post. I'm already learning first 6 algs for CO that don't affect edges but I still struggle to find suitable alg for case 7 which is the one where all corners need to get orientated (not the race car, the other one - two top layer stickers facing side and the other 2 facing forward), I found one with
Cubesolver but I don't like it : R U' R2 D' R U2 R' D U2 R2 U R2 U R
(executing last 6 moves is a pleasure, but the first 8 aren't that friendly)

don't know what algs are you referring to, those "all" 3-corner ones. I only have 2 cases that i need to learn to orient all 3 corners (unless i misunderstood something).

I had a look at Heise method as per your suggestion. Basically with my intuitive method I can get to the same point before his last step (last 3 corners to be orientated and permuted, although sometimes I'm left with 4 corners instead of 3). I can get there within 40-50 moves but then I start to struggle
while trying to finish the cube because I just use simple commutators to permute my remaining corners but I'm typically left with 2 or 3 or 4 corners correctly permuted but not orientated and it takes for ages to twist all corners. Sometimes I get lucky and all corners are already orientated and that's the only time I get sub 40 secs solve, otherwise with all those crazy corner twists I average 55 secs.

I like Heise approach in that last step actually - when the simple commutator doesn't resolve last remaining corners, he uses conjugates which typically involves just 1 or 2 setup moves, then applies commutator and reverses setup moves. this way instead of starting with simple commutator then twisting i.e. 3 corners and taking 35 moves (8 commutator moves + 27 twisting moves),
he would simply apply conjugate and completes the cube within 10 moves !

I like the logic behind his method but at this stage i'm not sure if it's possible to quicky recognize those conjugates. i can see simple commutators right away, but not if I have to do setup moves first. i think i would waste at least 30 seconds just to find correct setup moves by trial and error.

this is why i have decided to do my last 3-4 corners by orientating corners first and then simple corner permutation, not many algorithms to learn with
a guarantee that I finish within 20 moves in my last step (instead of doing endless corner twist and waiting for luck).

as i said before, i'm happy to do sub 40 secs solves (that's where i was 25 years ago) and i don't expect to get any better, i just waste too much time in my first 40-50 moves, i'm unable to execute those under 30 secs.

cheers

 

[pete]

Not many people can successfully act apathetic when you're spinning the faces at 3-4 turns per second and solving it Sub 30 :rolleyes:

There must be a few hundred people on the east coast of Australia alone that can solve a Rubik's Cube in less than a minute... It's just that nobody wants to risk a few hundred$$$ to try and setup an event nationally (at least no individual that I know).

*pokes Crown and Andrew for sponsorship*

Tim.

i think in Australia people are amazed when someone can solve the cube, never mind how long it takes
doing it under 30 secs as you do, people must think you're a genius.

maybe worth investigating trying to have a competition as a piggyback event with something else, i.e. when there is a big Poker tournament somewhere.
this way if only 5 people decided to show up, it wouldn't matter that much, competition would still go ahead and the Poker people would have a fun spectating during their half hour break or something like that.

or perhaps trying to combine it with some Chess tournament, just ideas...
i'm not sure how this works actually, whether WCA delegate needs to be present and whether Australia has any members on WCA board.

well hey at least if there was a comp it would be on the east coast, im on the west, but if it was big enough i would most likley go, i just remebered where going on a holiday to melbourne this year, so hopefully

could do it in Alice Springs so that it's approximately same distance from all the coasts

 

[Leviticus]

Yeh but some people arnt as dedicated and i would say majority of people that cube in Au do it on the east coast, so i say have it there, if it was big i would go

 

[abbracadiabra]

thx for the link, I had a look - most of it was already covered by Lofty in previous post. I'm already learning first 6 algs for CO that don't affect edges but I still struggle to find suitable alg for case 7 which is the one where all corners need to get orientated (not the race car, the other one - two top layer stickers facing side and the other 2 facing forward), I found one with Cubesolver but I don't like it : R U' R2 D' R U2 R' D U2 R2 U R2 U R
(executing last 6 moves is a pleasure, but the first 8 aren't that friendly)

You might want to try this alg, Pete.

Hold the cube so that corners you wish to orient are on the top layer, and the left layer has two of the top layer stickers pointing towards the left hand side, and the right layer has two top layer stickers pointing towards the front and the back. Then apply:

R U2 R2 U R2 U R2 U2 R

This move does affect edge permutation, but it doesn't affect edge orientation or corner permutation. The move you posted permutes the corners and the edges, as well as orienting the corners.

Hope that makes sense, and hope it helps.

By the way, what is your definition of "old"?

 

[hait2]

as a result of this thread i've finally learned acube (although i can't figure out how to gen rRU algs yet.. any help?)

here's my contribution, it's a bit long (15 moves) but it's a 2gen that flows a bit nicely i think - i really like the last part

(R U' R) U R' (U R' U' R2 U R' U' R U' R2)

(this is for the orienting corners without caring about corner permutation ,but preserving edges.. i think that's what you asked for, if not clarify please xD)

edit: i tried RMU algs but no dice.. i thought that would emulate rRU with moves like MR etc.. a lil' help please?:O

 

[pete]

You might want to try this alg, Pete.

R U2 R2 U R2 U R2 U2 R


By the way, what is your definition of "old"?

I would love that algorithm but can't use it unless I learn all 21 PLL cases.
i'm only learning 4 PLL cases where edge permutation is already solved and this
algorithm would mess up my edges (i solve LL edges by intuition when i'm doing my F2L).


definition of old, hmmm i guess that's relative.
if i go to bowling club where you find 90% of pensioners then i'm relatively young.
when a 12 year old visits kindergarden then he/she is relatively old.

i think i fit to any of these categories :
- old enough to have grandchildren
- more than double the average age of members on this forum
- brain outruns slow fingers
- old enough to remember the lyrics of Depeche Mode '81 hit "Just can't get enough"
-....

 

[pete]

as a result of this thread i've finally learned acube (although i can't figure out how to gen rRU algs yet.. any help?)

here's my contribution, it's a bit long (15 moves) but it's a 2gen that flows a bit nicely i think - i really like the last part

(R U' R) U R' (U R' U' R2 U R' U' R U' R2)

(this is for the orienting corners without caring about corner permutation ,but preserving edges.. i think that's what you asked for, if not clarify please xD)

edit: i tried RMU algs but no dice.. i thought that would emulate rRU with moves like MR etc.. a lil' help please?:O

thanks for that !
that's heaps better than the one i was trying to learn, this one is easy
for my hand coordination