## 6531Re: Working on learning a 3-look LL

Expand Messages
• Sep 5, 2003
Simple solution to your 4 corners twisted problem: double-sune and
bruno (R'U2R2UR2UR2U2R'). DanK/Mirek's Z Perm is the way to go, look
up DanK video on it... so you don't end up getting a bad habit of
the brute force alg.

I'm going through the same phase, teaching a friend of mine that's
at about 50s. (Just last night planning how to line up the next 8
PLLs: he knew the 3 edge ones, so I gave him the T and Y to start
off with... A and V might be the next step).

-Doug

--- In speedsolvingrubikscube@yahoogroups.com, jasmine_ellen
> What a very detailed reply Ryan!
>
> Here's where I'm up to at the moment... I've learnt several LL
edge
> algs -- ones that mess with the corners (which doesn't matter if
LL
> edges are oriented first). I already knew several LL edge algs
that
> preserve the corners -- I guess these are useful if the corners
are
> already done when you get to the LL.
>
> I already know the Sune and its mirror (can't imagine there'd be
any
> cuber who doesn't know this alg?!). I've learnt the algs for the
> three LL corner orientation positions where 2 are corrent and 2
> require twisting. Still need to learn the 2 algs for when all 4
> require twisting.
>
> I'm aiming to learn the 13 PLL if my enthusiasm lasts that long.
In
> case it doesn't, I planned the order that I would learn the algs.
> First I revised the algs that just move the edges since. I knew U
and
> H, but for some reason was a bit iffy with Z. After that, I just
> ranked them by 'probability of occurrence'. So, of the remaining
10,
> I've memorised the algs with a frequency of 1/9 or greater (A, R,
J,
> G). So, 7 down, 6 more to go.
>
> I think I need to take a break from new algs and spend some time
> getting the newly memorised algs firmly planted in my head. I'm
> finding that I can remember them without a problem, but I'm still
> taking a bit of time recognising the patterns. I'm not concerned
> though because I know that practise will fix this!
>
> Jasmine.
>
> --- In speedsolvingrubikscube@yahoogroups.com, Ryan Heise
> <rheise@p...> wrote:
> > On Mon, Sep 01, 2003 at 11:48:01AM -0000, jasmine_ellen wrote:
> > > I remembered that DanK had something like this on his site
> a
> > > beginner method, an intermediate method and an advanced
method.
> This
> > > is actually what I was after. Who else has something like this
on
> > > their site?
> >
> > I'll describe the beginner and intermediate methods that I once
> used.
> > They are simplications of the Fridrich last layer system. You
need
> to
> > first understand how that system works:
> >
> > http://www.ws.binghamton.edu/fridrich/system.html#last
> >
> > This general strategy, orienting first, then permuting, is good
> because
> > it is very easy to recognise the positions quickly. But instead
of
> > orienting all the pieces at once, you just orient the edges
first,
> and
> > the corners second.
> >
> > BEGINNER METHOD
> >
> > First, orient the 4 edges (ie. make a cross). There are only 4
> patterns,
> > and one algorithm to solve them all! To view this, you need to
use a
> > fixed width font:
> >
> > --- --- -x- -x-
> > -x- xxx xx- xxx
> > --- --- --- -x-
> >
> > For each pattern, apply the algorithm "R'U'F'U F R" and it will
> take you
> > to the next pattern. Eventually you will end up with a cross.
> >
> > Improvement: notice that if you apply the algorithm when you
> > have the cross, you will end up with a line. That means that you
can
> > jump immediately from the 2nd pattern to the 4th pattern by
> applying the
> > algorithm backwards! So, this step should take an average of 6
> moves.
> >
> > Next, orient the corners. There are only 6 patterns apart from
> solved.
> > Look through this complete list, and find all the patterns (and
> > algorithms) that have a cross already formed:
> >
> > http://www.ws.binghamton.edu/fridrich/Mike/orient.html
> >
> > For the case when 3 corners need to be twisted, I recommend the
> Sune TM
> > algorithm from Petrus. See the applet at the top/right of his
page:
> >
> > http://lar5.com/cube/fas6.html
> >
> > I think it's useful to learn to twist them forwards and
backwards:
> >
> > - R U R'U R U2 R'
> > - R'U'R U'R'U2 R
> >
> > To see that done fast, have a look at the quicktime movies at
the
> bottom
> >
> > http://lar5.com/cube/speed.html
> >
> > I recommend this algorithm because one of the other orientation
> cases
> > can be solved by just applying this algorithm twice. To know
which
> case
> > I'm talking about, apply this to a solved cube:
RU2R'U'RUR'U'RU'R'.
> >
> > Now, apply the sune (RUR'URU2R') twice:
> >
> > RUR'URU2R' + RUR'URU2R'
> >
> > Notice that the moves in the middle cancel out? So you can
actually
> do:
> >
> > RUR'URU'R'URU2R'
> >
> > which is an "optimal" solution.
> >
> > Next, permute the corners (yes, before the edges). Most of the
time
> > there will be a 3 cycle of corners. In this case, look at the 4
> sides of
> > the last layer. On one of the sides, the two last layer corners
> will be
> > matching in colour. Hold those two corners on the back side and
do:
> >
> > R'FR' B2 RF'R' B2 R2
> >
> > Now the corners should be solved. If you don't find any two
corners
> that
> > match, apply the above algorithm and that should result in a
> position
> > where two corners match. Don't worry, this case rarely happens.
If
> you
> > want, you can learn a special algorithm for that case by picking
the
> > most attractive one from:
> >
> > http://www.ws.binghamton.edu/fridrich/Mike/permute.html
> >
> > The "Y" pattern on that page has the desired corner effect.
> >
> > Next, permute the edges. Again, most of the time, there will be
a 3
> > cycle of edges. You need to learn to cycle the edges forwards and
> > backwards:
> >
> > - R2U FB'R2F'BU R2
> > - R2U'FB'R2F'BU'R2
> >
> > Sometimes, all 4 edges need to be swapped in opposite pairs.
That's
> an
> > easy case, so why not learn it:
> >
> > RLU2R'L' [U] R'L'U2RL
> >
> > The [U] means rotate the whole cube from the up side, 90 degrees.
> >
> > There's also a rare case where all 4 edges need to be swapped in
> > adjacent pairs. The algorithm's difficult so it's not worth
> learning.
> >
> > INTERMEDIATE METHOD
> >
> > First, orient everything like the beginner method. Then permute
> > everything like the full blown Fridrich method.
> >
> > How can you learn all the permutations? See my previous email:
> >
> > c
> >
> > There, I list which algorithms to learn first, and which
algorithms
> to
> > learn last.
> >
> >
> > You already know enough! It is not necessary to learn 40
orientation
> > algorithms. 6 is enough. Gilles Roux proved that it is possible
to
> > achieve sub-20 times with just these 6 orientation algorithms
and 13
> > permutation algorithms. He used the petrus method for the first
two
> > layers which gives you a cross automatically when you get to the
> last
> > layer. Using that strategy, those extra orientation algorithms
are
> not
> > needed.
> >
> > Ryan
• Show all 11 messages in this topic