Welcome to the Yahoo! Message Board for Speed Solving Rubik's Cube

Welcome to the speed solving rubik's cube club.
If you know how to speed solve then you can offer
hints to others who are just learning. If you don't
know how to speed solve then ask questions! That's
what this club is for.

In general there are two schools of thought about
speedsolving.<br>1. Know and learn many move sequences, so that
anything you encounter can be solved with very few
sequences.<br>2. Know and learn only a few sequences which can be
performed more quickly. Although you may have to apply
these sequences more often, and therefore make more
decisions, these decisions are easier to make so thinking
time is cut down.<br><br>These two ideas are
illustrated by the two main webpages about
speedsolving:<br><br>Jiri Friedrich's
page:<br><a href=http://www.ssie.binghamton.edu/fridrich/cube.html
target=new>http://www.ssie.binghamton.edu/fridrich/cube.html</a><br>Lars Petrus'
page:<br><a href=http://u1.netgate.net/~mette/lars/cubedude/
target=new>http://u1.netgate.net/~mette/lars/cubedude/</a><br><br>I am not
really a speedsolver (usually 85 secs or
so), and therefore much prefer Lars' method which has
few sequences to memorise.<br><br>Both pages above
use essentially a layer method (though Lars does the
first two layers differently). Are there any
speedcubers out there who do a corners first method like Minh
Thai did?<br><br>Jaap<br>Jaap's Puzzle Page:
<a href=http://www.org2.com/jaap/puzzles
target=new>http://www.org2.com/jaap/puzzles</a>

I use Jiri's method speed solving, or I know a
whole bunch of moves so anything that comes up I can
solve in a few moves. I have never tried the other
method, of learning fewer moves. I want to try to learn
Lars Petrus' solution but I've never really sat down
to do it. I can say though that decision time
doesn't take that long when you have to choose from a
larger number of moves that you've memorized. I've
broken them down into sections. For example the third
step of the solution I use is after the first two
layers are solved and the purpose of this step is to
orient all the bottom layer pieces so that the whole
face shows one color. There are 40 different
situations to choose from but my decision time is at longest
a little more than a second. Sometimes some of the
patterns are very similar so if you look for that similar
characteristic then you've automatically eliminated at least 30
of the situations if you see one of those similar
characterictics. The only drawback is that you have to practice
the moves a lot as they come up by the laws of
probability so the moves that are rare to come up you
especially need to practice or when they do come up you may
realize that you don't remember them so well which really
kills your time.<br><br>I've never heard how Minh Thai
solved the cube he used a corners first approach? Is
there anywhere to read up on that type of solution?
That sounds interesting.<br><br>I want to try to learn
Lars' solution so that I can have a basis of comparison
between the two main methods of speed solving, how many
moves does it take, on average, for his method? Jiri's
is about 55 or 56 moves on average.<br><br>Chris

I have a question for anyone out there that speed
solves. ilovemycue I know that you are faster than me so
maybe you can help me answer this. How can I solve the
first two layers faster? Yesterday I was timing myself
for solving just the first two layers, and once I
warmed up I could do it in 16-20 seconds on average,
with a few times at 14 seconds (not including any
lucky times). Does anyone have any tips for solving the
first two layers faster? I use Jiri Fridrich's solution
so preferably without having to learn a new solution
:-) It might also have been because I was tired (I
was taking my times right before I went to bed).
However after I practiced solving the first two layers my
times improved slightly, to about 24-25 seconds on
average for a period of a few minutes. Also should I use
any finger shortcuts and if so how often? If anyone
has any tips or hints or can help me then I'd really
appreciate it. Thanks in advance.<br><br>Chris

I also have good days (<23 avg) and bad days
(>25 avg). The difference is in the second phase (four
times a corner of the first layer and an edge of the
second). I think it is because you need to be really sharp
to immediately find the combination of those two
cubelets. On bad days I need relatively a lot of time to
find those two.<br><br>A method that also works fine,
is the following:<br>- solve the 4 edges of the
first layer (cross)<br>- solve 3 corners of the first
layer<br>- solve 3 edges of the second layer (using the not
solved corner of the first layer to "insert" the
edges)<br>- solve the last corner of the first layer together
with the last edges of the second<br><br>This method
will take about 5 or 6 moves extra, but you can focus
on solving one cubelet at a time. So it takes almost
the same time.<br><br>Good luck.<br><br>Ilovemycube

(note: most of this is from my home page)<br>I
use Jiri's method, and I usually do the first two
layers in 11-12 seconds. Here's what I recommend:
<br>solve the cube, stopping and writing down the number of
moves you took for each stage. <br>Try to solve as fast
as you can while still counting, that way it's more
like when you're speed-solving. Do this 50 times or
so, and then find out your average moves for each
stage. <br><br>I did this and here's what I got:<br>1.
(Cross) 6.0 moves<br>2. (4 corner/edge pairs) 27.28<br>3.
(orient) 8.74<br>4. (permute) 11.84<br>5. (total)
53.86<br><br>I think I can do better for #2. So make sure that
you are getting 27 or 28 for #2. <br><br>Suggestion
#2: when you are putting in your last pair in stage
#2, you can often put it in several ways. (like R U'
R' or R' F R F'). If you are careful, you can almost
always avoid the orientation cases with all edges
flipped.<br><br>Suggestion #3: One other important observation that I have
had is that during the first two layers, triggers
usually slow me down because they are too fast to watch
the pieces. I find my best times for the first two
layers are when I am doing really rythmic continuous
moves alternating between hands, at 3-3.5 moves per
second. <br><br>Suggestion #4: Another hugely helpful
observation that I've made several times this past year, and
then forgotten, is to really really try to look ahead
during the first two layers. This means that I force
myself to actively examine the whole cube, especially
the back side, while my hands are busy putting in the
current corner-edge pair. Whenever I remember to do this,
I seem to drop a whole second from my times. While
my hands are doing a move sequence, I am raising my
eyebrows and trying to peek around to the back of the cube
to see what pieces are there.<br><br>Suggestion #5.
As an exercize, try to solve the first two layers
without EVER changing the orientation of the cube. Do
this 100 times. This might be difficult at first, but
with some practice you will get used to it. Then take
these ideas and try to optimize your speed-cubing so
that you use as little turning of the whole cube as
possible, while still allowing some turning when it makes
the moves faster for your hands. The less you have to
turn the cube around, the easier it is for you to look
ahead.<br><br>Suggestion #6: It is also very important to minimize the
number of moves during your first two layers. There are
often shortcuts available that get rid of uneccessary
moves while you are inserting a corner-edge pair,
especially during the first few corner-edge pairs when you
have still have some "empty" slots left. Also, when
the last move of inserting one corner-edge pair is R'
and the first move of the next pair is going to be
R', try to anticipate this and just do R2.<br><br>Dan

I agree mostly with cubin4speed.<br><br>here are
some of the things that I have found most
helpful.<br><br>1. When finishing the cross make a very strong
effort to look around for your first corner edge pair.
This may require slowing down the last moves of the
cross. I have found that I can lose the most time here
if I am not on top of things.<br><br>2. from here
looking ahead is the name of the game. One rule to stick
with most of the time is: solve the first corner edge
pair that you find. If you see another pair that is
set up better just as you are moving into the pair
you saw first...changing could throw you off... just
track where those pieces are going and be ready for
them as soon as you finish the ce pair you started.
(exceptions may be the 3 move case vs. an 8 or 9 move case
(most of the others won't even mess up the second pair
very much)).<br><br>3. Really take advantage of empty
corner edge slots for putting pairs together before
putting them in the right place.<br><br>4. Practice a lot
going slow but with no stalls for
recognition.<br><br>5. Practice, Practice, Practice...This stage more
than any other just gets better and better the more
you do it.

I can usually solve the Rubik's Cube in 2-3 minutes.

I was thinking that it would be cool if we could
set up an averages board that people could post their
averages to. I think that this would be a pretty cool
thing. Does anyone know if we could set something like
this up with Yahoo Clubs?

Chris Hardwick has a great page set up for
posting
averages:<br><a
href=http://www.geocities.com/Area51/Zone/7362/rubik/recordtimes.html
target=new>http://www.geocities.com/Area51/Zone/7362/rubik/recordtimes.html</a><\
br><br>Also, I agreed with what rubixsux said about solving
the first two layers. Do you mind if I add your
comments to my home page? My address is in my
profile.<br>-Dan

I started out placing the cross on top, and then
solving top layers while holding cross on top. It was
hard to find the next pair of cubes that way. I am now
trying solving cross on top, flipping it over and
solving bottom up. What do others do? What's fastest?

There's not one answer.<br>Cubin4speed keeps the
cross at the bottom.<br>Some other very fast cubists
(below 20 secs average) keep the cross at the left, half
left half top, or even top. I keep the cross at the
left.<br>As long as you make sure that you have a good view
of what is going to be the top layer, either side is
OK. So you could keep the cross at the top, but then
hold the cube a bit backwards at the top.<br><br>I see
one big advantage of keeping the cross at the bottom:
you have the cube in the correct position for the
first part of the top layer. So you will need less
thinking time.<br><br>Both left and bottom have the
advantage that you can combine the moves of the right and
upper side in short cuts.<br><br>The best would be to
never turn the cube at all. So build the cross at the
bottom, and keep it there.<br><br>Good
luck.<br><br>Ilovemycube

How do you take advantage of empty slots to put
pieces together before placing them, as a few of you
have referred? I know it's probably hard to explain
and really easy to show.<br><br>And how do you
approach the situation where all 4 corner pieces are in
the wrong corners? This seems to really slow me down.

I'm also curious as to how people first started
memorizing the orientations. Did you draw them up to make
them most efficient for your hands, and then memorize,
or did you draw them so you could recognize the
patterns, memorize, and then change to fit your hands? Or
something else?<br><br>And when we talk of 40 cases;
shouldn't we say there are 57? Unless I'm missing
something, that's what you really have to memorize, right?

Well I had better show you a picture. So please
go to a part of my home page to be:
<a href=http://members1.chello.nl/~r.vanbruchem/corners1_edges2.html.
target=new>http://members1.chello.nl/~r.vanbruchem/corners1_edges2.html.</a><br>\
Now take a look at the second picture in the fourth
row. It shows two algorithms, one of seven moves and
one of six. But what if the corner at the front left
top (green orange white) would not yet be in its
place? Then it can be done in five moves: F'DF2DF'.
There are lots of situations in which you could win
even more than one move.<br><br>Good
luck.<br><br>Ilovemycube

There are 40 different cases of orientation in
the top layer. But if you include the cases which can
be formed by mirroring one of the 40 cases, then
there are 57 different cases. See my home page to
be:<br><a href=http://members1.chello.nl/~r.vanbruchem/top_layer.html
target=new>http://members1.chello.nl/~r.vanbruchem/top_layer.html</a><br><br>Abo\
ut memorizing the orientations: if I understand your
question correctly, you want to know how to start
memorizing the 40/57 algorithms.<br><br>You could start with
learning how to solve the orientation of the edges (3
cases) and then solve the orientation of the corners (7
cases), so in two steps.<br>Next thing would be to learn
the cases with 4 flipped edges (8 cases). This is
because there are only few cases, and they take the most
moves, if you do the orientation of the edges and the
corners separately.<br>Next learn the cases where the two
flipped edges are opposite each other.<br>Finally learn
the cases where the two flipped edges are next to
each other.<br><br>Hope this helps.<br>Good
luck.<br><br>Ilovemycube

Thanks for the ideas on this, and on the empty
slots, I appreciate it. I look forward to your website;
right now I couldn't get the pictures up, not done yet?
<br><br>When learning the orientations: Let's say you
recognize one of the patterns. But the required move is
awkward for your hands. Do you redraw the pattern so the
move is easier for your hands, and try to learn it
that way? Or do you continue recognizing the pattern
the first way, and then rotate the cube to make the
move easier for hands?<br><br>Not critical; I was just
curious as to how different people tackled this.
<br><br>Thanks again, <br><br>slowcuber

The address of my web page was not entered
correctly in my previous reply. You should remove the
closing "." in the URL.<br>I added my link to our club
links.<br><br>I think it is not a problem to learn the
orientation algorithms if they are awkward for your hands. In
the beginning I disliked moving the back side. Now,
using short cuts, it is one of my favorites.<br>Once
you have learned the algorithm, it is no problem to
start using the same algorithm but from a different
angle.<br><br>The easiest way to learn the algorithms is to use a
system to know which starting position to take. In that
sense it is important which angle to take.<br>But
again, you can easily switch later on.<br><br>Good
luck,<br><br>Ilovemycube

I found that the easiest way for me to learn the
moves was to print out Jiri's move list and to just go
down the list move by move and check them off on the
sheet when I had memorized it. He arranges the moves in
pretty good sequence for memorizing. When you're doing
it fast though I don't think it matters because no
matter how you memorized the moves it seems that if you
are used to doing each move then you can still decide
without losing much time. <br><br>I found out that a lot
of the times moves that felt bad for my hands got
better if I try it from a different angle. Just keep
trying different angles till you find the one that feels
the best. <br><br>By the way for those of you
wondering about my homepage I have no idea why it messed up
but it works now. I guess geocities' servers were
messing up.<br><br>I have a question about recording
record times. When is a time too lucky to consider? I
usually consider a time too lucky if it only takes one
algorithm to solve the third layer instead of the usual
two. Are there any other situations you should
consider too lucky or is it usually only the third layer
that you really look at to decide?<br><br>Also, on a
random note, recently I've been trying a move for fun
that I can do extremely fast. It has no purpose really
but it just feels good to my hands. The move is R' U
R' U' R U' R U It's kind of cool too because the
move can be repeated any number of times without
losing speed. I'm still trying to get my hands used to
it so I haven't timed myself but I'm pretty sure
I've done a few times in under 1.5 seconds. Does
anyone know any other moves that can be done really
fast?<br><br>Later,<br>Chris

here are some of my thoughts on
the<br>Orientation case:<br><br>learning the orientations is like
programming a good application. Design is a key element. I
would highly recommend starting with a good plan (you
have a good start by asking these questions). Most
important, before you start learning them know how you want
to learn them. Once you sort of know them, you won't
want to change. these are the recommendations I
have.<br><br>1. learn them in an order that allows you to
practice them in real solutions of the cube.<br>this means
learn the two six move cases for the edges and then all
of the corner cases. At this point, you can solve
the cube edges, corners, permutation. (cubin4speed
argues that just using this method you can solve the
cube in an average of sub 20 seconds) I agree for the
most part (you have to be solving the first two layers
in 10s). Anyways, this is a good start.<br>Then like
ilovemycube suggested, learn the cases where none of the
edges are solved (these are the hardest to learn and
most infrequent and have the longest number of moves)
However, seeing that none of the edges are solved is easy
and quick so you can incorporate these moves into
your solution while still speed solving.<br>Then
tackle the rest of the moves...<br><br>2. Take the time
to learn the moves the way that is best for you. Try
all of the different moves, experiment with the
different ways to hold the cube, try to find places where
you can use triggers (NOTE: I would not recommend
finishing any move with a trigger because it distracts your
recognition of the permutation). I would say that you should
let the move dictate your starting configuration and
then like you said, redraw the case for your way. It
doesn't really matter how you recognize the case and
where you start it.<br><br>3. practice each case
manytimes before you start learning the next (mor important
as you get further along)<br><br>4. learn the
mirrors (where neccessary) while you are learning the
other move.<br><br>5. have patience.<br><br>6. One
other thought is that when getting ready for the
orientation case (right after finishing the F2L) instead of
turning the whole cube to get it to the right position,
just turn the up face. I have found this quicker.

Here are some of the moves that are
fast:<br><br>Z permutation / (two and two edge switch)<br> /
(not opposite)<br>13 moves can be done in under 2
seconds<br><br>suni (three corner twist)<br><br>7 moves (.86
seconds)<br><br>I have been experimenting with a trigger that is as
follows:<br><br>RURUiRi turn R pull right pointer trigger<br> continue
with R pull left trigger while<br> lifting right thumb
off top<br> turn Ri back<br><br>this can be modified
and extended to a pattern of <br> -RiURUi<br>I think
that this could be done maybe as fast as ten moves a
second (that would be rad) (in Singmaster's book the
suggested maximum moves per second is something like two or
three) (granted this move doesn't do
anything)<br><br>Who knows other fast sequences?

Hi 'Ilovemycube',<br>The images on your homepage
don't work properly. You will have to change all the \
to / in the image
refs.<br><a href=http://members1.chello.nl/~r.vanbruchem
target=new>http://members1.chello.nl/~r.vanbruchem</a><br>Let me know when your
page is finished, so I can then
link to it from
mine.<br><a href=http://www.org2.com/jaap/puzzles
target=new>http://www.org2.com/jaap/puzzles</a><br><br>Jaap

hi SW1 here<br><br>With one of the
links<br><br><a href=http://ng.netgate.net/~mette/lars/cubedude/index.html
target=new>http://ng.netgate.net/~mette/lars/cubedude/index.html</a><br>Does
anyone know how to save that page onto hard disk?
(with those rotating cubes intact)<br><br>I keep on
forgetting some steps and its quite frustrating having to
connect everytime I want to look up the page. (I think
the rotating cube is some applets of some
sort)<br><br><br>Thanks

Hi SW1,<br><br>The rotating cube is a Java
applet.<br>I suppose you have Internet Explorer. If you save
the page using Internet Explorer's menu option "Save
as", it will not save the Java applet in the
directory.<br>You have to use the Find function of Windows in the
directory "c:\windows\Temporary Internet Files" to find the
applet. Look for lrubik.class.<br>Maybe it will be called
lrubik[1].class.<br>Copy (and rename if necessary to lrubik.class) the
class to the directory where you saved the
page.<br><br>Good luck,<br><br>Ilovemycube

Thanks<br>Ilovemycube  :)

I know a that there are people who do the first
two layers holding the cube so that the cross is on
the bottom so this isn't new I know. I've started
doing the cross and the corners and edge pieces with
the cube flipped over, I form the cross on the bottom
and then finish the first two layers. I'm still
trying to get used to it but from the times I've
practiced it my hands don't feel that they get in the way
of each other or anything so it seems like it would
go pretty well. I do like the fact that I never need
to flip the cube, which saves the sum .5 or .6
second it takes to flip it :-) I was curious though
becuase I used to hold the cross kind of half left and
half back, on a kind of diagonal. Is doing the first
two layers the same as the 3rd layer algorithms, you
just have to customize it for your hands or is there
one way that is better or faster than another? I
haven't gotten used to holding the cube upside down
enough to try it fast but once I do I'll probably know
the answer to my own question, whether one is faster
or not. But if anyone who does the cross upside down
likes it, or thinks it's terrible or anything I'd like
to hear about it. I'm trying to find the way that's
the fastest I can do but still comfortable to my
hands.

I solve with the F2L on the bottom and I like
it.<br><br>My impression is that you can be comfortable and
fast using any orientation of the cube once you get
used to it (probably after months of practice). So in
my opinion it's best to choose your orientation
based on factors like visibility, minimization of
whole-cube rotation, and overall continuity of the
solution.<br><br>Actually, I used to solve my cross on the top and then flip
it over for the rest of the solution, which caused
me to get lost sometimes when starting the F2L. I'm
currently practicing doing the whole thing from bottom to
top.<br><br>BTW, I don't solve the cube upside down, you do... : )

I agree with the comments of cubin4speed. I
started out forming the cross on top and keeping it there
while finishing the F2L. Main problem was finding the
cubie pairs quickly. This was particularly difficult
for the so called 'separated on bottom'
cases.<br><br> So I recently began forming cross on top and
flipping cube over and solving bottom to top. I haven't
yet been able to form cross on bottom and keep it
there; obviously this would be better. But so far, it
seems a bit easier to find the cubie pairs while
holding the cross on the bottom. Once I master the hand
movements, it should be a quicker method.<br><br>Question:
After I've got the cross on the bottom, I've been using
the right upper front corner and right front edge as
my "target slots" when placing the cubie pairs. Is
there a better approach? And for example, after I've
inserted a pair into those right "target slots", say the
next pair I see are separated and belong in the back
left slots, what do I do? Right now, I'm reduced to
turning the cube to get the "target slots" into front
right positions to orient myself for the memorized
move. I hope this makes sense. Any help? Thanks.

In regards to slowcuber's question:<br>Yes! I
know exactly what you mean about always using the
front right as your target slot. When your new target
slot is in the back or left, keep it there whenever
possible!<br>This may be hard at first, but soon the moves will
make sense from any position. <br>If you use this idea
AND take advantage of all empty slots to shorten your
moves (as explained by ilovemycube), then you will
almost never have to turn the cube or waste turns of the
U layer during the F2L. On a good run you can cube
straight through the F2L without any pauses.<br><br>I took
a month or two off from timing in order to practice
this, and when I started again, I had cut my times by
about 20%.<br><br><br>A while back I recommended trying
to solve the cube without (hardly) ever rotating it,
and then incorporating that into your solution - this
is what I was talking about.