Athefre
Member
- Joined
- Jul 25, 2006
- Messages
- 1,283
Some interesting method-related quotes from the now closed speedsolvingrubikscube group. I've known about these for years, but haven't posted them until now.
I thought Gilles Roux's CP line was an interesting find considering a recent topic and debate about the viability. The others show that Ryan Heise proposed EOLine three years before Zbigniew Zborowski, the first proposal for non-matching blocks in a speedsolve, and that Jessica Fridrich did more than just use the F2L method she learned from others - she developed it to a more advanced form.
I've always thought Ryan Heise is one of the greatest method developers. His completely intuitive Heise method with non-matching blocks and EO built in, HTA, Tripod, said that he experimented with an EO form of the Roux method before 2003, and the first to propose the EOLine/EOCross idea.
1/07/2003
This is another idea I had a while ago. It similar to the Petrus idea of
orienting edges early so that you only need to turn two sides.
Instead, we orient all edges right at the beginning, create a "line" on
the D side from front to back (or a cross if you wish), then we only
need to turn three sides to solve the rest of the cube (L, U, R). This
also means you don't ever need to rotate the cube.
To see what I mean, have a look at this video by AdaM:
Either by plan or by chance, all of your F2L edges must have been
oriented after your first few opening moves. We can deduce this because
after that you only used L,U,R turns to solve the rest of the F2L. On
the other hand, not all of your LL edges were oriented. I have to say
seeing you solve the cube without rotating it looks pretty cool.
Note, it only takes 4.6 moves on average to orient all edges at the
beginning, and maybe you can combine this with placing parts of the
"line" or "cross".
This is another idea I had a while ago. It similar to the Petrus idea of
orienting edges early so that you only need to turn two sides.
Instead, we orient all edges right at the beginning, create a "line" on
the D side from front to back (or a cross if you wish), then we only
need to turn three sides to solve the rest of the cube (L, U, R). This
also means you don't ever need to rotate the cube.
To see what I mean, have a look at this video by AdaM:
Either by plan or by chance, all of your F2L edges must have been
oriented after your first few opening moves. We can deduce this because
after that you only used L,U,R turns to solve the rest of the F2L. On
the other hand, not all of your LL edges were oriented. I have to say
seeing you solve the cube without rotating it looks pretty cool.
Note, it only takes 4.6 moves on average to orient all edges at the
beginning, and maybe you can combine this with placing parts of the
"line" or "cross".
1/07/2003
If you don't want to rotate the cube or move your left hand, here's
something you can do too:
- Make a 1x1x3 line: DLB, DL, DLF (4 moves)
- Make the 6 remaining corners solvable in <R,U> (3 moves)
- Then, solve everything using R, U, r and u only. (Hint: siamese cubes)
As an example, see my solution on this page:
Gilles.
PS: This method is useless for sub20 cubing.
If you don't want to rotate the cube or move your left hand, here's
something you can do too:
- Make a 1x1x3 line: DLB, DL, DLF (4 moves)
- Make the 6 remaining corners solvable in <R,U> (3 moves)
- Then, solve everything using R, U, r and u only. (Hint: siamese cubes)
As an example, see my solution on this page:
Gilles.
PS: This method is useless for sub20 cubing.
11/12/2002
I have thought of a possible improvement that may apply to more than
just one method.
To demonstrate, consider the petrus method:
1. solve a 2x2x3 block
2. complete the first two layers by adding a 3x2x1 slice
3. complete the last layer
One approach for step 2 is:
a. build a 2x2x1 section
b. add on a 1x2x1 section
For (a), you can choose between 4 starting points and pick the easiest one. If
you find a good one, it can be done in 2, 3 or 4 moves. Otherwise it can take
about 8 moves. Sometimes there are no good options among the 4 starting
points.
But now consider that there are actually 8 different 3x2x1 slices to choose
from in step 2. The slice you choose doesn't have to join with step 1 with
matching colours, it just needs to be the same shape. In the final step, you
can slide this slice back to where it should go in just one move.
This gives you 8 different options for step 2a, greatly increasing your
chances of finding a good option.
Once you get to the final layer, you have to be able to recognise all the
patterns with three of the pieces being a different colour, but it's not as
difficult as you might think - try it. Of course, some systems for the last
layer may be easier to handle than others.
This idea also applies, with more opportunities, in the tripod approach
described in my last email.
I have thought of a possible improvement that may apply to more than
just one method.
To demonstrate, consider the petrus method:
1. solve a 2x2x3 block
2. complete the first two layers by adding a 3x2x1 slice
3. complete the last layer
One approach for step 2 is:
a. build a 2x2x1 section
b. add on a 1x2x1 section
For (a), you can choose between 4 starting points and pick the easiest one. If
you find a good one, it can be done in 2, 3 or 4 moves. Otherwise it can take
about 8 moves. Sometimes there are no good options among the 4 starting
points.
But now consider that there are actually 8 different 3x2x1 slices to choose
from in step 2. The slice you choose doesn't have to join with step 1 with
matching colours, it just needs to be the same shape. In the final step, you
can slide this slice back to where it should go in just one move.
This gives you 8 different options for step 2a, greatly increasing your
chances of finding a good option.
Once you get to the final layer, you have to be able to recognise all the
patterns with three of the pieces being a different colour, but it's not as
difficult as you might think - try it. Of course, some systems for the last
layer may be easier to handle than others.
This idea also applies, with more opportunities, in the tripod approach
described in my last email.
23/07/2003
I am quite sure that there are many people who came up with this
idea for solving the F2L independently, so it will be next to
impossible to find the "inventor". I learned about it when I joined
the college. There were at least 5 guys (all coming from the same
high school) doing the F2L this way. However, they solved the F2L
kind of intuitively using lots of auxiliary moves with only a
handful of basic intuitive moves. As a result, they were not too
fast, and so I did not pay attention to this idea at first. Then,
one day just for fun I tried to solve the F2L using this approach
and quickly saw the _potential_ of this approach. However, it needed
a substantial "overhaul". Thus, I developed over a dozen new, less
obvious algorithms that you can see on my page.
It must have been somebody from that high school from which the 5
guys came who "invented" the idea. Perhaps, we could find out the
name if I contact those 5 guys and talk to them. Then, we would have
to ask Guus about the origin of his system and try to decide who
was "first" ... Any volunteers for this?
I am quite sure that there are many people who came up with this
idea for solving the F2L independently, so it will be next to
impossible to find the "inventor". I learned about it when I joined
the college. There were at least 5 guys (all coming from the same
high school) doing the F2L this way. However, they solved the F2L
kind of intuitively using lots of auxiliary moves with only a
handful of basic intuitive moves. As a result, they were not too
fast, and so I did not pay attention to this idea at first. Then,
one day just for fun I tried to solve the F2L using this approach
and quickly saw the _potential_ of this approach. However, it needed
a substantial "overhaul". Thus, I developed over a dozen new, less
obvious algorithms that you can see on my page.
It must have been somebody from that high school from which the 5
guys came who "invented" the idea. Perhaps, we could find out the
name if I contact those 5 guys and talk to them. Then, we would have
to ask Guus about the origin of his system and try to decide who
was "first" ... Any volunteers for this?
I thought Gilles Roux's CP line was an interesting find considering a recent topic and debate about the viability. The others show that Ryan Heise proposed EOLine three years before Zbigniew Zborowski, the first proposal for non-matching blocks in a speedsolve, and that Jessica Fridrich did more than just use the F2L method she learned from others - she developed it to a more advanced form.
I've always thought Ryan Heise is one of the greatest method developers. His completely intuitive Heise method with non-matching blocks and EO built in, HTA, Tripod, said that he experimented with an EO form of the Roux method before 2003, and the first to propose the EOLine/EOCross idea.
