Hello. My name is Anthony Deschamps. I am 16 years old and I live in
Windsor, Ontario, Canada. I have recently become the 82nd person to
solve the 3^4 cube.
I first solved the traditional cube when I was about 10 years old.=20
Since then I've learned how to solve the 4x4 and 5x5. However, it's
been a while and those are no longer challenging, so I was excited to
find MC4D.
I've always liked to challenge myself and I thoroughly enjoyed
puzzling over the 4d cube. Now I aim to solve the larger 4d cubes
and, if I can, MC5D.
I'm glad that I've found people who share my interests. I only know a
small group of people who can solve the traditional 3d cube, so I
sometimes feel alone when I am looking to bounce ideas off of others.
Hello everyone, my name is Matthew and I have recently returned to the prob=
lem of the 4D cube. I found them sometime last year after the advent of my=
cube obsession, but was too scared by what I saw to try to solve any!
It took a few days of messing around to figure out the mechanics and get so=
me ideas kicking around about how to solve it, and i had a practice at the =
2^4 first. My first (and currently only) 3^4 solution took a few days to w=
ork through, due to the time I had available, and also because I was figuri=
ng out a method as I went along, so it wasn't incredibly efficient. I then=
moved onto the 4^4, and it was simply a matter of adding a few steps onto =
my 3^4 solution in order to solve it.
I tried the all-purpose reduction tactic. First I solved the 1C pieces in =
blocks of 2, then I paired up the 2C pieces, first into 2 pairs from the sa=
me face, then into a block of 4. I then paired up 3C pieces like i would o=
n a 4^3. Melinda commented on my log file, saying I luckily avoided parity=
, but from my (perhaps naive) understanding of the cycles involved here, I =
don't think the 3D parities are seen on 4D.
I then used an improvement of my 3^4 solution to solve the rest. This invo=
lves first solving one of the cubes in columns of 3 pieces, apart from the =
last column consisting of a 3C and two 4C pieces which is left unsolved. I=
then solve what I call the "middle cube" (all the pieces not on the centra=
l or outside cube) by using the freedom from the unsolved column to first s=
olve the 8 2C pieces above and below, then making more comumns of 3 in the =
central cube and inserting them. This is equivalent of a keyhole F2L on a =
3D cube I believe. I then solved the block of 6 pieces to just leave the c=
entral cube, and a few moves before placing these 6 pieces left all 2C piec=
es facing the correct way. I then treat what is left almost like a 3x3x3, =
with a few setup moves required to help flip pieces in 4D using 3D algorith=
ms to create simple cycles. After the pieces all have the sticker matching=
the 1C piece in the remaining cube, I solve the rest as a 3^3. This is do=
ne with the realisation that as long as you use the same adjacent cube to p=
erform the face turns, the rest of the puzzle will be unscrambled by this a=
s long as these are overall a multiple of 4 turns, which is easy to account=
for on the LL, all that is required is a turn on the central cube between =
moves to change which face is being turned by the adjecent cube. I believe=
that the nature of the 4D cube eliminates the need to double odd permutati=
ons like the T-perm which would require an odd number of moves on the adjac=
ent cube and thus prevent the outer pieces from being solved as the adjacen=
t cube must be moved in mutilples of four.
I have now started looking at the 5D cube to get the hang of the mechanics,=
and though I wasn't going to, I am solving the 5^4 slowly in the backgroun=
d, between 5D and schoolwork.
As a new member, any comments would be appreciated, though please no spoile=
rs on the 5D puzzles and their solutions. I can have a mental breakdown of=
my own accord.
Matthew S
--0016364c6af98a90f8046692c478
Content-Type: text/plain; charset=ISO-8859-13
Content-Transfer-Encoding: quoted-printable
Welcome Matthew. Great to have you join us.
> Melinda commented on my log file, saying I luckily avoided parity, but fr=
om
> my (perhaps naive) understanding of the cycles involved here, I don't thi=
nk
> the 3D parities are seen on 4D.
Coincidentally, this question of parities came up and led to lively
discussion only a few months ago. With the paradoxical feeling of a
Buddhist k=F4an, let me just say you're both right. There was much more to
talk about than one might have guessed, and I think we fleshed things out
pretty well. Here is the thread in case you are interested to read it:
http://games.groups.yahoo.com/group/4D_Cubing/messages/620?xm=3D1&threaded=
=3D1&m=3Ds&l=3D1
Things started with message 620, and became solely devoted to parities at
message 624. However, the discussion was not limited to the 4D cubes (so i=
f
you want to avoid spoilers for a bit, you might want to hold off for now)..=
.
All the best,
Roice
--0016364c6af98a90f8046692c478
Content-Type: text/html; charset=ISO-8859-13
Content-Transfer-Encoding: quoted-printable
0px 0.8ex;border-left:#ccc 1px solid">Melinda commented on my log file, say=
ing I luckily avoided parity, but from my (perhaps naive) understanding of =
the cycles involved here, I don't think the 3D parities are seen on 4D.=
--- In 4D_Cubing@yahoogroups.com, Roice Nelson
> Coincidentally, this question of parities came up and led to lively
> discussion only a few months ago. With the paradoxical feeling of a
> Buddhist k=F4an, let me just say you're both right. There was much more =
to
> talk about than one might have guessed, and I think we fleshed things out
> pretty well.
I have looked in the database to find the parity present on the 4^4, and fo=
und the case which you have found in your solves, with the 2C pieces. I di=
dn't really think about this possibility, as my way of dealing with the las=
t few 2C pieces works differently from the first load and works around this=
, though I suppose it breifly crossed my mind. I leave all the 2C pieces w=
ith one colour (I chose blue as it happens) to the end of the step of match=
ing up 2C pieces. I bring them to one hyperface then flip them all first s=
o that the one colour faces inwards, then match like centres on a 4^3. If =
the number of 2C pieces isnt a multiple of 4, I simply flip an odd number o=
f these with the alg I use which flips 8 2C pieces, and this leaves a multi=
ple of 4.
Thanks for bringing this to my attention, the discussion before and the par=
ity log file you submitted were very interesting.
Matthew
Gosh it's been ages since the last code update! The new version fixes a
few bugs and adds the ability to automatically undo macro set-up moves.
Here's how to use it: Macros > Begin Macro Setup Moves (
the current state and enters a new mode (yellow sky) until a macro is
applied or until cancelled (
reverse of the setup moves are applied. Repeat selections of this
command simply reset the mark to the current point. Undo/redo still
function properly in this mode though undoing past the setup mark
cancels the setup. The reverse move playback is treated the same as
macro moves in that they will undo/redo as a group. So if you apply a
macro with set-up moves and then click Edit > Undo (
reversed set-up moves will be undone. The next undo will undo all the
macro moves, and subsequent undos will step through your set-up moves.
Redo (
confusing. The net result is that macros should become even easier and
safer to use because you will no longer need to memorize how to undo
your set-up moves and perform them correctly.
The new release has not been exhaustively tested, so please help by
trying it out as soon as you can and exercising the new functionality
along with the other features that you typically use. You can reply with
public comments here, private email to me, or open new issues at the new
repository here: https://github.com/cutelyaware/magiccube4d/issues/
Thanks and happy puzzling!
-Melinda
Indeed. I looked at the wiki page after my first =C2=A0 The last 3D face can be done in at most twice as many =C2=A0 The human Thitlethwaite is not particularly efficient =C2=A0 Block-building methods seem the thing to do indeed. It =C2=A0 Thank you for your answer! Thomas
--------------070802030304060007090601
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable
Welcome Thomas indeed! There's nothing like breaking one of our most=20
cherished records for waking us from blissful slumber.
I'm happy to create new solution categories though I'd like to make sure=20
we have a rough idea how to determine where potential new methods will=20
fall. Tools like reset, undo & redo are forms of computer assistance but=20
they seemed gentle enough to not detract from an otherwise pure human=20
solution. Macros OTOH required quite a bit more introspection and we=20
ended up categorizing them as another valid human solution, though not=20
directly comparable to no-macro solutions.
How should we now consider more sophisticated computer assistance that=20
goes beyond simple mechanistic aid into some forms of real thinking? I=20
definitely want to create a "shortest computer solution" category. So=20
far we've only had one by Don Hatch which produces solutions with around=20
1,500 twists. Perhaps when we have a more efficient solution, we should=20
create a "first" and "shortest" computer solution category pair like the=20
others. I can also imagine a time in which a human only guides the broad=20
strokes of an otherwise completely computer-driven solution. Would that=20
be comparable to this new computer-assisted solution or should we then=20
create a new "human-assisted" category in anticipation of our=20
obsolescence? These are obviously more philosophical questions than=20
practical ones but they might inform our immediate choices.
I'm really hoping that someone will implement an efficient computer=20
solution that can be integrated into MC4D and replace Don's version=20
which I sadly broke. It was nice having a true solution integrated,=20
though 1,500 twists turned out to be just a bit too tedious to watch.=20
Something around 200 twists seems much more practical, plus it could be=20
merged with the current "cheating" solution such that states involving=20
less than that number of twists will simply reverse all those twists=20
(ideally with some compression which I also broke), and solutions longer=20
than that will use the full computer solution. In fact if we do this=20
right, the real prize for the shortest computer solution could be its=20
integration into MC4D!
And speaking of human obsolescence, are any of you also Go players, and=20
have you been following the news regarding the upcoming human vs. AI=20
match? Go has been a very difficult game to program due to the roughly=20
200 possible moves at every turn. Google's AlphaGo AI recently crushed=20
the European Go champion and next month will play against the legendary=20
Lee Se-dol who has dominated the game for the last decade. That match=20
has a $1,000,000 prize and could well be a watershed moment even bigger=20
than when computers became the best chess players in 1996. The match=20
begins on March 9th. Details here
Happy puzzling!
-Melinda
On 2/23/2016 9:35 AM, Roice Nelson roice3@gmail.com [4D_Cubing] wrote:
>
>
> Welcome Thomas!
>
> My vote is that we mark this new solution as a first class record in a=20
> new category, "shortest computer assisted solve". I've considered=20
> trying to use the computer for attacks on the shortest competition in=20
> the past, and it'd be great if more folks were motivated to do this.=20=20
> With advances in this area we could, for example, come closer to=20
> intuiting what God's Number for the 3^4 might be. We don't even have=20
> a rough idea of what it is right now, upper OR lower bounds (as far as=20
> I know).
>
> Congrats on your impressive solve, and happy to have you posting here.
>
> Cheers,
> Roice
>
>
>
> On Tue, Feb 23, 2016 at 10:21 AM, Thomas Leh=C3=A9ricy=20
> thomas.lehericy78@orange.fr
> [4D_Cubing] <4D_Cubing@yahoogroups.com=20
>
>
>
>
> Indeed. I looked at the wiki page after my first solves, but
> didn't understand everything and preferred to keep going with the
> method I was designing - at this point using an analogue of CFOP
> was just the thing to do, seing how intuitive it is when you know
> it well. Now that I read it again it looks very clear, and it
> indeed looks the same as my own method up to the last layer.
>
> The last 3D face can be done in at most twice as many moves as one
> would need for the 3D cube. To do that, one can simply "regrip"
> (rotate) the cube so that the face one turns is always the same.
> Of course it can be improved, for instance when using URU'R': a
> single regrip in the middle and all moves will cancel... So it's
> only an upper bound. I don't know of any general method to
> optimize this step, although I would be extremely interested.
>
> The human Thitlethwaite is not particularly efficient at giving
> low-move counts solutions, but still better than CFOP. I think you
> can hope for a 40-50 move counts on average if you know all cases
> (which I don't), without optimizing it for too long. What is good
> is that each step is rather intuitive, and it can be optimized and
> yields extremely good results: Kociemba's algorithm is derived
> from it.
>
> Block-building methods seem the thing to do indeed. It seems to me
> that Matthew Sheerin built his first two layers like this in his
> record. It is not as optimized nor as flexible as Heise, and it
> would be indeed interesting to see how well Heise translates into
> 4D - but that's far beyond my abilities right now.
>
> Thank you for your answer!
>
> Thomas
>
>
>
>
>
>
>=20
--------------070802030304060007090601
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
">
Welcome Thomas indeed! There's nothing like breaking one of our most
cherished records for waking us from=C2=A0 blissful slumber.
I'm happy to create new solution categories though I'd like to make
sure we have a rough idea how to determine where potential new
methods will fall. Tools like reset, undo & redo are forms of
computer assistance but they seemed gentle enough to not detract
from an otherwise pure human solution. Macros OTOH required quite a
bit more introspection and we ended up categorizing them as another
valid human solution, though not directly comparable to no-macro
solutions.
How should we now consider more sophisticated computer assistance
that goes beyond simple mechanistic aid into some forms of real
thinking? I definitely want to create a "shortest computer solution"
category. So far we've only had one by Don Hatch which produces
solutions with around 1,500 twists. Perhaps when we have a more
efficient solution, we should create a "first" and "shortest"
computer solution category pair like the others. I can also imagine
a time in which a human only guides the broad strokes of an
otherwise completely computer-driven solution. Would that be
comparable to this new computer-assisted solution or should we then
create a new "human-assisted" category in anticipation of our
obsolescence? These are obviously more philosophical questions than
practical ones but they might inform our immediate choices.
I'm really hoping that someone will implement an efficient computer
solution that can be integrated into MC4D and replace Don's version
which I sadly broke. It was nice having a true solution integrated,
though 1,500 twists turned out to be just a bit too tedious to
watch. Something around 200 twists seems much more practical, plus
it could be merged with the current "cheating" solution such that
states involving less than that number of twists will simply reverse
all those twists (ideally with some compression which I also broke),
and solutions longer than that will use the full computer solution.
In fact if we do this right, the real prize for the shortest
computer solution could be its integration into MC4D!
And speaking of human obsolescence, are any of you also Go players,
and have you been following the news regarding the upcoming human
vs. AI match? Go has been a very difficult game to program due to
the roughly 200 possible moves at every turn. Google's AlphaGo AI
recently crushed the European Go champion and next month will play
against the legendary Lee Se-dol who has dominated the game for the
last decade. That match has a $1,000,000 prize and could well be a
watershed moment even bigger than when computers became the best
chess players in 1996. The match begins on March 9th. Details
Happy puzzling!
-Melinda
com"
type=3D"cite">
record in a new category, "shortest computer assisted solve".=C2=
=A0
I've considered trying to use the computer for attacks on the
shortest competition in the past, and it'd be great if more
folks were motivated to do this.=C2=A0 With advances in this area
we could, for example, come closer to intuiting what God's
Number for the 3^4 might be.=C2=A0 We don't even have a rough ide=
a
of what it is right now, upper OR lower bounds (as far as I
know).
posting here.
Thomas Leh=C3=A9ricy href=3D"mailto:thomas.lehericy78@orange.fr">thomas.lehericy78@o=
range.fr
[4D_Cubing] < href=3D"mailto:4D_Cubing@yahoogroups.com" target=3D"_blank">4=
D_Cubing@yahoogroups.com>
wrote:
.8ex;border-left:1px #ccc solid;padding-left:1ex">
solves, but didn't understand everything and preferred
to keep going with the method I was designing - at this
point using an analogue of CFOP was just the thing to
do, seing how intuitive it is when you know it well. Now
that I read it again it looks very clear, and it indeed
looks the same as my own method up to the last layer.
moves as one would need for the 3D cube. To do that, one
can simply "regrip" (rotate) the cube so that the face
one turns is always the same. Of course it can be
improved, for instance when using URU'R': a single
regrip in the middle and all moves will cancel... So
it's only an upper bound. I don't know of any general
method to optimize this step, although I would be
extremely interested.
at giving low-move counts solutions, but still better
than CFOP. I think you can hope for a 40-50 move counts
on average if you know all cases (which I don't),
without optimizing it for too long. What is good is that
each step is rather intuitive, and it can be optimized
and yields extremely good results: Kociemba's algorithm
is derived from it.
seems to me that Matthew Sheerin built his first two
layers like this in his record. It is not as optimized
nor as flexible as Heise, and it would be indeed
interesting to see how well Heise translates into 4D -
but that's far beyond my abilities right now.
=20=20=20=20=20=20
--------------070802030304060007090601--
------=_NextPart_000_0012_01D16EE7.D94EB5D0
Content-Type: text/plain;
charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
I'm very amused to read "human-assisted" or "our obsolescence". This are ne=
cessary new humble terms after "godlike humans" and "from human invented go=
ds".
Is there a 4D fan which doesn't play go? Go alpha is a little bit sad.
Kind regards
Ed
----- Original Message -----=20
From: Melinda Green melinda@superliminal.com [4D_Cubing]=20
To: 4D_Cubing@yahoogroups.com=20
Sent: Wednesday, February 24, 2016 2:10 AM
Subject: Re: [MC4D] Re: Greetings
=20=20=20=20
Welcome Thomas indeed! There's nothing like breaking one of our most cher=
ished records for waking us from blissful slumber.=20
I'm happy to create new solution categories though I'd like to make sure =
we have a rough idea how to determine where potential new methods will fall=
. Tools like reset, undo & redo are forms of computer assistance but they s=
eemed gentle enough to not detract from an otherwise pure human solution. M=
acros OTOH required quite a bit more introspection and we ended up categori=
zing them as another valid human solution, though not directly comparable t=
o no-macro solutions.
How should we now consider more sophisticated computer assistance that go=
es beyond simple mechanistic aid into some forms of real thinking? I defini=
tely want to create a "shortest computer solution" category. So far we've o=
nly had one by Don Hatch which produces solutions with around 1,500 twists.=
Perhaps when we have a more efficient solution, we should create a "first"=
and "shortest" computer solution category pair like the others. I can also=
imagine a time in which a human only guides the broad strokes of an otherw=
ise completely computer-driven solution. Would that be comparable to this n=
ew computer-assisted solution or should we then create a new "human-assiste=
d" category in anticipation of our obsolescence? These are obviously more p=
hilosophical questions than practical ones but they might inform our immedi=
ate choices.
I'm really hoping that someone will implement an efficient computer solut=
ion that can be integrated into MC4D and replace Don's version which I sadl=
y broke. It was nice having a true solution integrated, though 1,500 twists=
turned out to be just a bit too tedious to watch. Something around 200 twi=
sts seems much more practical, plus it could be merged with the current "ch=
eating" solution such that states involving less than that number of twists=
will simply reverse all those twists (ideally with some compression which =
I also broke), and solutions longer than that will use the full computer so=
lution. In fact if we do this right, the real prize for the shortest comput=
er solution could be its integration into MC4D!
And speaking of human obsolescence, are any of you also Go players, and h=
ave you been following the news regarding the upcoming human vs. AI match? =
Go has been a very difficult game to program due to the roughly 200 possibl=
e moves at every turn. Google's AlphaGo AI recently crushed the European Go=
champion and next month will play against the legendary Lee Se-dol who has=
dominated the game for the last decade. That match has a $1,000,000 prize =
and could well be a watershed moment even bigger than when computers became=
the best chess players in 1996. The match begins on March 9th. Details her=
e.
Happy puzzling!
-Melinda
On 2/23/2016 9:35 AM, Roice Nelson roice3@gmail.com [4D_Cubing] wrote:
Welcome Thomas!=20
My vote is that we mark this new solution as a first class record in a =
new category, "shortest computer assisted solve". I've considered trying t=
o use the computer for attacks on the shortest competition in the past, and=
it'd be great if more folks were motivated to do this. With advances in t=
his area we could, for example, come closer to intuiting what God's Number =
for the 3^4 might be. We don't even have a rough idea of what it is right =
now, upper OR lower bounds (as far as I know).
Congrats on your impressive solve, and happy to have you posting here.
Cheers,
Roice
On Tue, Feb 23, 2016 at 10:21 AM, Thomas Leh=C3=A9ricy thomas.lehericy7=
8@orange.fr [4D_Cubing] <4D_Cubing@yahoogroups.com> wrote:
Indeed. I looked at the wiki page after my first solves, but didn't u=
nderstand everything and preferred to keep going with the method I was desi=
gning - at this point using an analogue of CFOP was just the thing to do, s=
eing how intuitive it is when you know it well. Now that I read it again it=
looks very clear, and it indeed looks the same as my own method up to the =
last layer.
The last 3D face can be done in at most twice as many moves as one wo=
uld need for the 3D cube. To do that, one can simply "regrip" (rotate) the =
cube so that the face one turns is always the same. Of course it can be imp=
roved, for instance when using URU'R': a single regrip in the middle and al=
l moves will cancel... So it's only an upper bound. I don't know of any gen=
eral method to optimize this step, although I would be extremely interested=
.
The human Thitlethwaite is not particularly efficient at giving low-m=
ove counts solutions, but still better than CFOP. I think you can hope for =
a 40-50 move counts on average if you know all cases (which I don't), witho=
ut optimizing it for too long. What is good is that each step is rather int=
uitive, and it can be optimized and yields extremely good results: Kociemba=
's algorithm is derived from it.
Block-building methods seem the thing to do indeed. It seems to me th=
at Matthew Sheerin built his first two layers like this in his record. It i=
s not as optimized nor as flexible as Heise, and it would be indeed interes=
ting to see how well Heise translates into 4D - but that's far beyond my ab=
ilities right now.
Thank you for your answer!
Thomas
=20=20
------=_NextPart_000_0012_01D16EE7.D94EB5D0
Content-Type: text/html;
charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
=EF=BB=BF
Welcome Thomas indeed! There's nothing like breaking one of our most=20
cherished records for waking us from blissful slumber.
I'm =
happy=20
to create new solution categories though I'd like to make sure we have a =
rough=20
idea how to determine where potential new methods will fall. Tools like r=
eset,=20
undo & redo are forms of computer assistance but they seemed gentle e=
nough=20
to not detract from an otherwise pure human solution. Macros OTOH require=
d=20
quite a bit more introspection and we ended up categorizing them as anoth=
er=20
valid human solution, though not directly comparable to no-macro=20
solutions.
How should we now consider more sophisticated computer=
=20
assistance that goes beyond simple mechanistic aid into some forms of rea=
l=20
thinking? I definitely want to create a "shortest computer solution" cate=
gory.=20
So far we've only had one by Don Hatch which produces solutions with arou=
nd=20
1,500 twists. Perhaps when we have a more efficient solution, we should c=
reate=20
a "first" and "shortest" computer solution category pair like the others.=
I=20
can also imagine a time in which a human only guides the broad strokes of=
an=20
otherwise completely computer-driven solution. Would that be comparable t=
o=20
this new computer-assisted solution or should we then create a new=20
"human-assisted" category in anticipation of our obsolescence? These are=
=20
obviously more philosophical questions than practical ones but they might=
=20
inform our immediate choices.
I'm really hoping that someone will=
=20
implement an efficient computer solution that can be integrated into MC4D=
and=20
replace Don's version which I sadly broke. It was nice having a true solu=
tion=20
integrated, though 1,500 twists turned out to be just a bit too tedious t=
o=20
watch. Something around 200 twists seems much more practical, plus it cou=
ld be=20
merged with the current "cheating" solution such that states involving le=
ss=20
than that number of twists will simply reverse all those twists (ideally =
with=20
some compression which I also broke), and solutions longer than that will=
use=20
the full computer solution. In fact if we do this right, the real prize f=
or=20
the shortest computer solution could be its integration into MC4D!
speaking of human obsolescence, are any of you also Go players, and have =
you=20
been following the news regarding the upcoming human vs. AI match? Go has=
been=20
a very difficult game to program due to the roughly 200 possible moves at=
=20
every turn. Google's AlphaGo AI recently crushed the European Go champion=
and=20
next month will play against the legendary Lee Se-dol who has dominated t=
he=20
game for the last decade. That match has a $1,000,000 prize and could wel=
l be=20
a watershed moment even bigger than when computers became the best chess=
=20
players in 1996. The match begins on March 9th. Details
Happy=20
puzzling!
-Melinda
Indeed. I looked at the wiki page after my first solves, but didn'=
t=20
understand everything and preferred to keep going with the method I w=
as=20
designing - at this point using an analogue of CFOP was just the thin=
g to=20
do, seing how intuitive it is when you know it well. Now that I read =
it=20
again it looks very clear, and it indeed looks the same as my own met=
hod=20
up to the last layer.
The last 3D face can be done in at most twice as many moves as one=
=20
would need for the 3D cube. To do that, one can simply "regrip" (rota=
te)=20
the cube so that the face one turns is always the same. Of course it =
can=20
be improved, for instance when using URU'R': a single regrip in the m=
iddle=20
and all moves will cancel... So it's only an upper bound. I don't kno=
w of=20
any general method to optimize this step, although I would be extreme=
ly=20
interested.
The human Thitlethwaite is not particularly efficient at giving=20
low-move counts solutions, but still better than CFOP. I think you ca=
n=20
hope for a 40-50 move counts on average if you know all cases (which =
I=20
don't), without optimizing it for too long. What is good is that each=
step=20
is rather intuitive, and it can be optimized and yields extremely goo=
d=20
results: Kociemba's algorithm is derived from it.
Block-building methods seem the thing to do indeed. It seems to me=
that=20
Matthew Sheerin built his first two layers like this in his record. I=
t is=20
not as optimized nor as flexible as Heise, and it would be indeed=20
interesting to see how well Heise translates into 4D - but that's far=
=20
beyond my abilities right now.
Thank you for your answer!
Thomas
Indeed. I looked at the wiki page after my first =C2=A0 The last 3D face can be done in at most twice as many =C2=A0 The human Thitlethwaite is not particularly efficient =C2=A0 Block-building methods seem the thing to do indeed. It =C2=A0 Thank you for your answer! Thomas Indeed. I looked at the wiki page after my =C2=A0 The last 3D face can be done in at most =C2=A0 The human Thitlethwaite is not particularly =C2=A0 Block-building methods seem the thing to do =C2=A0 Thank you for your answer! Thomas
Hello Art is one of my pleasures of which I've done many Finally I also like puzzle solving and can solve 3D rubik I used the LBL method which I find to be aesthetically This methodology worked and I have done this a few times
--001a113e51121a9468052c873cd3
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
These are great questions that I wish I had answers for. I will say that
even though I'm marked as the "first no macro solution" in the Hall of
Fame, I've always felt that not using that feature isn't quite different
enough to warrant its own category. I think this just emphasizes what
you're saying about computer assisted categories being fuzzy. Maybe a
useful approach here would be not to find the perfect categories right now,
but to figure a way for them to grow or change easily over time? You seem
to have laid out a good start in any case, with possible beginning
categories like:
1. Human-only
2. Computer-assisted (>50% human solve, some computer state space
searching)
3. Human-assisted (>50% computer solve, I question whether there will be
enough here since implementations would likely just jump to the next
category like Don did.)
4. Computer-only (humans are allowed to write the software)
5. Singularity award (computers must write the software)
...with first and shortests for each. I'd vote to start with 1,2, and 4 I
think. We might want to consider putting 2/4 on a completely separate page=
.
On the topic of Go, I've had a background itch to learn for a long time,
and when I read the news about AlphaGo beating the European champion, I
finally decided it was time! I bought a board and some books, and have
been playing a lot the last few weeks. In fact, I was at my first meeting
with the Austin Go Club when your email came through last night. I had
high handicaps and was still getting crushed, but as expected I'm finding
the game extremely elegant, interesting, addictive, and fun - I can tell
this is a hobby that will stick. What are your predictions for the match
with Lee Se-dol? An Austin player I was talking to last night is betting
on the human. I'm rooting for AlphaGo, and am really looking forward to
following the drama!
Roice
On Tue, Feb 23, 2016 at 7:10 PM, Melinda Green melinda@superliminal.com
[4D_Cubing] <4D_Cubing@yahoogroups.com> wrote:
>
>
> Welcome Thomas indeed! There's nothing like breaking one of our most
> cherished records for waking us from blissful slumber.
>
> I'm happy to create new solution categories though I'd like to make sure
> we have a rough idea how to determine where potential new methods will
> fall. Tools like reset, undo & redo are forms of computer assistance but
> they seemed gentle enough to not detract from an otherwise pure human
> solution. Macros OTOH required quite a bit more introspection and we ende=
d
> up categorizing them as another valid human solution, though not directly
> comparable to no-macro solutions.
>
> How should we now consider more sophisticated computer assistance that
> goes beyond simple mechanistic aid into some forms of real thinking? I
> definitely want to create a "shortest computer solution" category. So far
> we've only had one by Don Hatch which produces solutions with around 1,50=
0
> twists. Perhaps when we have a more efficient solution, we should create =
a
> "first" and "shortest" computer solution category pair like the others. I
> can also imagine a time in which a human only guides the broad strokes of
> an otherwise completely computer-driven solution. Would that be comparabl=
e
> to this new computer-assisted solution or should we then create a new
> "human-assisted" category in anticipation of our obsolescence? These are
> obviously more philosophical questions than practical ones but they might
> inform our immediate choices.
>
> I'm really hoping that someone will implement an efficient computer
> solution that can be integrated into MC4D and replace Don's version which=
I
> sadly broke. It was nice having a true solution integrated, though 1,500
> twists turned out to be just a bit too tedious to watch. Something around
> 200 twists seems much more practical, plus it could be merged with the
> current "cheating" solution such that states involving less than that
> number of twists will simply reverse all those twists (ideally with some
> compression which I also broke), and solutions longer than that will use
> the full computer solution. In fact if we do this right, the real prize f=
or
> the shortest computer solution could be its integration into MC4D!
>
> And speaking of human obsolescence, are any of you also Go players, and
> have you been following the news regarding the upcoming human vs. AI matc=
h?
> Go has been a very difficult game to program due to the roughly 200
> possible moves at every turn. Google's AlphaGo AI recently crushed the
> European Go champion and next month will play against the legendary Lee
> Se-dol who has dominated the game for the last decade. That match has a
> $1,000,000 prize and could well be a watershed moment even bigger than wh=
en
> computers became the best chess players in 1996. The match begins on Marc=
h
> 9th. Details here
>
> Happy puzzling!
> -Melinda
>
> On 2/23/2016 9:35 AM, Roice Nelson roice3@gmail.com [4D_Cubing] wrote:
>
> Welcome Thomas!
>
> My vote is that we mark this new solution as a first class record in a ne=
w
> category, "shortest computer assisted solve". I've considered trying to
> use the computer for attacks on the shortest competition in the past, and
> it'd be great if more folks were motivated to do this. With advances in
> this area we could, for example, come closer to intuiting what God's Numb=
er
> for the 3^4 might be. We don't even have a rough idea of what it is righ=
t
> now, upper OR lower bounds (as far as I know).
>
> Congrats on your impressive solve, and happy to have you posting here.
>
> Cheers,
> Roice
>
>
>
> On Tue, Feb 23, 2016 at 10:21 AM, Thomas Leh=C3=A9ricy
> thomas.lehericy78@orange.fr [4D_Cubing] <4D_Cubing@yahoogroups.com> wrote=
:
>
>>
>>
>> Indeed. I looked at the wiki page after my first solves, but didn't
>> understand everything and preferred to keep going with the method I was
>> designing - at this point using an analogue of CFOP was just the thing t=
o
>> do, seing how intuitive it is when you know it well. Now that I read it
>> again it looks very clear, and it indeed looks the same as my own method=
up
>> to the last layer.
>>
>>
>>
>> The last 3D face can be done in at most twice as many moves as one would
>> need for the 3D cube. To do that, one can simply "regrip" (rotate) the c=
ube
>> so that the face one turns is always the same. Of course it can be
>> improved, for instance when using URU'R': a single regrip in the middle =
and
>> all moves will cancel... So it's only an upper bound. I don't know of an=
y
>> general method to optimize this step, although I would be extremely
>> interested.
>>
>>
>>
>> The human Thitlethwaite is not particularly efficient at giving low-move
>> counts solutions, but still better than CFOP. I think you can hope for a
>> 40-50 move counts on average if you know all cases (which I don't), with=
out
>> optimizing it for too long. What is good is that each step is rather
>> intuitive, and it can be optimized and yields extremely good results:
>> Kociemba's algorithm is derived from it.
>>
>>
>>
>> Block-building methods seem the thing to do indeed. It seems to me that
>> Matthew Sheerin built his first two layers like this in his record. It i=
s
>> not as optimized nor as flexible as Heise, and it would be indeed
>> interesting to see how well Heise translates into 4D - but that's far
>> beyond my abilities right now.
>>
>>
>>
>> Thank you for your answer!
>>
>> Thomas
>>
>>
>>
>
>
>
>=20
>
--001a113e51121a9468052c873cd3
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
=C2=A0 I will say that even though I'm marked as the "first no mac=
ro solution" in the Hall of Fame, I've always felt that not using =
that feature isn't quite different enough to warrant its own category.=
=C2=A0 I think this just emphasizes what you're saying about computer a=
ssisted categories being fuzzy.=C2=A0 Maybe a useful approach here would be=
not to find the perfect categories right now, but to figure a way for them=
to grow or change easily over time?=C2=A0 You seem to have laid out a good=
start in any case, with possible beginning categories like:
man-only
state space searching)
I question whether there will be enough here since implementations would li=
kely just jump to the next category like Don did.)
y (humans are allowed to write the software)
(computers must write the software)
and shortests for each.=C2=A0 I'd vote to start with 1,2, and 4 I think=
.=C2=A0 We might want to consider putting 2/4 on a completely separate page=
.
ch to learn for a long time, and when I read the news about AlphaGo beating=
the European champion, I finally decided it was time!=C2=A0 I bought a boa=
rd and some books, and have been playing a lot the last few weeks.=C2=A0 In=
fact, I was at my first meeting with the Austin Go Club when your email ca=
me through last night.=C2=A0 I had high handicaps and was still getting cru=
shed, but as expected I'm finding the game extremely elegant, interesti=
ng, addictive, and fun - I can tell this is a hobby that will stick.=C2=A0 =
What are your predictions for the match with=C2=A012.8px">Lee Se-dol?=C2=A0 An Austin player I was talking to last night is b=
etting on the human.=C2=A0 I'm rooting for AlphaGo, and am really looki=
ng forward to following the drama!
Melinda Green melinda@superlim=
inal.com [4D_Cubing] <yahoogroups.com" target=3D"_blank">4D_Cubing@yahoogroups.com>=
wrote:der-left:1px #ccc solid;padding-left:1ex">
=20=20=20=20=20=20=20=20
=20=20
=20=20=20=20
=20=20
Welcome Thomas indeed! There's nothing like breaking one of our mos=
t
cherished records for waking us from=C2=A0 blissful slumber.
I'm happy to create new solution categories though I'd like to =
make
sure we have a rough idea how to determine where potential new
methods will fall. Tools like reset, undo & redo are forms of
computer assistance but they seemed gentle enough to not detract
from an otherwise pure human solution. Macros OTOH required quite a
bit more introspection and we ended up categorizing them as another
valid human solution, though not directly comparable to no-macro
solutions.
How should we now consider more sophisticated computer assistance
that goes beyond simple mechanistic aid into some forms of real
thinking? I definitely want to create a "shortest computer solutio=
n"
category. So far we've only had one by Don Hatch which produces
solutions with around 1,500 twists. Perhaps when we have a more
efficient solution, we should create a "first" and "shor=
test"
computer solution category pair like the others. I can also imagine
a time in which a human only guides the broad strokes of an
otherwise completely computer-driven solution. Would that be
comparable to this new computer-assisted solution or should we then
create a new "human-assisted" category in anticipation of our
obsolescence? These are obviously more philosophical questions than
practical ones but they might inform our immediate choices.
I'm really hoping that someone will implement an efficient computer
solution that can be integrated into MC4D and replace Don's version
which I sadly broke. It was nice having a true solution integrated,
though 1,500 twists turned out to be just a bit too tedious to
watch. Something around 200 twists seems much more practical, plus
it could be merged with the current "cheating" solution such =
that
states involving less than that number of twists will simply reverse
all those twists (ideally with some compression which I also broke),
and solutions longer than that will use the full computer solution.
In fact if we do this right, the real prize for the shortest
computer solution could be its integration into MC4D!
And speaking of human obsolescence, are any of you also Go players,
and have you been following the news regarding the upcoming human
vs. AI match? Go has been a very difficult game to program due to
the roughly 200 possible moves at every turn. Google's AlphaGo AI
recently crushed the European Go champion and next month will play
against the legendary Lee Se-dol who has dominated the game for the
last decade. That match has a $1,000,000 prize and could well be a
watershed moment even bigger than when computers became the best
chess players in 1996. The match begins on March 9th. Details =3D"http://www.deepmind.com/alpha-go.html" target=3D"_blank">here.
Happy puzzling!
-Melinda
=20=20=20=20=20=20
=20=20=20=20=20=20
record in a new category, "shortest computer assisted solve&=
quot;.=C2=A0
I've considered trying to use the computer for attacks on the
shortest competition in the past, and it'd be great if more
folks were motivated to do this.=C2=A0 With advances in this area
we could, for example, come closer to intuiting what God's
Number for the 3^4 might be.=C2=A0 We don't even have a rough=
idea
of what it is right now, upper OR lower bounds (as far as I
know).
posting here.
Thomas Leh=C3=A9ricy r" target=3D"_blank">thomas.lehericy78@orange.fr
[4D_Cubing] <oogroups.com" target=3D"_blank">4D_Cubing@yahoogroups.com>
wrote:
er-left:1px #ccc solid;padding-left:1ex">
solves, but didn't understand everything and preferred
to keep going with the method I was designing - at this
point using an analogue of CFOP was just the thing to
do, seing how intuitive it is when you know it well. Now
that I read it again it looks very clear, and it indeed
looks the same as my own method up to the last layer.
moves as one would need for the 3D cube. To do that, one
can simply "regrip" (rotate) the cube so that the=
face
one turns is always the same. Of course it can be
improved, for instance when using URU'R': a single
regrip in the middle and all moves will cancel... So
it's only an upper bound. I don't know of any gener=
al
method to optimize this step, although I would be
extremely interested.
at giving low-move counts solutions, but still better
than CFOP. I think you can hope for a 40-50 move counts
on average if you know all cases (which I don't),
without optimizing it for too long. What is good is that
each step is rather intuitive, and it can be optimized
and yields extremely good results: Kociemba's algorithm
is derived from it.
seems to me that Matthew Sheerin built his first two
layers like this in his record. It is not as optimized
nor as flexible as Heise, and it would be indeed
interesting to see how well Heise translates into 4D -
but that's far beyond my abilities right now.
=20=20=20=20=20=20
=20=20
--001a113e51121a9468052c873cd3--
From: Melinda Green <melinda@superliminal.com>
Date: Thu, 25 Feb 2016 00:07:04 -0800
Subject: Re: [MC4D] Re: Greetings
--------------040903040200090306070001
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable
On 2/24/2016 9:17 AM, Roice Nelson roice3@gmail.com [4D_Cubing] wrote:
> These are great questions that I wish I had answers for. I will say=20
> that even though I'm marked as the "first no macro solution" in the=20
> Hall of Fame, I've always felt that not using that feature isn't quite=20
> different enough to warrant its own category.=20
You don't mind it being removed? Maybe that makes sense after Andrey=20
convinced us that he had done it earlier in an implementation of his=20
own, but it sure seemed to Don and I like a big milestone at the time.
> I think this just emphasizes what you're saying about computer=20
> assisted categories being fuzzy. Maybe a useful approach here would=20
> be not to find the perfect categories right now, but to figure a way=20
> for them to grow or change easily over time? You seem to have laid=20
> out a good start in any case, with possible beginning categories like:
>
> 1. Human-only
> 2. Computer-assisted (>50% human solve, some computer state space
> searching)
> 3. Human-assisted (>50% computer solve, I question whether there will
> be enough here since implementations would likely just jump to the
> next category like Don did.)
> 4. Computer-only (humans are allowed to write the software)
> 5. Singularity award (computers must write the software)
>
> ...with first and shortests for each. I'd vote to start with 1,2, and=20
> 4 I think. We might want to consider putting 2/4 on a completely=20
> separate page.
You mean 3 & 5?
Regarding 3, if the computer were in charge of a MC4D solution, I agree=20
that human help will only become less useful over time. I was just=20
imagining that the user would still be directing the overall solution=20
but the software would do all of the grunt work. So for example, it=20
might suggest a list of the shortest sequences it finds to move pieces=20
into their proper places. The user can select one and have it perform=20
all the work, or completely override it; sort of like how Google Maps=20
driving instructions find new routes when you've convinced it that=20
you're not going to stick with the previously chosen one.
As for 5, I would dearly love to award that prize someday! I for one am=20
ready to become a pampered pet of our new AI overlords.
>
> On the topic of Go, I've had a background itch to learn for a long=20
> time, and when I read the news about AlphaGo beating the European=20
> champion, I finally decided it was time! I bought a board and some=20
> books, and have been playing a lot the last few weeks. In fact, I was=20
> at my first meeting with the Austin Go Club when your email came=20
> through last night. I had high handicaps and was still getting=20
> crushed, but as expected I'm finding the game extremely elegant,=20
> interesting, addictive, and fun - I can tell this is a hobby that will=20
> stick. What are your predictions for the match with Lee Se-dol? An=20
> Austin player I was talking to last night is betting on the human.=20=20
> I'm rooting for AlphaGo, and am really looking forward to following=20
> the drama!
That is so cool that you've just gotten into my favorite game! I used to=20
really enjoy chess but once I took up Go, I completely switched. While=20
it's a great social activity, I also recommend finding good Go software=20
to play against, especially if it can adapt it's handicap to your rapid=20
progress in the beginning. I started with IGoWin=20
to play hundreds of 9x9 games as fast as I liked. The better I got, the=20
lower handicap I got, and eventually it started taking larger and larger=20
handicaps for itself. That kept me challenged and helped me quickly=20
learn the most important patterns and a sense of what's safe and what's=20
not. I could just try any crazy thing that I'd be curious about but=20
unlikely to try against another person.
While I'm also rooting for AlphaGo, I honestly have no idea who's going=20
to win. AlphaGo's strength when it beat the European champion was=20
definitely not strong enough to beat LSD, but they wouldn't have offered=20
a million dollar prize if they didn't think they would be ready by=20
March. They're certainly going to throw every bit of hardware they can=20
at it, and they're tuning the algorithms and playing millions of=20
training games in the meantime. Here's as much of a prediction as I=20
would bet on: The result will not be 5-0 in either direction, and if LSD=20
wins this time, he will lose 5-0 in a rematch the following year.=20
Exciting stuff!
-Melinda
>
>
> On Tue, Feb 23, 2016 at 7:10 PM, Melinda Green=20
> melinda@superliminal.com
> <4D_Cubing@yahoogroups.com
>
>
>
> Welcome Thomas indeed! There's nothing like breaking one of our
> most cherished records for waking us from blissful slumber.
>
> I'm happy to create new solution categories though I'd like to
> make sure we have a rough idea how to determine where potential
> new methods will fall. Tools like reset, undo & redo are forms of
> computer assistance but they seemed gentle enough to not detract
> from an otherwise pure human solution. Macros OTOH required quite
> a bit more introspection and we ended up categorizing them as
> another valid human solution, though not directly comparable to
> no-macro solutions.
>
> How should we now consider more sophisticated computer assistance
> that goes beyond simple mechanistic aid into some forms of real
> thinking? I definitely want to create a "shortest computer
> solution" category. So far we've only had one by Don Hatch which
> produces solutions with around 1,500 twists. Perhaps when we have
> a more efficient solution, we should create a "first" and
> "shortest" computer solution category pair like the others. I can
> also imagine a time in which a human only guides the broad strokes
> of an otherwise completely computer-driven solution. Would that be
> comparable to this new computer-assisted solution or should we
> then create a new "human-assisted" category in anticipation of our
> obsolescence? These are obviously more philosophical questions
> than practical ones but they might inform our immediate choices.
>
> I'm really hoping that someone will implement an efficient
> computer solution that can be integrated into MC4D and replace
> Don's version which I sadly broke. It was nice having a true
> solution integrated, though 1,500 twists turned out to be just a
> bit too tedious to watch. Something around 200 twists seems much
> more practical, plus it could be merged with the current
> "cheating" solution such that states involving less than that
> number of twists will simply reverse all those twists (ideally
> with some compression which I also broke), and solutions longer
> than that will use the full computer solution. In fact if we do
> this right, the real prize for the shortest computer solution
> could be its integration into MC4D!
>
> And speaking of human obsolescence, are any of you also Go
> players, and have you been following the news regarding the
> upcoming human vs. AI match? Go has been a very difficult game to
> program due to the roughly 200 possible moves at every turn.
> Google's AlphaGo AI recently crushed the European Go champion and
> next month will play against the legendary Lee Se-dol who has
> dominated the game for the last decade. That match has a
> $1,000,000 prize and could well be a watershed moment even bigger
> than when computers became the best chess players in 1996. The
> match begins on March 9th. Details here
>
>
> Happy puzzling!
> -Melinda
>
> On 2/23/2016 9:35 AM, Roice Nelson roice3@gmail.com
>
>> Welcome Thomas!
>>
>> My vote is that we mark this new solution as a first class record
>> in a new category, "shortest computer assisted solve". I've
>> considered trying to use the computer for attacks on the shortest
>> competition in the past, and it'd be great if more folks were
>> motivated to do this. With advances in this area we could, for
>> example, come closer to intuiting what God's Number for the 3^4
>> might be. We don't even have a rough idea of what it is right
>> now, upper OR lower bounds (as far as I know).
>>
>> Congrats on your impressive solve, and happy to have you posting
>> here.
>>
>> Cheers,
>> Roice
>>
>>
>>
>> On Tue, Feb 23, 2016 at 10:21 AM, Thomas Leh=C3=A9ricy
>> thomas.lehericy78@orange.fr
>> [4D_Cubing] <4D_Cubing@yahoogroups.com
>>
>>
>>
>>
>> Indeed. I looked at the wiki page after my first solves, but
>> didn't understand everything and preferred to keep going with
>> the method I was designing - at this point using an analogue
>> of CFOP was just the thing to do, seing how intuitive it is
>> when you know it well. Now that I read it again it looks very
>> clear, and it indeed looks the same as my own method up to
>> the last layer.
>>
>> The last 3D face can be done in at most twice as many moves
>> as one would need for the 3D cube. To do that, one can simply
>> "regrip" (rotate) the cube so that the face one turns is
>> always the same. Of course it can be improved, for instance
>> when using URU'R': a single regrip in the middle and all
>> moves will cancel... So it's only an upper bound. I don't
>> know of any general method to optimize this step, although I
>> would be extremely interested.
>>
>> The human Thitlethwaite is not particularly efficient at
>> giving low-move counts solutions, but still better than CFOP.
>> I think you can hope for a 40-50 move counts on average if
>> you know all cases (which I don't), without optimizing it for
>> too long. What is good is that each step is rather intuitive,
>> and it can be optimized and yields extremely good results:
>> Kociemba's algorithm is derived from it.
>>
>> Block-building methods seem the thing to do indeed. It seems
>> to me that Matthew Sheerin built his first two layers like
>> this in his record. It is not as optimized nor as flexible as
>> Heise, and it would be indeed interesting to see how well
>> Heise translates into 4D - but that's far beyond my abilities
>> right now.
>>
>> Thank you for your answer!
>>
>> Thomas
>>
>>
>>
>>
>
>
>
>
>
>
>=20
--------------040903040200090306070001
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
">
On 2/24/2016 9:17 AM, Roice Nelson " href=3D"mailto:roice3@gmail.com">roice3@gmail.com [4D_Cubing]
wrote:
l.com"
type=3D"cite">These are great questions that I wish I had answers
for.=C2=A0 I will say that even though I'm marked as the "first no
macro solution" in the Hall of Fame, I've always felt that not
using that feature isn't quite different enough to warrant its own
category.=C2=A0
You don't mind it being removed? Maybe that makes sense after Andrey
convinced us that he had done it earlier in an implementation of his
own, but it sure seemed to Don and I like a big milestone at the
time.
l.com"
type=3D"cite">I think this just emphasizes what you're saying about
computer assisted categories being fuzzy.=C2=A0 Maybe a useful approa=
ch
here would be not to find the perfect categories right now, but to
figure a way for them to grow or change easily over time?=C2=A0 You
seem to have laid out a good start in any case, with possible
beginning categories like:
state space searching)
whether there will be enough here since implementations
would likely just jump to the next category like Don did.)
t
with 1,2, and 4 I think.=C2=A0 We might want to consider puttin=
g
2/4 on a completely separate page.
You mean 3 & 5?
Regarding 3, if the computer were in charge of a MC4D solution, I
agree that human help will only become less useful over time. I was
just imagining that the user would still be directing the overall
solution but the software would do all of the grunt work. So for
example, it might suggest a list of the shortest sequences it finds
to move pieces into their proper places. The user can select one and
have it perform all the work, or completely override it; sort of
like how Google Maps driving instructions find new routes when
you've convinced it that you're not going to stick with the
previously chosen one.
As for 5, I would dearly love to award that prize someday! I for one
am ready to become a pampered pet of our new AI overlords.
l.com"
type=3D"cite">
for a long time, and when I read the news about AlphaGo
beating the European champion, I finally decided it was
time!=C2=A0 I bought a board and some books, and have been
playing a lot the last few weeks.=C2=A0 In fact, I was at my
first meeting with the Austin Go Club when your email came
through last night.=C2=A0 I had high handicaps and was still
getting crushed, but as expected I'm finding the game
extremely elegant, interesting, addictive, and fun - I can
tell this is a hobby that will stick.=C2=A0 What are your
predictions for the match with Lee Se-dol?=C2=A0 An Austin play=
er
I was talking to last night is betting on the human.=C2=A0 I'm
rooting for AlphaGo, and am really looking forward to
following the drama!
That is so cool that you've just gotten into my favorite game! I
used to really enjoy chess but once I took up Go, I completely
switched. While it's a great social activity, I also recommend
finding good Go software to play against, especially if it can adapt
it's handicap to your rapid progress in the beginning. I started
with IGoWin
because it was great to be able to play hundreds of 9x9 games as
fast as I liked. The better I got, the lower handicap I got, and
eventually it started taking larger and larger handicaps for itself.
That kept me challenged and helped me quickly learn the most
important patterns and a sense of what's safe and what's not. I
could just try any crazy thing that I'd be curious about but
unlikely to try against another person.
While I'm also rooting for AlphaGo, I honestly have no idea who's
going to win. AlphaGo's strength when it beat the European champion
was definitely not strong enough to beat LSD, but they wouldn't have
offered a million dollar prize if they didn't think they would be
ready by March. They're certainly going to throw every bit of
hardware they can at it, and they're tuning the algorithms and
playing millions of training games in the meantime. Here's as much
of a prediction as I would bet on: The result will not be 5-0 in
either direction, and if LSD wins this time, he will lose 5-0 in a
rematch the following year. Exciting stuff!
-Melinda
l.com"
type=3D"cite">
Melinda Green href=3D"mailto:melinda@superliminal.com">melinda@superliminal.c=
om
[4D_Cubing] < href=3D"mailto:4D_Cubing@yahoogroups.com" target=3D"_blank">4=
D_Cubing@yahoogroups.com>
wrote:
.8ex;border-left:1px #ccc solid;padding-left:1ex">
Welcome Thomas indeed! There's nothing like breaking one
of our most cherished records for waking us from=C2=A0 blissf=
ul
slumber.
I'm happy to create new solution categories though I'd
like to make sure we have a rough idea how to determine
where potential new methods will fall. Tools like reset,
undo & redo are forms of computer assistance but they
seemed gentle enough to not detract from an otherwise pure
human solution. Macros OTOH required quite a bit more
introspection and we ended up categorizing them as another
valid human solution, though not directly comparable to
no-macro solutions.
How should we now consider more sophisticated computer
assistance that goes beyond simple mechanistic aid into
some forms of real thinking? I definitely want to create a
"shortest computer solution" category. So far we've only
had one by Don Hatch which produces solutions with around
1,500 twists. Perhaps when we have a more efficient
solution, we should create a "first" and "shortest"
computer solution category pair like the others. I can
also imagine a time in which a human only guides the broad
strokes of an otherwise completely computer-driven
solution. Would that be comparable to this new
computer-assisted solution or should we then create a new
"human-assisted" category in anticipation of our
obsolescence? These are obviously more philosophical
questions than practical ones but they might inform our
immediate choices.
I'm really hoping that someone will implement an efficient
computer solution that can be integrated into MC4D and
replace Don's version which I sadly broke. It was nice
having a true solution integrated, though 1,500 twists
turned out to be just a bit too tedious to watch.
Something around 200 twists seems much more practical,
plus it could be merged with the current "cheating"
solution such that states involving less than that number
of twists will simply reverse all those twists (ideally
with some compression which I also broke), and solutions
longer than that will use the full computer solution. In
fact if we do this right, the real prize for the shortest
computer solution could be its integration into MC4D!
And speaking of human obsolescence, are any of you also Go
players, and have you been following the news regarding
the upcoming human vs. AI match? Go has been a very
difficult game to program due to the roughly 200 possible
moves at every turn. Google's AlphaGo AI recently crushed
the European Go champion and next month will play against
the legendary Lee Se-dol who has dominated the game for
the last decade. That match has a $1,000,000 prize and
could well be a watershed moment even bigger than when
computers became the best chess players in 1996. The match
begins on March 9th. Details href=3D"http://www.deepmind.com/alpha-go.html"
target=3D"_blank">here.
Happy puzzling!
-Melinda
href=3D"mailto:roice3@gmail.com" target=3D"_blank">lass=3D"moz-txt-link-abbreviated" href=3D"mailto:roice3@gmail.com">roice3@g=
mail.com
[4D_Cubing] wrote:
first class record in a new category, "shortest
computer assisted solve".=C2=A0 I've considered tryin=
g
to use the computer for attacks on the shortest
competition in the past, and it'd be great if more
folks were motivated to do this.=C2=A0 With advances =
in
this area we could, for example, come closer to
intuiting what God's Number for the 3^4 might be.=C2=
=A0
We don't even have a rough idea of what it is
right now, upper OR lower bounds (as far as I
know).
have you posting here.
10:21 AM, Thomas Leh=C3=A9ricy
href=3D"mailto:thomas.lehericy78@orange.fr"
target=3D"_blank">ted" href=3D"mailto:thomas.lehericy78@orange.fr">thomas.lehericy78@orange.f=
r
[4D_Cubing] <
href=3D"mailto:4D_Cubing@yahoogroups.com"
target=3D"_blank">iated" href=3D"mailto:4D_Cubing@yahoogroups.com">4D_Cubing@yahoogroups.com<=
/a>>
wrote:
0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
first solves, but didn't understand
everything and preferred to keep going with
the method I was designing - at this point
using an analogue of CFOP was just the thing
to do, seing how intuitive it is when you
know it well. Now that I read it again it
looks very clear, and it indeed looks the
same as my own method up to the last layer.
twice as many moves as one would need for
the 3D cube. To do that, one can simply
"regrip" (rotate) the cube so that the face
one turns is always the same. Of course it
can be improved, for instance when using
URU'R': a single regrip in the middle and
all moves will cancel... So it's only an
upper bound. I don't know of any general
method to optimize this step, although I
would be extremely interested.
efficient at giving low-move counts
solutions, but still better than CFOP. I
think you can hope for a 40-50 move counts
on average if you know all cases (which I
don't), without optimizing it for too long.
What is good is that each step is rather
intuitive, and it can be optimized and
yields extremely good results: Kociemba's
algorithm is derived from it.
indeed. It seems to me that Matthew Sheerin
built his first two layers like this in his
record. It is not as optimized nor as
flexible as Heise, and it would be indeed
interesting to see how well Heise translates
into 4D - but that's far beyond my abilities
right now.
=20=20=20=20=20=20
--------------040903040200090306070001--
From: Roice Nelson <roice3@gmail.com>
Date: Thu, 25 Feb 2016 12:09:29 -0600
Subject: Re: [MC4D] Re: Greetings
--001a113db22473ee3e052c9c15ae
Content-Type: text/plain; charset=UTF-8
On Thu, Feb 25, 2016 at 2:07 AM, Melinda Green melinda@superliminal.com
[4D_Cubing] <4D_Cubing@yahoogroups.com> wrote:
>
> You don't mind it being removed? Maybe that makes sense after Andrey
> convinced us that he had done it earlier in an implementation of his own,
> but it sure seemed to Don and I like a big milestone at the time.
>
>
Nope, not at all. I feel it is an outdated category at this point, and we
get daily solves by folks on this list of crazy puzzles that represent so
much more!
>
> ...with first and shortests for each. I'd vote to start with 1,2, and 4
> I think. We might want to consider putting 2/4 on a completely separate
> page.
>
>
> You mean 3 & 5?
>
>
I meant 2 & 4, because of worries of 3 and the timeframe of 5. But I was
just brainstorming. 1, 2, & 4 in the existing HOF would be great too.
> Regarding 3, if the computer were in charge of a MC4D solution, I agree
> that human help will only become less useful over time. I was just
> imagining that the user would still be directing the overall solution but
> the software would do all of the grunt work. So for example, it might
> suggest a list of the shortest sequences it finds to move pieces into their
> proper places. The user can select one and have it perform all the work, or
> completely override it; sort of like how Google Maps driving instructions
> find new routes when you've convinced it that you're not going to stick
> with the previously chosen one.
>
That's a nice analogy. Maybe 3 will end up being a bigger category than
I'm imagining!
> As for 5, I would dearly love to award that prize someday! I for one am
> ready to become a pampered pet of our new AI overlords.
>
>
:D
> That is so cool that you've just gotten into my favorite game! I used to
> really enjoy chess but once I took up Go, I completely switched. While it's
> a great social activity, I also recommend finding good Go software to play
> against, especially if it can adapt it's handicap to your rapid progress in
> the beginning. I started with IGoWin
>
> to play hundreds of 9x9 games as fast as I liked. The better I got, the
> lower handicap I got, and eventually it started taking larger and larger
> handicaps for itself. That kept me challenged and helped me quickly learn
> the most important patterns and a sense of what's safe and what's not. I
> could just try any crazy thing that I'd be curious about but unlikely to
> try against another person.
>
>
Sweet, I'm glad to hear I picked a good one because I've been using IGoWin
too! I have about a hundred 9x9 and fifty 13x13 games under my belt at
this point. It's interesting that while ratings and handicaps change, I
always win about half the games. It's tuned quite well for progressing
that way, but I'm not yet to the point where the computer is taking
handicaps! I can already tell I like Go better than chess too, which I've
played quite a bit in the past. (It feels analogous to a
skiing-snowboarding switch I made many years ago. After riding a snowboard
once, I never skied again.)
By the way, I forgot to mention. My username on both Pandanet
to play a game sometime, let me know!
While I'm also rooting for AlphaGo, I honestly have no idea who's going to
> win. AlphaGo's strength when it beat the European champion was definitely
> not strong enough to beat LSD, but they wouldn't have offered a million
> dollar prize if they didn't think they would be ready by March. They're
> certainly going to throw every bit of hardware they can at it, and they're
> tuning the algorithms and playing millions of training games in the
> meantime. Here's as much of a prediction as I would bet on: The result will
> not be 5-0 in either direction, and if LSD wins this time, he will lose 5-0
> in a rematch the following year. Exciting stuff!
>
I like those predictions. I went to the site you sent around, and am
excited that the games will be going at a somewhat reasonable hour here
too. Watching some of them live will be fun, and perhaps a Go-watching
party is in order when the competition is nearing the end.
Roice
--001a113db22473ee3e052c9c15ae
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
On Thu, Feb 25, 2016 at 2:07 AM, Melinda Green perliminal.com">melinda@superliminal.com [4D_Cubing] =
<4D_Cubin=
g@yahoogroups.com> wrote:" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color=
:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
=20=20=20=20=20=20=20=20
=20=20
=20=20=20=20
=20=20
You don't mind it being removed? Maybe that makes sense after Andre=
y
convinced us that he had done it earlier in an implementation of his
own, but it sure seemed to Don and I like a big milestone at the
time.
=A0 I feel it is an outdated category at this point, and we get daily solve=
s by folks on this list=C2=A0of crazy puzzles=C2=A0that represent so much m=
ore!0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);=
border-left-style:solid;padding-left:1ex">
ol>
te to start
with 1,2, and 4 I think.=C2=A0 We might want to consider puttin=
g
2/4 on a completely separate page.
You mean 3 & 5?
f worries of 3 and the timeframe of 5.=C2=A0 But I was just brainstorming. =
=C2=A01, 2, & 4 in the existing HOF would be great too.
<=
/div> 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);bor=
der-left-style:solid;padding-left:1ex">
Regarding 3, if the computer were in charge of a MC4D solution, I
agree that human help will only become less useful over time. I was
just imagining that the user would still be directing the overall
solution but the software would do all of the grunt work. So for
example, it might suggest a list of the shortest sequences it finds
to move pieces into their proper places. The user can select one and
have it perform all the work, or completely override it; sort of
like how Google Maps driving instructions find new routes when
you've convinced it that you're not going to stick with the
previously chosen one.=C2=A0
That's a nice analogy.=C2=A0 Maybe 3 will end up being a bigger categor=
y than I'm imagining!
px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:=
1ex">
As for 5, I would dearly love to award that prize someday! I for one
am ready to become a pampered pet of our new AI overlords.=3D"">
x 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);bo=
rder-left-style:solid;padding-left:1ex">
That is so cool that you've just gotten into my favorite game! I
used to really enjoy chess but once I took up Go, I completely
switched. While it's a great social activity, I also recommend
finding good Go software to play against, especially if it can adapt
it's handicap to your rapid progress in the beginning. I started
with k">IGoWin
because it was great to be able to play hundreds of 9x9 games as
fast as I liked. The better I got, the lower handicap I got, and
eventually it started taking larger and larger handicaps for itself.
That kept me challenged and helped me quickly learn the most
important patterns and a sense of what's safe and what's not. I
could just try any crazy thing that I'd be curious about but
unlikely to try against another person.
I picked a good one because I've been using IGoWin too!=C2=A0 I have ab=
out a hundred 9x9 and fifty 13x13 games under my belt at this point.=C2=A0 =
It's interesting that while ratings and handicaps change, I always win =
about half the games.=C2=A0 It's tuned quite well for progressing that =
way, but I'm not yet to the point where the computer is taking handicap=
s!=C2=A0 I can already tell I like Go better than chess too, which I've=
played quite a bit in the past. =C2=A0(It feels analogous to a skiing-snow=
boarding switch I made many years ago.=C2=A0 After riding a snowboard once,=
I never skied again.)
tion.=C2=A0 My username on both=C2=A0munities/pandanet">Pandanet (live games) and go.com/">Online Go=C2=A0(correspondence games) is roice3.=C2=A0 If anyo=
ne wants to play a game sometime, let me know!
br>x;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-styl=
e:solid;padding-left:1ex">
While I'm also rooting for AlphaGo, I honestly have no idea who'=
;s
going to win. AlphaGo's strength when it beat the European champion
was definitely not strong enough to beat LSD, but they wouldn't hav=
e
offered a million dollar prize if they didn't think they would be
ready by March. They're certainly going to throw every bit of
hardware they can at it, and they're tuning the algorithms and
playing millions of training games in the meantime. Here's as much
of a prediction as I would bet on: The result will not be 5-0 in
either direction, and if LSD wins this time, he will lose 5-0 in a
rematch the following year. Exciting stuff!
around, and am excited that the games will be going at a somewhat reasonabl=
e hour here too.=C2=A0 Watching some of them live will be fun, and perhaps =
a Go-watching party is in order when the competition is nearing the end.iv>
--001a113db22473ee3e052c9c15ae--
From: alexanderrouse@yahoo.co.uk
Date: 01 Apr 2016 13:13:58 -0700
Subject: Greetings
From: alexanderrouse@yahoo.co.uk
Date: Fri, 1 Apr 2016 15:52:50 -0700
Subject: Greetings
--------------070703060906040604000305
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable
Hello Alex,
You're only 90 minutes away from Liam, the latest new solver. Maybe you=20
can start a London Area hyper-puzzling club!
I like photography too and participate in a club, but since I need to=20
take everything one dimension further so it's a 3D photography club of=20
course.
Since you like large puzzles, you may be interested in the 120-cell.=20
It's only fractionally larger than the Examinx. MC4D supports it but=20
Roice's special purpose Magic120Cell program is a better choice for a=20
serious attempt.
Thank you for your kind praise of MC4D and have fun whatever you do!
-Melinda
On 4/1/2016 1:13 PM, alexanderrouse@yahoo.co.uk [4D_Cubing] wrote:
>
>
> Hello
>
> My name's Alex. I'm 42 and I live in England in a small town called=20
> Canvey a few miles to the east of London.
> I work as a picture framer which involves matching many colours of=20
> frame and mounts to pictures to customer taste or my own suggestions.
>
>
> Art is one of my pleasures of which I've done many pictures in mediums=20
> from pencil to acrylics and my subjects are mostly landscapes and also=20
> I do computer generated art. I also like digital photography and going=20
> on long walks snapping shots of landscapes and so get my own source=20
> photos for painting or drawing. I'm also interested in a few=20
> scientific subjects which range from cosmology to qauntum & M-theory=20
> and higher dimensions of space.
>
>
> Finally I also like puzzle solving and can solve 3D rubik cubes of any=20
> size. The largest I done is 10x10x10. I also bea t the minxes up to=20
> examinx size using the Android app Magic Puzzle Pro. The latest puzzle=20
> I beat is the 3x3x3x3 hypercube with Mc4d. I did'nt find this too=20
> difficult to beat and it took me about the same time as it did for me=20
> to beat the 3x3x3 original 3D cube. I did have to put more work into=20
> it though for the same time frame.
>
>
> I used the LBL method which I find to be aesthetically pleasing and a=20
> combination of sequences adapted
> from the original cube to 4D plus new sequences found by=20
> experimentation. I completed the first 2 layers and then made twists=20
> to turn the last face of the third layer all 1 colour. The logic of=20
> this was that if the first 2 layers were solved and the last face on=20
> the third layer was 1 colour then the rest of the puzzle could be=20
> solved by treating the last face like a normal 3D rubiks and this was=20
> where my adaptation from 3D rubik sequences came in.
>
>
> This methodology worked and I have done this a few times so it seems=20
> to be proven. I think MC4D is a great program with lots of puzzles I=20
> look forward to solving and think it would be good for a 4D geometry=20
> class as well as its puzzling value.
>
> Hope to hear soon.
> Alex.
>
>
>
>
>
>=20
--------------070703060906040604000305
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
">
Hello Alex,
You're only 90 minutes away from Liam, the latest new solver. Maybe
you can start a London Area hyper-puzzling club!
I like photography too and participate in a club, but since I need
to take everything one dimension further so it's a 3D photography
club of course.
Since you like large puzzles, you may be interested in the 120-cell.
It's only fractionally larger than the Examinx. MC4D supports it but
Roice's special purpose Magic120Cell program is a better choice for
a serious attempt.
Thank you for your kind praise of MC4D and have fun whatever you do!
-Melinda
=C2=A0
My name's Alex. I'm 42 and I live in England in a small town
called Canvey a few miles to the east of London.
I work as a picture framer which involves matching many
colours of frame and mounts to pictures to customer taste or
my own suggestions.
pictures in mediums from pencil to acrylics and my subjects
are mostly landscapes and also I do computer generated art. I
also like digital photography and going on long walks snapping
shots of landscapes and so get my own source photos for
painting or drawing. I'm also interested in a few scientific
subjects which range from cosmology to qauntum & M-theory
and higher dimensions of space.
cubes of any size. The largest I done is 10x10x10. I also bea
t the minxes up to examinx size using the Android app Magic
Puzzle Pro. The latest puzzle I beat is the 3x3x3x3 hypercube
with Mc4d. I did'nt find this too difficult to beat and it
took me about the same time as it did for me to beat the 3x3x3
original 3D cube. I did have to put more work into it though
for the same time frame.
pleasing and a combination of sequences adapted
from the original cube to 4D plus new sequences found by
experimentation. I completed the first 2 layers and then made
twists to turn the last face of the third layer all 1 colour.
The logic of this was that if the first 2 layers were solved
and the last face on the third layer was 1 colour then the
rest of the puzzle could be solved by treating the last face
like a normal 3D rubiks and this was where my adaptation from
3D rubik sequences came in.
so it seems to be proven. I think MC4D is a great program with
lots of puzzles I look forward to solving and think it would
be good for a 4D geometry class as well as its puzzling value.
Hope to hear soon.
Alex.
=20=20=20=20=20=20
--------------070703060906040604000305--
From: alexanderrouse@yahoo.co.uk
Date: 02 Apr 2016 12:21:54 -0700
Subject: Re: [MC4D] Greetings
From: alexanderrouse@yahoo.co.uk
Date: Sat, 2 Apr 2016 14:41:54 -0700
Subject: Re: [MC4D] Greetings
--------------070607060001020707030803
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: quoted-printable
I should be clear that I'm not strongly encouraging anyone to attempt=20
the 120 cell as it involves a tremendous amount of tedium. It's just a=20
special object that seems to really grab some people. You've got a good=20
idea to start with the dodecahedral prism first. Surprisingly there are=20
many fewer solutions to to that than the 120 cell. Only two solutions=20
lengths 2 and 3. They also look very odd when fully scrambled. Try to=20
guess what they will look like before you scramble them if you haven't=20
seen that already. If you go looking for a really hard 4D puzzle, the=20
face-turning 24 cell=20
that immediately comes to mind. You'd probably want to warm up with some=20
smaller deep-cut puzzles before tackling that monster.
On 4/2/2016 12:21 PM, alexanderrouse@yahoo.co.uk [4D_Cubing] wrote:
>
>
> Hello Melinda
>
> Nice to hear from you again. I just installed the 120 Cell and it=20
> seems quite a puzzle but I think I'll have a warm up with the=20
> dodecahedral prism first. On both puzzles I think my system with the=20
> hypercube can be used but with translation of minx moves instead of=20
> cube moves this time.
>
> I'll be giving Liam a message soon. It's nice to know there's another=20
> solver so close just across the Thames from me and joining so soon=20
> after I did. I may be able to help him with his parity issue though I=20
> don't think it is that with the division of the faces being less than=20
> 4. It's the two colour pieces and I have a couple of simple algorithms=20
> that could help with that.
>
> -Alex
>
>
>
>=20
--------------070607060001020707030803
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable
">
I should be clear that I'm not strongly encouraging anyone to
attempt the 120 cell as it involves a tremendous amount of tedium.
It's just a special object that seems to really grab some people.
You've got a good idea to start with the dodecahedral prism first.
Surprisingly there are many fewer solutions to to that than the 120
cell. Only
solutions for each of edge lengths 2 and 3. They also look
very odd when fully scrambled. Try to guess what they will look like
before you scramble them if you haven't seen that already. If you go
looking for a really hard 4D puzzle, the
-turning
24 cell is the one that immediately comes to mind. You'd
probably want to warm up with some smaller deep-cut puzzles before
tackling that monster.
Hello Melinda
Nice to hear from you again. I just installed the 120 Cell and it
seems quite a puzzle but I think I'll have a warm up with the
dodecahedral prism first. On both puzzles I think my system with
the hypercube can be used but with translation of minx moves
instead of cube moves this time.
=C2=A0I'll be giving Liam a message soon. It's nice to know there's
another solver so close just across the Thames from me and joining
so soon after I did. I may be able to help him with his parity
issue though I don't think it is that with the division of the
faces being less than 4. It's the two colour pieces and I have a
couple of simple algorithms that could help with that.
-Alex
=20=20=20=20=20=20
--------------070607060001020707030803--
From: programagor@gmail.com
Date: 14 Jul 2016 06:52:33 -0700
Subject: Greetings
From: programagor@gmail.com
Date: 14 Jul 2016 06:53:20 -0700
Subject: Greetings
--9brsnTS9ocNmxdhsK3FKVLWiM6a5jwOaA8HvrLV
Content-Type: text/plain; charset=utf-8
Enter your vote now! https://groups.yahoo.com/neo/groups/4D_Cubing/polls/poll/13267602
If you play music, what is your main instrument?
Created by: cutelyaware
1. Theremin https://groups.yahoo.com/neo/groups/4D_Cubing/polls/poll/13267602#122673127 (added by: programagor . July 14, 2016)
Thanks!
--9brsnTS9ocNmxdhsK3FKVLWiM6a5jwOaA8HvrLV
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: 7bit
Thanks!
--9brsnTS9ocNmxdhsK3FKVLWiM6a5jwOaA8HvrLV--