• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Is there really no 2-gen notation?!

ostracod

Member
Joined
Nov 21, 2007
Messages
135
Location
USA
I just did a Google search for "2-gen notation" hoping to find some useful resources on how to abbreviate sequences such as:

(R U2 R' U) R2 U2 (R' U' R U') R2

The repetition of "R" and "U" in such a sequence seems useless and distracting to what the meaning of the sequence truly conveys. The solver KNOWS to alternate between R and U faces, but needs to know only the direction in which to turn each face.

To my amazement, the Google search did not yield any results. 2-gen algorithms are fairly common (at least nowadays they are), so I would have expected someone to think of a better scheme for conveying them.

I think for sure this needs to be done. A new notation would not only make such 2-gen algorithms take up less space on a page, but they would also be easier to memorize. I NEVER memorize 2-gen algorithms by thinking "AHR YEW TWO AHR INVERTED YEW AHR TWO YEW TWO......" This is just silly. I always end up relying on following corner-edge pairs, like I do for most winter variation algorithms, or I do the algorithm over and over until it makes a bond in my finicky muscle memory. Muscle memory is my least favorite kind, since the SLIGHTEST amount of conscious thought about the sequence results in its failure. In order to let your muscle memory work, you have to turn your brain off and trigger a reflex, and the reflex is not guaranteed to turn out correctly.

As I have started to learn CLS, I have learned that many of the algs break up and manipulate corner-edge pairs in ways that I cannot easily follow. For this reason my WV strategy of memorization becomes ineffective. I intend to use my proposed notation for learning CLS...

In addition, I don't think about turning faces in terms of clockwise and counterclockwise rotation. Although I understand the principle of clockwise versus counterclockwise, I typically think of individual moves as "turn the right side up", "turn the top side left", etc etc: a face plus the DIRECTION in which to turn that face. When I see "L", it translates in my head to "turn the left side down" because I MEMORIZED how to translate it. I know it really means "turn the left side clockwise", but when I think that way I can't quickly find out which way to turn the face. This preference of mine may originate from the first notation I learned (a face-direction notation), or it may be that I find clockwise/counterclockwise to be more awkward. Again, I don't know if you all find the same to be true, but from my experience it is more natural to think of moves like "turn the bottom face left" instead of "turn the bottom face counterclockwise".

So to summarize the problem:
1. Standard notation conveys a lot of useless information about 2-gen algorithms.
2. Such a clunky notation may force the cuber to rely on muscle memory which is not always reliable.
3. For at least SOME people, the notion of clockwise/counterclockwise is awkward.

So here is my proposal for a 2-gen algorithm notation system.
- Each algorithm consists of a sequence of characters.
- Each character represents a face, either U or R, and the direction in which to turn that face.
- The characters are:
R = turn the top side to the RIGHT (= U')
L = turn the top side to the LEFT (= U)
U = turn the right side UP (= R)
D = turn the right side DOWN (= R')
2 = turn the right or top side TWICE (= R2/U2); the face depends on whether the preceding face was R or U.

As an example, the clunky algorithm:

(R U2 R' U) R2 U2 (R' U' R U') R2

becomes:

(U2DL)22(DRUR)2

This is, at least for me, much easier to memorize than the original, and is much more useful for capturing what I need to know to perform the 2-gen algorithm. Again, I KNOW I am alternating between the R and U faces (because it is 2-gen, after all), so all I need is the direction in which to turn each face, which the condensed form succinctly gives me.

I know it is possible to condense the notation further by including more letters of the alphabet, but then the problem becomes decompressing a memorized code into moves, which takes time for the solver. I also know that using the letters U, D, R, and L to represent directions may seem confusing, since they are normally used to represent a face, but this is simply my preference.

I want your feedback- what kind of 2-gen notation would you propose? How do you like to memorize 2-gen algorithms? The sky's the limit. ;)

(This message should provide people a lengthly read! XP)
(The example algorithm is used for solving a CLS case. I found it at: http://cube.garron.us/MGLS/)
 

Hammond

Member
Joined
Dec 2, 2009
Messages
20
Just seems confusing and unnecessary to me. I mean why describe something that's two gen, with multiple notation, that's just over complicating it, imo.
 

Zarxrax

Member
Joined
Jan 7, 2009
Messages
1,282
Location
North Carolina
It seems complicated to me, but I guess because I would have to spend some time getting used to it.
Instead of (U2DL)22(DRUR)2, why not something like:
(↑2↓←)22(↓→↑→)2
 

jms_gears1

Premium Member
Joined
Jun 10, 2009
Messages
1,303
WCA
2009MAUP01
YouTube
Visit Channel
lol honestly heres how i would do it:
(R U2 R' U) R2 U2 (R' U' R U') R2
[R,U](1,2,',1)2,2,(''1')2

1 means one turn clockwise
2 means 180
' means 3 or counterclock wise

the thing is people would have to learn it. and its easier to just keep it how it is, however if your learning a bunch of 2GAs then use your way that way its easier to keep in one place.

EDIT: the [F1,F2] would be according to which face you move fisrt so if it was RUR'U' itd be [R,U] if it was URU'R' itd be [U.R]
 
Last edited:

ostracod

Member
Joined
Nov 21, 2007
Messages
135
Location
USA
For those who prefer clockwise/counterclockwise, yours might be better, jms_gears1.

I am not trying to make any "official" notation, but it would be nice for each person to have their own way to list and remember their algorithms. If one finds the need to learn some 30 2-gen algorithms, they better have some good way to do it. The standard notation does not cut it.

I think the standard notation should be used as an "Esperanto" of sorts for cubers to share their algorithms with others. But for their own convenience they would use a condensed form when memorizing.
 

Ryanrex116

Member
Joined
Aug 9, 2008
Messages
464
WCA
2008JEWR01
I have some shortenings for 2 gen algs, but not for every combination. Here are the ones I can think of now:

SM=R U R' U' (Sexy Move)

RT= R U R' (Right Trigger)
RT'= R U' R' (R Trigger inverse)
R'T= R' U R
R'T'= R' U' R
 

DavidWoner

The Punchmaster
Joined
May 7, 2008
Messages
3,936
Location
Kansas City, MO, USA
WCA
2008WONE01
YouTube
Visit Channel
I am not trying to make any "official" notation, but it would be nice for each person to have their own way to list and remember their algorithms. If one finds the need to learn some 30 2-gen algorithms, they better have some good way to do it. The standard notation does not cut it.

Who learns by notation? If you are learning by muscle memory or by following pieces then different notation is not really needed. When I am learning an alg I only need to look at it 5 times, max.
 
Last edited:

qqwref

Member
Joined
Dec 18, 2007
Messages
7,834
Location
a <script> tag near you
WCA
2006GOTT01
YouTube
Visit Channel
Overall impression: This is an interesting idea, and if it works for you that's great, but it's not suitable for most people and IMO is not an improvement over the normal notation.

I think for sure this needs to be done. A new notation would not only make such 2-gen algorithms take up less space on a page, but they would also be easier to memorize. I NEVER memorize 2-gen algorithms by thinking "AHR YEW TWO AHR INVERTED YEW AHR TWO YEW TWO......" This is just silly.
Yeah, uh, nobody memorizes algs that way. That's why the normal notation is still used for everything. Try saying it out loud like that for any other alg - it'll sound just as silly. People don't memo algs by memorizing the move notations and then executing them one by one, but instead by performing the alg a lot until they have internalized how it feels to do the moves (muscle memory) and/or how the pieces and blocks on the cube move around. Triggers - sequences of a few moves that are done in essentially one motion and memorized as a group - are a big part of this. Let's look at your alg:
(R U2 R' U) R2 U2 (R' U' R U') R2
Here's one way to think about it:
(R U2 R') (U R2 U2) (R' U' R) (U' R2)
First thing is the R U2 R' trigger, so that just moves the edge and corner out of the way of the R face. Then U R2 U2 moves the DR group over to the far left while keeping those pieces on U; the R' U' R trigger joins up the DFR corner with that group, and the U' R2 moves everything back into the F2L. So you don't know the moves of the alg, but instead you have an idea of what you do to solve the case. There are a few ways to do each alg, generally. You can also just learn the entire thing with muscle memory, by doing it over and over until your hands know what to do, but that's more mistake-prone (if you mess up, good luck fixing it).

In addition, I don't think about turning faces in terms of clockwise and counterclockwise rotation. Although I understand the principle of clockwise versus counterclockwise, I typically think of individual moves as "turn the right side up", "turn the top side left", etc etc: a face plus the DIRECTION in which to turn that face. When I see "L", it translates in my head to "turn the left side down" because I MEMORIZED how to translate it. I know it really means "turn the left side clockwise", but when I think that way I can't quickly find out which way to turn the face.
Personally, I don't think about direction or counterclockwise or clockwise at all. I know what an L is, I know what an L' is, I know what a B is, etc. all without thinking about it, just from having used them so much in scrambles and such. So you could tell me "do L U L'" and I'd do it easily, but if you told me "turn L clockwise and U clockwise and L counterclockwise" I'd have to figure it out. What I'm saying is that the problem isn't in the notation but in the way you convert notated moves to physical moves - it should be unconscious and immediate, not a matter of figuring out which way clockwise is.

I know it is possible to condense the notation further by including more letters of the alphabet, but then the problem becomes decompressing a memorized code into moves, which takes time for the solver. I also know that using the letters U, D, R, and L to represent directions may seem confusing, since they are normally used to represent a face, but this is simply my preference.
Yeah, see, that's the problem with nonstandard notations. The normal one might have a few extra moves compared to an ultra-condensed or specialized one, but the advantage is that it is totally uniform and (for most people) does not actually require any translating, decompression, etc. An R' isn't converted into "arr prime" or "right face counterclockwise", it's converted into the idea in my head of what an R' is, and your notation would have exactly the same effect once I learned it. The key in all cases is that you memorize the moves themselves (singly, in combination, or by feel), not the notation itself. Also, if you ever need the notation for an alg written out, you don't have to have it memorized - you can just do the alg slowly and write down the names of all the moves you do. Again this is pretty much instantaneous if you are familiar with the notation you're using.
 

ostracod

Member
Joined
Nov 21, 2007
Messages
135
Location
USA
Perhaps if the algorithms were easier to remember, people would memorize them. It might help to use an easy algorithm in conjunction with muscle memory- After a lot of repetition, muscle memory works OK, but it would be helpful to have some recall aid for the first week or so of using the algorithm.

Another problem to keep in mind is keeping multiple 2-gen algorithms from becoming mixed up when they are recalled. Because each 2-gen algorithm may have such a restricted set of moves, you may fool your muscle memory into leading from one algorithm into another, or confusing bits and pieces between each. By memorizing algorithms, I think it would be easier to keep the algorithms untangled.
 

ostracod

Member
Joined
Nov 21, 2007
Messages
135
Location
USA
Personally, I don't think about direction or counterclockwise or clockwise at all. I know what an L is, I know what an L' is, I know what a B is, etc. all without thinking about it, just from having used them so much in scrambles and such. So you could tell me "do L U L'" and I'd do it easily, but if you told me "turn L clockwise and U clockwise and L counterclockwise" I'd have to figure it out. What I'm saying is that the problem isn't in the notation but in the way you convert notated moves to physical moves - it should be unconscious and immediate, not a matter of figuring out which way clockwise is.

I had the feeling someone would have this reaction. Experienced cubers, such as you, Qqwref, have used the standard notation quite frequently for scrambles and so forth. As soon as you see a sequence in standard notation, you can rattle it off on the cube without any intermediary translation or thought. Even though I have used the standard notation to memorize maybe 50 algorithms (over a couple years, not all of which I remember), I don't use the notation for scrambles. My understanding (which is certainly prone to inaccuracy) is that those who compete in official competitions tend to use generated scrambles frequently, and thus they can fluently scramble the cube based on a set of moves in standard notation. I virtually never do this, because I do not compete in competitions; I am a casual speed cuber. So, in effect, I am less experienced at standard notation.

Not being completely fluent in standard notation, I do not have the advantage of instantly turning a symbol into a move. Thus using a notation system which makes more natural sense to me- namely, a direction based system- helps me execute an algorithm. And since 2-gen algorithms are especially based on the way in which the R or U faces are turned, they seem to be prime subjects for such a notation.

Regardless of comfort with standard notation, a more compact notation makes it easier to turn an algorithm into a kind of chant- "Up Two Down Left, Two Two, (Down Right Up Right), Two". If it becomes possible to memorize this kind of line, then you need to rely less heavily on muscle memory. THEN after you have done the algorithm maybe 20-40 times using the chant, it will become engraved in your muscle memory more soundly... At least I think.
 

qqwref

Member
Joined
Dec 18, 2007
Messages
7,834
Location
a <script> tag near you
WCA
2006GOTT01
YouTube
Visit Channel
Even though I have used the standard notation to memorize maybe 50 algorithms (over a couple years, not all of which I remember), I don't use the notation for scrambles. My understanding (which is certainly prone to inaccuracy) is that those who compete in official competitions tend to use generated scrambles frequently, and thus they can fluently scramble the cube based on a set of moves in standard notation. I virtually never do this, because I do not compete in competitions; I am a casual speed cuber. So, in effect, I am less experienced at standard notation.
That's right, but even as a casual speedcuber it's a good idea to use computer-generated scrambles, and pretty much all scramble generators use the normal notation. At the very least, it gives you consistent randomization, and is a very good way to keep track of times. You're not the only person I've heard say something like 'I'm not a serious speedcuber, so I don't need to do X' (where X is something most speedcubers do), but I honestly think that even though you don't consider yourself a serious solver it is still helpful to do stuff like this. It usually doesn't take a lot of effort to switch - and things like computer scramblers wouldn't be used by so many people if they weren't helpful. Getting down to the 20 second range IS something you can do with casual practice, but if you're not willing to make small changes for the better in the way you practice, you're not going to get there. Being familiar with notation does not require an hour of practice a day or anything like that, it is just something that develops over time if you use it, and IMO it's an extremely helpful skill if you're interested in fewest moves, theory, or learning algs.
 

ostracod

Member
Joined
Nov 21, 2007
Messages
135
Location
USA
I have not seen the difference between a scramble generated by a lot of random finger tricks and one made by a computer. But if you think it will help me as a cuber, then I'll give it a try.
 

blade740

Mack Daddy
Joined
May 29, 2006
Messages
851
WCA
2007NELS01
YouTube
Visit Channel
I think a trigger-based 2gen notation would be better. Define names for common triggers (for example, A=RUR', B=RU'R', C=RU2R', etc) and then link them with U/R turns in easy-to-memorize ways. For example, sune would be something like A U C.
 

miniGOINGS

Member
Joined
Feb 27, 2009
Messages
3,049
I have some shortenings for 2 gen algs, but not for every combination. Here are the ones I can think of now:

SM=R U R' U' (Sexy Move)

RT= R U R' (Right Trigger)
RT'= R U' R' (R Trigger inverse)
R'T= R' U R
R'T'= R' U' R

Yea that makes a lot of sense, that's how I learn most algorithms.
 
Top