[Help Thread] Algorithm Memorization Discussion

[ExoCorsair]

@ any moderator: can this maybe be moved?

Moved.

@OP: I would suggest learning fingertricks first to get used to certain move sequences that you will see over and over again. I personally used Joël's site (http://solvethecube.110mb.com/fingertricks.html). I learned these tricks before learning F2L/OLL/PLL algs (although you may not want to learn F2L algs).

 

[Krazy Jeff]

Tips on how to remember long algorithms?

I'd like to ask the experts if they could share tips on how they go about learning to remember the long algorithms.

Around December 2007, I started to browse the internet on how to solve the 3x3. I found a youtube video that helped me solve the puzzle. It taught me the following algorithms to help me solve the puzzle: (I noted them here, since I really don't know what system of algorithms I am using.)

3x3 algorithm I use
Fi U Li Ui
Ri Di R D
Ui Li U L U F Ui Fi
U R ui Ri Ui Fi U F
F R U Ri Ui Fi
R U Ri U R U U Ri
U R Ui Li U Ri Ui L
Ri Di R D

Since these algorithms for the 3x3 weren't too long (each), it took me less than 48 hrs to remember the steps by memory without having to read them from a sheet of paper. My best time on the 3x3 is less than a minute. I am not using any speedcubing finger tricks (maybe at a later time, I will learn how to do that--since I think it looks so cool!)

anyway, now to my current problem. I am currently working the 4x4 cube. I was able to get help from another youtube.com tutorial on how to solve it (He is the same person that teaches how to use the 3x3).

anyway, the codes I must use (not always, but sometimes) goes as follows (after I have everything in the proper position, then I just solve the puzzle as I would a 3x3 using the algorithm system I noted above): (Again I don't know the name of the algorithms system I am using--hopefully you recognize them.) I highlighted in red the algorithms I am having difficulty remembering by memory--I have to look at my sheet of paper to be able to follow them.)

4x4 algorithm I use
R* - 2U - R'*
L'* - 2U - L*
L' - R - U - F - U' - L
(solve as 3x3)
D* - R - F' - U - R' - F - D'*
r2 - U2 - r2 - U2 - u2 - r2 - u2
r2 - B2 - U2 - l - U2 - r' - U2 - r - U2 - F2 - r - F2 - l' - B2 - r2

So my question to you guys, is there a method or tip you could give me to be able to remember the steps above I must remember by memory to solve my 4x4 cube? I think since they are many steps that are similar, plus the fact that not always you need to use all those algorithms steps during the puzzle solving, and also having to deal with inner and outer layers, it gets a bit difficult to remember the patterns. Well, at least for me at the moment.

I've had the 4x4 cube for at least a week and I still find myself having to look at my notes to proceed using the algorithms (I highlighted in red above) to continue with my puzzle solving (of course when I need to use those steps--since it's not always that I need to use them all or in a sequence.) Please keep in mind that I know when to use each algorithm when it's called for--that is not my problem, my issue is trying to remember the above three (in red) algorithms by memory, so I don't have to look at my notes). -- this way I look cool among my friends.

I really appreciate any tips you guys can provide and I do appreciate this forum.

 

[ImNOTnoob]

For me, I just track the already solved pieces of the cube as they get seperated, moved around, and rejoined with the other pieces. if the pieces just dont make sense, I just break up the alg into parts and find a pattern for it.

 

[Kian]

practice, practice, practice.

it's easier and faster to remember them with muscle memory than having to think about individual moves. you'll get used to them all.

 

[Krazy Jeff]

Thanks guys for the good tips and the words of encouragement. Just hoping to learn if there is a faster way or a better method.

If you got one, or just want to tell us how you ended up learning all those long algorithms, please note your .02 cents down here (even if it's the same as everyone else that has already replied or a very different method nobody has heard before.--thanks in advance!)

 

[cookingfat]

the first 2 are easy, don't learn the notation, just simply move either 2 sides to the top, spin the top twice and bring back down. I go on what it looks like rather than the actual algorithm. If you watch what happens it becomes clear.

D* - R - F' - U - R' - F - D'*

I split it up like this > D* (R F') U (R' F) D'*. just remember that you move the D* and the beginning and then move it back at the end. That way, you only really need to remember the middle bit (R F' U R' F)

the second red one should actually be r2 U2 r2 U2* r2 u2 (but yours might be a different way, I don't know). This one feels so funky to do, it's really easy to remember.

the last one, r2 - B2 - U2 - l - U2 - r' - U2 - r - U2 - F2 - r - F2 - l' - B2 - r2 is the hardest to remember. I learnt this before I actually got my cube, so I did it over and over again on gabbasoft cube demo which is here > http://www.gabbasoft.com

just do it over and over again on a solved cube. eventually you will see a pattern to it. each person has their own different way of remembering things, I remember it like this, do the first 2 moves, r2 B2 (you will be doing the reverse at the end). after this, I know that before I move a middle slice, I must do a U2. so the next bit goes, U2 l, then U2 r', (slices move towards you) then before you put the r' back, U2 again, the r needs to go in further but this time it will take a U2 & F2 before I can do this, so U2, F2 r, then just F2 and l', then reverse the first 2 moves, B2 and r2 (this probably makes no sense to you at all, but it's just the way I remember it.

I'm no expert cuber by no means, I can only average under 5 minutes on this, but as I'm new it's easy for me to help you too.

 

[Carson]

Using finger tricks will also help you remember them. For example: R U R' done as a fingertrick is really just one motion, where if you do it without fingertricks, it is threee seperate motions. This way, you aren't cutting down on the number of turns, but you are cutting down the amount of movement that you make, which effectively gives you less to commit to muscle memory.

Jeff, that is a cool avatar. That makes me think that a clock mod would be possible on a square-1... hmm, ideas ideas.

 

[Kian]

jeff- you'll be fine, no worries. that oll parity alg for the 4x4 is the most annoying and slowest alg you may ever learn, so try and take solace in that. there's no good way around it, just considerable practice.

 

[Mike Hughey]

Jeff, I had great fun teaching my 7-year-old daughter to solve 4x4x4. For her, the strategy that worked best for teaching her the harder algorithms was this:

1. I showed her the algorithm by holding it so she could see the moves from her perspective, a bunch of times.
2. I talked her through the algorithm a number of times, so she could do it herself.
3. I had her practice the algorithm over and over again, and I watched every move and corrected her whenever she went wrong.

After just a handful of times, she got where I only needed to correct her once or twice even during OLL parity. And after another 20 or so tries, she could get it without me needing to correct her even once.

She doesn't practice very often, sometimes going a month or more between solves. So when she picks it back up, I usually have to go through step 3 with her for 10 or so more tries on each difficult algorithm so she can solve it again. I'll be going through that again soon for the Lexington competition, no doubt. She gets faster every time, though.

 

[cookingfat]

Jeff, I had great fun teaching my 7-year-old daughter to solve 4x4x4. For her, the strategy that worked best for teaching her the harder algorithms was this:

1. I showed her the algorithm by holding it so she could see the moves from her perspective, a bunch of times.
2. I talked her through the algorithm a number of times, so she could do it herself.
3. I had her practice the algorithm over and over again, and I watched every move and corrected her whenever she went wrong.

After just a handful of times, she got where I only needed to correct her once or twice even during OLL parity. And after another 20 or so tries, she could get it without me needing to correct her even once.

She doesn't practice very often, sometimes going a month or more between solves. So when she picks it back up, I usually have to go through step 3 with her for 10 or so more tries on each difficult algorithm so she can solve it again. I'll be going through that again soon for the Lexington competition, no doubt. She gets faster every time, though.

that's amazing your daughter can remember all that stuff. I suppose kids learn stuff like that easier.

How fast can she actually solve it?

 

[Mike Hughey]

Jeff, I had great fun teaching my 7-year-old daughter to solve 4x4x4. For her, the strategy that worked best for teaching her the harder algorithms was this:

1. I showed her the algorithm by holding it so she could see the moves from her perspective, a bunch of times.
2. I talked her through the algorithm a number of times, so she could do it herself.
3. I had her practice the algorithm over and over again, and I watched every move and corrected her whenever she went wrong.

After just a handful of times, she got where I only needed to correct her once or twice even during OLL parity. And after another 20 or so tries, she could get it without me needing to correct her even once.

She doesn't practice very often, sometimes going a month or more between solves. So when she picks it back up, I usually have to go through step 3 with her for 10 or so more tries on each difficult algorithm so she can solve it again. I'll be going through that again soon for the Lexington competition, no doubt. She gets faster every time, though.

that's amazing your daughter can remember all that stuff. I suppose kids learn stuff like that easier.

How fast can she actually solve it?

About 5 minutes now. This is her:
http://www.worldcubeassociation.org/results/p.php?i=2007HUGH04

I'm especially proud of her managing an average at the last competition. On the DNF, she temporarily forgot the OLL parity, but she successfully did the OLL parity on a couple of the other solves.

 

[hawkmp4]

Your daughter is faster than I am at Magic. T_T
What I do if I'm teaching someone an algorithm, is I do the first move of the inverse of the algorithm, and give them the cube to solve. Then I give them the first 2 moves, and so on until they're doing the whole algorithm because they've seen how the algorithm looks, step by step.

 

[Mike Hughey]

Your daughter is faster than I am at Magic. T_T
What I do if I'm teaching someone an algorithm, is I do the first move of the inverse of the algorithm, and give them the cube to solve. Then I give them the first 2 moves, and so on until they're doing the whole algorithm because they've seen how the algorithm looks, step by step.

That sounds like a very nice approach - nice because you get used to how the cube looks after each move. I like it!

 

[hawkmp4]

Yeah, and its not so intimidating because you're not trying to learn a 13 or 14 move algorithm all at once, that was my goal, to try and break up algorithms into smaller chunks so that people learning to cube (who couldn't divide algorithms into fingertricks/triggers) would see it as a more manageable task to learn algorithms.

 

[waffle=ijm]

When I was learning algs specifically for COLL and CMLL, I had one page for each orientation. I'd dedicate about 5 hours getting the pattern memorized and my muscle memory used to the movements.

 

[BarCode]

The biggest thing you need to do is see what the cube looks like after each move. You can sit there starting at the screen and just do the algorithm over and over, which lets you memorize it. The instant you try to use it during a solve, you "forget" how to do it. What I do is read the algorithm, do it a few times glancing back at the algorithm every once in a while to keep it "fresh," then completely ignore the written work, and set up the cases on my cube so I can see exactly what is happening while I'm executing. Not only will you memorize it faster, you won't even need to think about it.

 

[CAT13]

for an algorithm like this:
R U R' U' R' F R2 U' R' U' R U R' F'
I would pair it up into easily performed fingertricks... like this:
(R U R' U') (R' F) (R2 U') (R' U') (R U) (R' F)
After doing that. Just remember each of these little sections one after another. So I would remember R U R' U'. Do that without looking at the algorithm, then just finish the algorithm while looking at the paper. Once you have that down, remember R' F. Do all of R U R' U' R' F without looking at the paper, then finish the rest of the algorithm while looking. Just rinse and repeat for the rest of the algorithm.
Some crazy people think it's fun. I don't. But it gets the job done.

PS: it doesn't really take that long to memorize an algorithm... it is just boring as ****.

 

[TimMc]

1) Break down the algorithm into finger tricks.
2) Execute the algorithm several times using those finger tricks.
3) Say the notation in your head a few times.
4) Note how the algorithm effects the cube and how it solves it.
5) Stop thinking about the notation and just do it.
6) Try executing it on the opposite face with your other hand. ;-)

Also, if you're finding it difficult to execute the algorithm because you're not comfortable with the finger trick then use a different one. However, if this algorithm is *very* commonly used I'd learn it anyway; chances are that there's room for improvement with regards to finger-tricks and that the algorithm itself has potential for fast execution.

Currently my favourite algorithm is Y Permutation. Simply because it only took me 5 minutes to remember. I'm still working on executing it quickly though...

Tim.

 

[Raffael]

For me it always helps when I try to break long algorithms into smaller (logic) parts.

For this one
r2 - B2 - U2 - l - U2 - r' - U2 - r - U2 - F2 - r - F2 - l' - B2 - r2
it would be sth like this:

r2 B2 in the beginning, [--> reverse in the end (B2 r2)]
next part is
l towards you, then U2
then it's
r towards you, U2, r back, U2 again
then it's
F2, right half away from you, F2, left half away from you
now you do the reverse of the beginning part: B2 r2

Until I have a long algorithm lilke this in my muscle memory, I write it down like this:
(r2 B2)
(l U2)
(r' U2)(r U2)
(F2 r)(F2 l')
(B2 r2)

and execute it like 20-30 times.
then i just do normal solves and when i come across this situation and still can't remember, i look at the piece of paper where i've written it down.

hope this made any sense.

another two things:
- learning algorithms becomes a lot easier over time
- Mike Hughey's daughter is truly amazing

 

[Krazy Jeff]

Everyone thank you very very much for all your tips and suggestions. I am gonna try each one to see which one works the best for me.

I believe my avatar is a picture of a popular clock/alarm that was sold in the 80s that looked like a rubik's cube.

Wow, your daughter's memory is impressive!

Thanks everyone once again. For others, please if you come across this thread, leave your .02 cents on how you remember long algorithms.