[Help Thread] Algorithm Memorization Discussion

[Selkie]

The system that suits me and is the way I learned full PLL/OLL was to learn the recognition first. I would do a group at a time. So after all edges oriented I did the dot cases so I learned to recognise all the dot cases initially and gave them names too to aid the recognition. Once this was done I would learn the notation for a single algorithm before work and during the morning I would repeat it in my head. The first time I would actually do the algorithm was at lunchtime where I would drill it for 5 minutes. Then in the evening I would drill it for another 5 minutes, it if had a mirror I would also drill that, as well as drilling those algs learned the previous 3 days. By the time the notation I had initially learned had dropped from short term memory the algs were already burned into long term muscle memory. The only place you can store a large number of algs easily is muscle memory, not by notation.

Using this system I never had a recognition recall problem and have adopted exactly the same system in the past week learning square 1 algs. I have learned all VHF2L and half of COLL (halted until after my first comp)

By learning how the cases looked first, by utilising memory at a time I could not cube and by only learning 1-2 a day I found a system that worked for me. But you may find one that works for you better. The main challenges are ensuring you can link cases to algs if you learn too many at a time and ensuring that each one is firmed memorised by muscle memory.

 

[Tao Yu]

I learned the OLLs and PLLs by seeing how they break up the restore the F2L. For example I learned the "T perm OLL" by seeing it took a pair out with RUR' and it inserted it again with the sledgehammer (R'FRF'). I made a few of my own OLLs this way. They were easier to learn but not as good as the standard OLL algs
Later the algs got into my muscle memory

 

[Jorghi]

umm I just learned them.. by doing them until it resolves the cube

 

[tx789]

JUst do it over and over again in parts can help

 

[cubersmith]

And?

If you do learn algorithms then you really should learn them only 1 or 2 at a time. I rarely learn algorithms but I noticed about a week ago that I had only ten OLLs left to learn. I learned them all that day and the algorithms stuck to me fine, the problem was that I can't remember what algs go to what case. Only about a day ago did I get use to what algs go with what case.

So the moral of the story is to learn algorithms slowly so you can get your barings with them.

BTW: Don't type like a damn troll. Even if you don't have proper grammar(I don't), you can still make it look ok if you correctly capitalize and punctuate.

Now that's some irony right there.

 

[EnterPseudonym]

BTW: Don't type like a damn troll. Even if you don't have proper grammar(I don't), you can still make it look ok if you correctly capitalize and punctuate.

Obvious troll is obvious

 

[Bizarro]

I'm learning full PLL right now, and I find that two a day is working for me. I try to find a way to help me recognize the case, then continually do the algorithm until it sticks.

 

[guitarman960]

i have learned the 7 2 look OLL algs. in 2 days

 

[uberCuber]

The important thing is that, as you learn more algs, you will get better at learning algs. I remember that learning the PLLs took me forever. I would feel like I had done plenty of repetitions of an alg and knew it well enough, but then after taking a break and returning to the cube later, I had forgotten it in part and would need to refer to the wiki page once more. I didn't learn two a day; I learned two every few days.
Then, when I practiced OH solving and computer solving for the first times, I could never remember algs at all. That's when I realized how it is better to learn algs by a combination of muscle memory, and actually looking at how certain pieces are moving around. By recognizing patterns in the motion of pieces, you have a consciously-remembered picture of an alg to reference, rather than pure muscle memory, which happens to not be perfect (though it does get better as well, with practice).

 

[michaelfivez]

The important thing is that, as you learn more algs, you will get better at learning algs. I remember that learning the PLLs took me forever. I would feel like I had done plenty of repetitions of an alg and knew it well enough, but then after taking a break and returning to the cube later, I had forgotten it in part and would need to refer to the wiki page once more. I didn't learn two a day; I learned two every few days.
Then, when I practiced OH solving and computer solving for the first times, I could never remember algs at all. That's when I realized how it is better to learn algs by a combination of muscle memory, and actually looking at how certain pieces are moving around. By recognizing patterns in the motion of pieces, you have a consciously-remembered picture of an alg to reference, rather than pure muscle memory, which happens to not be perfect (though it does get better as well, with practice).

Yeah I'm unable to perform some algs slowly enough so I can actually see what I'm doing so I can use it for one handed wich sucks, I always have to look the algs up again at home

 

[Arkwell]

Memorizing Algorithms

Guys, at what rate & how many new algs do you memorize in a day. What is your memorization process?

 

[Mal]

I memorize one a day. I just repeat it, and learn it intuitively. I just remember it like this: bring that up there and bring that there.

 

[Petezorzz]

I remember the steps of how the alg works, not the letters.
For e.g. T Perm:
R U R' U' = Sexy move
R F R2 U' R' = Putting a different F2L pair in place
U' = Random, but easy to remember
R U R' F' = Putting the F2L pair into correct place

Easy as that ;]

 

[a cool guy]

One a day usually does the trick for me. I find that if I try to memorize more than one a day, I usually end up remembering none since I spread my practice of the algs too thin.

 

[Sarahjdes]

It always takes me forever to learn algorithms. I'm currently learning 2-look OLL and PLL (gotta start somewhere!).

For the Blinker, I'm using this alg : (R U2) (R2 U') (R2 U') (R2 U2 R)

R U2 : I put one of the top edge piece on the 2nd layer
R2 U' : I end up with a T on the side
R2 U' : I end up with a vertical line
R2 U2 R : I put my edges back on the top.

Weird, hard to explain, and maybe not that efficiant, but it works for me!

 

[fastcubesolver]

When I first see a new alg, I repeat it a few times, and see how it works visually. Then I do an avg12 with whichever cube I'm learning algs for, and do the alg I just learned between every solve. Seems to work pretty well.

 

[cubecraze1]

i find not to remember more than one a day you still remember the algorithm but you don't remember what it does
(i remember learning 3 g perms in one day)

 

[Michael1026]

Depends on the algorithm. It it's something hard, I usually work with it for a couple days, if it's extremely easy, I'll learn about 2 a day.

 

[Kirjava]

I wouldn't advise learning more than 48 in a single day. After doing this it's also useful to let them sink in for one or two days before learning any more.

For learning, just find fingertricks for an alg and by that point you should already know it.

 

[onlyleftname]

I learned like 6-7 a day on average. Protip, memorize the algorithms/repeat them right before you sleep. It commits it to long term memory.