I need some advice on how to learn OLLs

[Ellis]

So... I would really like to make the transition to full OLL, but I'm experiencing a little problem. What happens is that Ill take a day or two to sit down and practice and memorize new OLLs. Once I get about 10-15 new OLLs memorized Ill stop and go back to speed solves. Ill use these new algs while they are fresh in my mind and Im concentrating on utilizing them. But if I don't continue to learn them and keep going on with speed solves, Ill end up going back to 2-look OLL even when its a case I know. I think this is partly because it takes a few seconds to distinguish if its a case that I know or a similar looking one that I don't know yet and I don't like wasting time if Im going for speed. And if enough time passes, Ill have to go back and re-learn algorithms that I had learned really solidly.

I realize how ridiculous this is, but how can I best remedy it? Should I give up going for speed for like a week?

I was thinking about just going through every OLL and practicing each for about 5-10 minutes. I know that at the end I wont be able to execute each of them when the time comes... but I think that it would really help because I would never come to a case which I had never performed the algorithm for, and I could more easily stop and think "hey, I know I've done this algorithm before" and go back and re-learn it if I need to.

So what do you think the best way to tackle this is? How did you learn OLL? Did you find any problems while learning them where you really had to stop doing what you wanted to do and force yourself to learn a certain way?

 

[Faz]

badmephisto's oll video. nuff said.

 

[McWizzle94]

I would say just learn 1 OLL everyday. Don't do 5-10, because then you get mixed up with all of the algorithms. Just learn 1 a day, and you will be done in 57 days =]

 

[hawkmp4]

Erm...well, 10-15 OLLs in 1-2 days is your problem, if you're having trouble with recognizing them. Slow it down. Do an algorithm over and over (You do know if you do any algorithm enough times it returns the cube to the solved state?) Get it ingrained in your head so deeply you don't have to think about it. Then, after you learn even one new algorithm, go back and review (do each algorithm 10 times at least, just to make sure the muscle memory is still there for the algorithms you've already learned)

 

[Ellis]

badmephisto's oll video. nuff said.

thats where I went to first

Do an algorithm over and over (You do know if you do any algorithm enough times it returns the cube to the solved state?) Get it ingrained in your head so deeply you don't have to think about it. Then, after you learn even one new algorithm, go back and review (do each algorithm 10 times at least, just to make sure the muscle memory is still there for the algorithms you've already learned)

Yea I know... when I would do 10 algs, thats how I would do it.... I would spend almost 10 minutes on each algorithm until I fully got the muscle memory down, and even go back to the first alg and go through them all again. The problem would come when I went to speedsolve, I knew in the back of my head that there were so many OLLs that I hadn't done yet that I just automatically went to 2-look OLL without giving a full pause to see exactly what case it was.

Should I be trying to integrate them into speedsolves or should I not go for speed at all?

something tells me that no matter how well I learn them... that If I dont have a good number of them down that I will inevitably go back to 2-look if im going for speed, which is where im failing

 

[hawkmp4]

Oh. Well, it'll increase your times, but even when going for speed, FORCE yourself to look and see if its a 1-look OLL that you know. If it is, obviously, do it. 2-look is a habit that you have and if you don't fight it a little its not going to change.

 

[ScottKidder]

Take this as an oppurtunity to do a lot of slow turning solves so you can work on F2L. After finishing F2L see if it's an OLL you know, if it is do it, then scramble. If not, scramble and do it again. slow turning solves are veerry good.

 

[williamg]

Erm...well, 10-15 OLLs in 1-2 days is your problem

1-2 days ought to be enough for anybody.

 

[Ellis]

2-look is a habit that you have and if you don't fight it a little its not going to change.

yea im only forcing myself to check for so long, I guess from now until I get them all down I really need to be forcing everytime I come to the last layer

Take this as an oppurtunity to do a lot of slow turning solves so you can work on F2L. After finishing F2L see if it's an OLL you know, if it is do it, then scramble. If not, scramble and do it again.

Thats what Ive been doing all day today actually- slow turning through f2l, and completely skipping PLL, re-scrambling if I dont know the OLL. Would it be a good idea to literally do this until Ive learned the majority of OLLs?

btw, im new to the forum and I just want to say its awesome. People are always on here and are willing to help... i mean, hawk, you replied 4 minutes after my second post, thats so cool

 

[Ellis]

Erm...well, 10-15 OLLs in 1-2 days is your problem

1-2 days ought to be enough for anybody.

Thats kinda what I was thinking... OLLs are easier than PLLs and a lot of them are similar. thats kinda why I thought... well just go through all of them so that Ive at least done them so I can easily come back to any one whenever I get to the last layer

 

[hawkmp4]

I don't doubt your ability to memorize them, not at all- I just think it gives rise to problems like these, actually recognizing and executing the algorithms.
But whatever you end up doing, let us know how it goes.

 

[DavidWoner]

Erm...well, 10-15 OLLs in 1-2 days is your problem

1-2 days ought to be enough for anybody.

Thats kinda what I was thinking... OLLs are easier than PLLs and a lot of them are similar. thats kinda why I thought... well just go through all of them so that Ive at least done them so I can easily come back to any one whenever I get to the last layer

of course you can learn the algorithms in a few days. learning the algs is not the problem, nor is commiting them to muscle memory.

the problem is that when you see a case that you just learned, your brain will automatically associate it with the movements for 2-look. It takes a while for your brain to rewire these synapses to perform the 1-look alg. it is impossible for your brain to rewire the synapses for 15 algs in only two days.

you need to learn them in pairs or quads(like the squares and the two associated lightning bolts). ive been doing 3-4 every week and I have had no problems with recognizing them. I practice them for about an hour total over two days, and then i have them down.

 

[Ellis]

you need to learn them in pairs or quads(like the squares and the two associated lightning bolts). ive been doing 3-4 every week and I have had no problems with recognizing them. I practice them for about an hour total over two days, and then i have them down.

That actually sounds like a good idea, learning a certain group every week. I think I might actually try it like that and not rush it as much. Seems like it could be good if I were do it like that along with an hour or so of slow turning to make sure Im retaining all the algorithms that I've learned previously. thanks for the idea vault, and everyone else. I think Im set on that for now, if anyone else had any advice, Id like to hear it but I think Im ok as of now. thanks again!

 

[amostay2004]

take things slowly..seriously
memorise 2-3 algos in a day..probably similar looking ones...
like for a couple of days try and memorise the algos for small L shapes..or only a dot in the centre..stuff like that
it is gonna make you slower in the recognition in the beginning but like they said, FORCE yourself to recognise..heh
one thing about getting faster is you first have to go slower...

 

[Kenneth]

My advice is don't, go for COLL instead, it's much easier to learn (despite what most people says), is faster (VH, COLL, EPLL compared to OLL, PLL) and got lesser cases.

All my experiences, naturally I can't prove it

 

[Skrato]

I'm in the middle of changing from 2-look to 1-look as well, I go by learning all the OLLS that have a dot, then all that have a line, then all that have the funny thing made into a cross but FURU'R'F'. That way, you immediately know if you have gone over that algorithm. E.G. It took me one day to learn all dot algorithms, now if a dot appears with no other edges, I know the OLL so immediately know the algorithm I need. If a line comes up with no edges I know that as well. Then you can still speedcube, get those algorithms into your head before learning more.

 

[Ellis]

My advice is don't, go for COLL instead, it's much easier to learn (despite what most people says), is faster (VH, COLL, EPLL compared to OLL, PLL) and got lesser cases.

could you refresh my memory on exactly what COLL is? I know it stands for corner orientation of the last layer, but I dont really see how that would help me. I keep thinking thats when you orient the corners as you're doing the last f2l slot but thats MGLS right? What exactly would be the process of a solve using fridrich+COLL?

EDIT: Actually I think that MGLS is orienting the edges with the last f2l slot, not corners... I probably should be looking this stuff up

 

[hawkmp4]

COLL orients and permutes the last layer corners at the same time. 41 cases, if I remember correctly. MGLS has to deal with the last F2L slot, anything more than that I don't remember >.<
Fridrich + COLL would be a normal cross and first 3 F2L pairs, some kind of edge control for the 4th pair so that you have all LL edges oriented, COLL and the PLL of edges only.

 

[Ellis]

COLL orients and permutes the last layer corners at the same time. 41 cases, if I remember correctly.

Yea... I dont really know if thats something I want to pursue. While the idea of only having to do an edge permutation PLL to finish seems nice... I just cant see it being faster than OLL+PLL for me. For one, I think the recognition would be more difficult, and what happens when you have no edges correctly oriented when you come to your last f2l slot? wouldn't there have to be algorithms for something like that? And trying to orient LL edges when im filling in a BL or BR slot last doesn't seem too easy.

Plus, I've already got about 16 OLL algs down, so 41 new algs which completely redefines my LL doesnt seem too appealing, especially when I've got to basically throw away ~17 PLL algs that Ive already learned. I'm kind of over that idea unless someone has a seriously strong argument to convince me.

 

[hawkmp4]

Yeah, me too, the only reason I could see it used with Fridrich is as a stepping stone to ZB.
If I'm not mistaken, COLL was originally conceived for Petrus, yeah?