1-look, 2-alg: an intermediate step

[Zarxrax]

A lot of people hate learning algorithms, me included. A lot of the reason for it, I think, is that it can be really hard to keep all the cases straight, AND have to memorize the move sequences. So then it hit me! Why even bother learning to recognize the cases along with the algorithms. I could learn them separately, and it would make things much easier.

For example, let's take OLL. For each case, just memorize the 2 2-look OLLs needed to solve it. I think it should be possible to easily memorize all 57 cases in just a few days.
Now, you've learned how to recognize every OLL case, so learning full OLL is just a matter of attaching a new alg to each one.

I think this has a lot of advantages to it.
- First off, as a stepping stone to full OLL, it can give you a small speed improvement without the effort of having to learn full OLL right away. If you just learn cases and algs one at a time, I think it would probably slow down your times overall, until you finish learning all the cases and algs.
- Next, it makes the learning easier, because you are only learning 1 new piece of information at a time, rather than 2.
- Finally, the recognition part is usually much more difficult to practice than the execution. You can easily get your algs fast by just spamming them repeatedly for a few hours. But you can only really practice your recognition by solving the cube, or using some sort of OLL trainer, or something. By learning recognition first, by the time you get around to learning the algs, you have already had lots of chances to practice your recognition.

This method isn't only for OLL, but could be used for anything that has a 2-look version. PLL, CMLL, maybe even zbf2l?

 

[4Chan]

Lucas had something similar on his website?

http://cube.garron.us/algs/compOLL/index.htm
Lol, 2006.

 

[Zarxrax]

I'm apparently a little late to the party then!
But, I think more people should try this. I never see anyone talking about it.

 

[James Ludlow]

Like this?

 

[Anthony]

But, I think more people should try this. I never see anyone talking about it.

People don't talk about it because it's slower than full OLL and people generally want to be fast. If you're willing to put in the work to learn this "system", you might as well just learn full OLL. You make it sound like it is insanely difficult when it really isn't. You can learn most algs by simply tracking F2L pairs and such. Seriously, it's not that hard.

 

[Zarxrax]

But, I think more people should try this. I never see anyone talking about it.

People don't talk about it because it's slower than full OLL and people generally want to be fast. If you're willing to put in the work to learn this "system", you might as well just learn full OLL. You make it sound like it is insanely difficult when it really isn't. You can learn most algs by simply tracking F2L pairs and such. Seriously, it's not that hard.

And you are making it sound like I proposed this as an alternative to full OLL, which I clearly didn't.

 

[Anthony]

But, I think more people should try this. I never see anyone talking about it.

People don't talk about it because it's slower than full OLL and people generally want to be fast. If you're willing to put in the work to learn this "system", you might as well just learn full OLL. You make it sound like it is insanely difficult when it really isn't. You can learn most algs by simply tracking F2L pairs and such. Seriously, it's not that hard.

And you are making it sound like I proposed this as an alternative to full OLL, which I clearly didn't.

You suggested to use this as a stepping stone, but imo it's not worth the effort to learn all the combinations before tackling full OLL.

 

[Zarxrax]

You suggested to use this as a stepping stone, but imo it's not worth the effort to learn all the combinations before tackling full OLL.

But why not? I think you are overestimating the amount of effort needed to learn to recognize the cases (which I might add, is effort that you have to expend REGARDLESS). I think all of the cases could be memorized in just a few days (wheras learning algs will likely take months), which would afford you with a number of advantages.

I'll be doing it myself over the next few days to see for sure.

 

[Anthony]

You suggested to use this as a stepping stone, but imo it's not worth the effort to learn all the combinations before tackling full OLL.

But why not? I think you are overestimating the amount of effort needed to learn to recognize the cases (which I might add, is effort that you have to expend REGARDLESS). I think all of the cases could be memorized in just a few days (wheras learning algs will likely take months), which would afford you with a number of advantages.

I'll be doing it myself over the next few days to see for sure.

I'm not "overestimating the amount of effort needed to learn to recognize cases", if anything I'm underestimating the effort it takes to learn all the OLL algorithms. Most algs are fairly easy to learn, honestly.

 

[JL58]

It still looks like a daunting task to learn the 45+ algs you need on top of the 10 or so you need for the 2LOLL. This being said I am opting for a hybrid solution between 1LOLL and Zarxax proposal.

To save moves I learned all the "dot" OLL's. Only 8 of them to learn. Then a few more selected for easiness of execution and I am almost halfway (24/57). The rest of it I will probably do what Zarxax suggests:

- Learn recognition and remember the 2 required OLLs. This can be done rather quickly.
- Take all my time to learn the optimized algs.

Take the the example of the small lightning bolts. Two of them are easy - Sunes with a wide move at the beginning and another one at the end. And they save many turns compare to a 2LOLL solution. The two other lightning bolts are more difficult to discriminate/recognize from the first, they require a different AUF and their 1LOLL algs don't save much. I prefer, for the time being, to learn the sune-like and the continuation for the 2 more complex cases. In this specific case the simplification of the AUF should compensate for the lost of 3-4 turns. If that makes any sense.

 

[Zarxrax]

I think it's not all that difficult to just "learn" an alg, but for me, I spend a lot of time trying out the different ones, and trying to figure out the fingertricks, and then I still can't even decide on which one to learn, so I can spend hours working on a single algorithm before I even start to learn it. At the rate I go at, it would probably take me 6 months to learn all of the OLLs, Maybe 2-3 if I REALLY buckle down and do it.

But if you just learn them one at a time, during that intermediate phase while you are learning, it really slows you down, because every single time you get to the OLL phase, you have to stop and think for a minute: "ok, do I know this one?"

I've already learned to recognize about 1/4 of the OLL cases now, with only about 1 hour of studying them.

 

[jms_gears1]

You suggested to use this as a stepping stone, but imo it's not worth the effort to learn all the combinations before tackling full OLL.

But why not? I think you are overestimating the amount of effort needed to learn to recognize the cases (which I might add, is effort that you have to expend REGARDLESS). I think all of the cases could be memorized in just a few days (wheras learning algs will likely take months), which would afford you with a number of advantages.

I'll be doing it myself over the next few days to see for sure.

I'm not "overestimating the amount of effort needed to learn to recognize cases", if anything I'm underestimating the effort it takes to learn all the OLL algorithms. Most algs are fairly easy to learn, honestly.

Anthony, stop your making your self look stupid,
same with you Zar

let me preface this by saying i dont use fridrich so I cold be wrong but:

Zar, when doing 2LOLL, for the most part you already do this, just you dont actively work on doing it, you begin to recog patterns and know what 2LOLL algs you need which is why you see a lot of people who look like they do 1LOLL, its because they dont need to pause they already know what corner OLL alg they are going to need.

Anthony, your points make no sense. Again you do this already. Your suggesting that you should have a pause between E-OLL and C-OLL. If you learn to recog the patterns as OLL then you know what 2 algs to use, w/o a pause why not try and practice like this....

 

[Zarxrax]

Wut?
I'm pretty sure I don't do this already. I don't usually have a pause during 2LOLL because I can usually recognize the 2nd case immediately as it comes up, but, I have no idea what 2nd case I'm going to get until I see it.

 

[4Chan]

Anthony isn't making himself look stupid.

He's right.

 

[Anthony]

Anthony, stop your making your self look stupid,
same with you Zar

let me preface this by saying i dont use fridrich so I cold be wrong but:

Don't try to tell me I'm the one making myself look stupid when you're the one using improper grammar.

Anthony, your points make no sense. Again you do this already. Your suggesting that you should have a pause between E-OLL and C-OLL. If you learn to recog the patterns as OLL then you know what 2 algs to use, w/o a pause why not try and practice like this....

First of all, when did I "suggest that you should have a pause between E-OLL and C-OLL?"

Anyway, here's a good example of what I'm talking about.
I was cubing with Sam Boyles at US Nationals and he did exactly what Zarxrax is describing. He had memorized what C-OLL he would get for every OLL algorithm if he did E-OLL first. He showed me several examples and said he thought it was much easier than actually learning one algorithm for every case. I then proceeded to show him what algorithms I used to solve those few cases and he was surprised at how quickly he learnt them. I'm paraphrasing now, but he said something along the lines of "It's really easy to track the F2L pairs in order to learn the algorithm. Why didn't I just learn these in the first place?"

I'm going to restate what I said before. If you're going to memorize which C-OLL case you are going to get for every OLL, you might as well just learn the algorithm, it's not much more difficult.

 

[Zarxrax]

You keep saying "it's not much more difficult", but as I have already said, in my case, learning all of the algorithms will take me many MONTHS, whereas learning to recognize the cases will only take a few hours.
It seems to me that what you are saying IS in fact quite a bit more difficult, or at the least, more time consuming.

 

[jms_gears1]

Anthony, stop your making your self look stupid,
same with you Zar

let me preface this by saying i dont use fridrich so I cold be wrong but:

Don't try to tell me I'm the one making myself look stupid when you're the one using improper grammar.

Anthony, your points make no sense. Again you do this already. Your suggesting that you should have a pause between E-OLL and C-OLL. If you learn to recog the patterns as OLL then you know what 2 algs to use, w/o a pause why not try and practice like this....

First of all, when did I "suggest that you should have a pause between E-OLL and C-OLL?"

Anyway, here's a good example of what I'm talking about.
I was cubing with Sam Boyles at US Nationals and he did exactly what Zarxrax is describing. He had memorized what C-OLL he would get for every OLL algorithm if he did E-OLL first. He showed me several examples and said he thought it was much easier than actually learning one algorithm for every case. I then proceeded to show him what algorithms I used to solve those few cases and he was surprised at how quickly he learnt them. I'm paraphrasing now, but he said something along the lines of "It's really easy to track the F2L pairs in order to learn the algorithm. Why didn't I just learn these in the first place?"

I'm going to restate what I said before. If you're going to memorize which C-OLL case you are going to get for every OLL, you might as well just learn the algorithm, it's not much more difficult.

I am horrible with grammar, i'm not going to lie. I really only know, for the most part, where to put commas and periods, oh and correct capitalization.
Either way you understood it right? so meh.

Sure its going to be faster to learn OLL by just learning them the plain ol' fashion way, however what zar said works just as well, if not better for recog.

This is how I am going about learning CLL, and was suggested to me by BigGreen. So far i have had no problems with recognition, and in fact when i go to learn CLL algs, its much easier because i just have to put an algorithm with the case.

as he said in his earlier post, learning the algorithm isn't as big a problem once you have that down as learning to recognize them.

 

[Anthony]

You keep saying "it's not much more difficult", but as I have already said, in my case, learning all of the algorithms will take me many MONTHS, whereas learning to recognize the cases will only take a few hours.
It seems to me that what you are saying IS in fact quite a bit more difficult, or at the least, more time consuming.

If you feel that way then by all means, learn this system. All I said is that I feel like it really isn't much more difficult to simply learn the algorithm. Most OLL algorithms have triggers or "play with" a F2L pair that make them easy for me to learn them. If that isn't the case for you, then I suppose you may be better off with the other "system." You've already made up your mind that's what you're going to do anyway. So go for it.

Sure its going to be faster to learn OLL by just learning them the plain ol' fashion way, however what zar said works just as well, if not better for recog.

With practice I suppose it can be just as good, but in no way is it "better for recog."

This is how I am going about learning CLL, and was suggested to me by BigGreen. So far i have had no problems with recognition, and in fact when i go to learn CLL algs, its much easier because i just have to put an algorithm with the case.

as he said in his earlier post, learning the algorithm isn't as big a problem once you have that down as learning to recognize them.

Wait, you mean you learnt CLL by memorizing what PLL you would get for each case? lol. You can do it your way, but in my opinion I think that is a bit silly. I think it is better to simply learn the algorithm. That's what I've done with OLL, PLL, CLL, EG, etc.

Anyway, you guys can obviously do it anyway you want. That's just my two cents.

 

[jms_gears1]

Sure its going to be faster to learn OLL by just learning them the plain ol' fashion way, however what zar said works just as well, if not better for recog.

With practice I suppose it can be just as good, but in no way is it "better for recog."

I think its better, because you practice recog without worrying about having to try and learn the alg at this point.

This is how I am going about learning CLL, and was suggested to me by BigGreen. So far i have had no problems with recognition, and in fact when i go to learn CLL algs, its much easier because i just have to put an algorithm with the case.

as he said in his earlier post, learning the algorithm isn't as big a problem once you have that down as learning to recognize them.

Wait, you mean you learnt CLL by memorizing what PLL you would get for each case? lol. You can do it your way, but in my opinion I think that is a bit silly. I think it is better to simply learn the algorithm. That's what I've done with OLL, PLL, CLL, EG, etc.

Anyway, you guys can obviously do it anyway you want. That's just my two cents.

and no i didnt learn the PLL, i just remembered if i was going to have to do a diagonal or adjacent corner swap.

its worked really well so far.

 

[endless_akatsuki]

Uh. I don't see the point in this thread. when you're using 2 look OLL, you should be automatically memorizing the 2 algorithms to a case anyway, just from sheer repetition. It's a "stepping stone" in itself. And it's NOT faster that full OLL (assuming your recognition is good).