Possibilities of Compound OLL

[tenderchkn]

After wasting time executing.

 

[pjk]

Umm, yeah. And I'm saying that compound OLL doesn't speed things up.

It does, if don't know all the OLLs.

 

[tenderchkn]

Originally posted by PJK@May 24 2007, 03:12 PM

Umm, yeah. And I'm saying that compound OLL doesn't speed things up.

It does, if don't know all the OLLs.

But as compared to what?

Preserving CO while flipping edges is slower and does not offer any additional values that are not already present in doing it any other way.

 

[dChan]

As I keep saying, it teaches you to evaluate the whole LL situation. As compared to edges first then look at corner orientation. Comp OLL makes you handle OLL as if it were ONE case.

 

[pjk]

But as compared to what?

Preserving CO while flipping edges is slower and does not offer any additional values that are not already present in doing it any other way.

It is quicker over time as if you just treated each new situation as a new case. If you have done the algs enough to know the effect, then you can treat a multiple look OLL stage as 1 look.

I am not understanding where preserving CO comes from... I must be missing something.

 

[dChan]

The H-OLL and inverse of the H-OLL are used to flip edges and preserve CO.

 

[tenderchkn]

Alright, so everyone seems to be a little confused. Let me summarize the case and the arguments for both sides.

"Traditional" 2-look OLL
-Faster EO
-Longer recognition for CO
-Slighty faster overall?

Compound OLL
-Slower EO
-No recognition for CO
-Slightly slower overall?
-Helps one "learn" to recognize OLL?

Items without question marks are pretty much agreed upon; questioned-marked items are speculations. The summary of my argument is that 1) there is no gain in speed from using compound, and quite possibly a lost in speed, and 2) compound does not really teach you how to recognize the entire OLL, nor does it help you transition into a 1-look. I, for one, do not see the OLL as a combination of edge case + corner case, but as an entire shape. And "over time" you're going to learn to recognize the following corner case whether you consciously learn it or not. It's just a matter of repetition. So therefore, my conclusion is that compound OLL does not offer anything over the traditional approach.

 

[Erik]

I'm a bit baffled about the fact that you guys keep saying that it doesn't learn anything and that it doesn't let you treat OLL as ONE. With both systems (let's call it the 'advanced 2 look') and the compound OLL you treat it as ONE case. Treating something as one case means: you recognize the whole thing, you execute it as one thing, you merely use 2 connected algorithms which merge in to one long. I still have for instance an OLL case I perform as: FURU'R'F' R'U2RUR'UR. I also use the reverse of this case. Although I know about all other OLL cases.

 

[pjk]

That is exactly what I would say, Erik.

 

[dChan]

But that is still after a lot of practice. And secondly, you may treat it as one case but say you have 2 edges flipped and 0 corners oriented. You would flip the 2 edges and end up have 3 corners oriented. Or two corners, etc. This means you are treating it as two DIFFERENT cases in one look. And you honestly wouldn't look at the whole LL in a 2-look would you? You'd just do the edges then sort of 'predict' what the next case is. And that is still with lots of practice. But when you are supposed to be learning OLL you don't need a lot of practice on a 2-look LL, right? If you are serious about learning OLL you'll probably never get to the point where you are actually predicting the next case of CO.

With Comp OLL you have 56 algorithms for each case made up of two algorithms which makes it slightly longer. But it is better than the 2-look algorithm where you need to predict the next case. In reality you will never have a case where you look at edge orientation and predict what case you'll get afterwards for CO. Isn't that right? I see no arguement the can possibly formulate against that because it is a FACT. OLL is supposed to be a one-look thing with no such thing as predictions in it. You look at a case apply the rule/algorithm and you are finished. None of this predicting goes around. And Comp OLL isn't slower than 2-look OLL. When you look at it from an objective poitn-of-view it saves you one-look and you know exactly what to do unlike in 2-look where you are predicting what to do. But what if you predict wrong? Then you have to look at the case again and figure out the next algorithm. And it requires a lot more practice that should be focused on learning OLL. I think this really can't be argued.

 

[tenderchkn]

Originally posted by Erik@May 25 2007, 03:19 AM
I'm a bit baffled about the fact that you guys keep saying that it doesn't learn anything and that it doesn't let you treat OLL as ONE. With both systems (let's call it the 'advanced 2 look') and the compound OLL you treat it as ONE case. Treating something as one case means: you recognize the whole thing, you execute it as one thing, you merely use 2 connected algorithms which merge in to one long. I still have for instance an OLL case I perform as: FURU'R'F' R'U2RUR'UR. I also use the reverse of this case. Although I know about all other OLL cases.

I'm not saying anything different. I use the same algs. But it's not like you do FRUR'U'F' to flip the edges then the sune to orient corners - granted, it's a compound algorithm, but it's still one algorithm.

When you do a 2-look, it's a 2-look. It really doesn't matter how you go about doing it. When you get fluent to the point where you can see the OLL and close your eyes and do it in two algorithms, then it becomes a 1-look. It seems to be the goal of both of these methods to eventually achieve a 1-look, correct? Now, the argument that dChan is making is that using the compound method allows you to reach this 1-look stage faster and teaches you something that the traditional approach does not. My counter to that is that no matter what algs you use, it takes the same amount of time to learn to recognize the OLL as a whole and learn to combine two algorithms and execute them as one. It does not matter what algorithms you choose to use.

Even if you do remember and preserve the corner orientation, the resulting case is probably not going to be in the angle that you like, so you'd have to put a U turn in there on the fly. Is anybody (a beginner, especially) really expected to be able to combine algorithms like that on the fly? It takes learning either way, and there's no empirical evidence that one method has a easier learning curve over the other.

 

[Erik]

Ah finaly, you totaly understand it :lol:
However on the last part: the compound OLL is easier to master, cause you already know how the corner orientations will be, in the advanced 2 look OLL thing you only know that after practise... :lol:

 

[shadowclad002]

tenderchkn, have you tried using it (ComOLL)?
I use compound OLL.
IF you use the finger trick versions by Macky (as suggested in the link on the first post) , the algs are pretty fast, for me about as fast as the conventional 2 look OLL.

My counter to that is that no matter what algs you use, it takes the same amount of time to learn to recognize the OLL as a whole and learn to combine two algorithms and execute them as one. It does not matter what algorithms you choose to use.

I believe it does not take the same amount of time... I actually sat down and tried to figure out what would happen to the corners in the normal 2 look OLL before i learned ComOLL. IT was terribly hard(for me, that is).
Now which is a better stepping stone to full OLL? A 2 look that takes more time to learn *..after practice*.., or a two look which gives you almost instantly a one look OLL feel and at the same time teaches you an extra 3 algorithms?
I think it should be this way: First learn 2look OLL, then Compound OLL, then Full OLL.

 

[mazei]

I have around 10 more OLL algs to learn so I do them like this. I'm starting to get too comfortable with it and I might not even learn them

 

[d4m4s74]

I combined the 3 beginner OLLs I know to get a very-long-alged 2look oll
as soon as I know all 2look OLL algs I will probably combine those to a long 1look oll

I bet after that I will combine the OLL and PLL algs I know for a 1lll