Developing a better LL system

[Kirjava]

Ok. Since I'm currently focusing on getting faster at 3x3x3 at the moment I don't intend to learn this system myself until some point in the future.

However, I would still like to keep up development of it - which I will cease to do for now if I'm the only person planning to learn this.

I don't think this can truly be 'completed' unless it's done in conjunction with learning it - because of this, I'd like to find someone intending to learn this to work with me at developing it. When I did OLLCP, I basically had to go through each subset and produce some information that is currently unpublished (and only exists in my head) to ease learning by a huge factor. Stuff like this can't be discovered without basically learning the entire thing. Also it'd give me huge motivation to know that it was actually being used :U.

Basically, the base material is there for someone to learn this if they put their own effort in. I'm willing to put the effort in if someone is willing to learn it. srs offers only plz.

 

[Ickathu]

How many algs are we talking about? (too lazy to check)
I don't do 3x3 that often, but I'll learn the algs (or at least try to as long as it's not an obnoxious amount) and practice it. How would you learn the algs? Just pick a set and learn all of them, then continue? How is it going to be broken down? By orientation, permutation, or both?

 

[Kirjava]

How many algs are we talking about? (too lazy to check)
I don't do 3x3 that often, but I'll learn the algs (or at least try to as long as it's not an obnoxious amount) and practice it. How would you learn the algs? Just pick a set and learn all of them, then continue? How is it going to be broken down? By orientation, permutation, or both?

You are completely the opposite of what I need. You should probably understand what I'm asking before you offer to take part.

 

[Godmil]

How many algs are we talking about? (too lazy to check)
I don't do 3x3 that often, but I'll learn the algs (or at least try to as long as it's not an obnoxious amount) and practice it. How would you learn the algs? Just pick a set and learn all of them, then continue? How is it going to be broken down? By orientation, permutation, or both?

nonono, there are only a handful of algs. What is needed is learning 1000+ cases of when to use which two (and I pressume which AUF to do in between them).

 

[Kirjava]

You don't need to learn which AUF and second alg unless you're planning to learn it as a 1 look system - you only need to learn the first alg.

 

[jeff081692]

I can put 2 hours a day into learning this. Do you really think it is better to learn by alg instead of take like a ZBLL case and learn the 40 or something algs then practice the recognition for it then move on to another case? The second option is how I always imagined I would learn it.

Edit: you just answered that. I was planning on learning it as a 1 look but I understand what you mean now.

 

[JonnyWhoopes]

You don't need to learn which AUF and second alg unless you're planning to learn it as a 1 look system - you only need to learn the first alg.

This just made me understand how it's even possible to learn this system. I was thinking this was a 1 look system. Now this makes me wish I had the time to dedicate to this...

 

[Kirjava]

I can put 2 hours a day into learning this. Do you really think it is better to learn by alg instead of take like a ZBLL case and learn the 40 or something algs then practice the recognition for it then move on to another case? The second option is how I always imagined I would learn it.

If you'd like to learn it like that then no more work needs to be done - the full table has the data in this format ready to be learnt.

However! I just had a batshit brainwave. A combination of the two would be ****ing delicious.

So you will have the LL split up into CLL cases, and WITHIN that subset, the cases are listed by alg. This could make learning it actually achievable - as it's practisable.

I think this might just be the format that works for the idea. Why didn't I realise this before doing the entire thing ;-;.

So I guess I'll do an LL subset and release it, see how people get on.

God damn this is so much work :U

 

[Ickathu]

You are completely the opposite of what I need. You should probably understand what I'm asking before you offer to take part.

I've been following this thread since the beginning, I'm just a bit confused how it's broken down. I understand what it is: 1 alg to give an easy 1 alg finish. I've been doing this some recently when I get a CLL skip and edges aren't permuted. I'll use a TuRBo alg to do EO and set up into an easy EPLL, or even if corners aren't solved I'll solve into an A-perm, and stuff like that. But you're probably right; I'd probably be a bad person to learn this first haha. Maybe once it's a bit more refined and some other people have done it I'll try and throw in my input

 

[jeff081692]

So you will have the LL split up into CLL cases, and WITHIN that subset, the cases are listed by alg. This could make learning it actually achievable - as it's practisable.

Yea that sounds really cool. Since I was learning ZB like cases first I was using this page
http://www.speedsolving.com/wiki/in...?mode=view&view=default&puzzle=3&group=ZBLL-H
since the cases are organized and I was making a text document with corresponding numbers and the alg from your table.
I didn't get very far though yet.

 

[Kirjava]

Yeah, I'm just gonna go through the damn thing and do it one by one, so I can reduce cases myself.

I'll do the 'A' CLL set first (the one with corners oriented). Then maybe someone can use WV to practise it to check how feasable learning the subsets is.

Once it's shown to actually work, I'll document the other CLLs.

This could be organised in other ways, too - OLL, ELL. However, reducing the number of cases from the 'true' number relies on the way the cases are organised. I think CLL is the best route to take to minimise number of cases (OLL would give almost no reductions).

 

[A Leman]

Sorry Thom, this seems a bit too much for me at the moment. About 4000 cases would be overwhelming for me as I am now. I looked into it; This would be like learning 8000 digits of pi for me and I have trouble seeing how much faster it would be than the more standard large systems like OLLCP or ZB

 

[MaeLSTRoM]

I am considering learning this. Seriously.
Also, Like I said before I don't mind helping with the organising of the cases. Might learn it a bit slowly to start with though, just because 3x3 is much of a focus for me.
Also, this could be pretty boss for K4, just like reduce to L2E in 1 look

 

[Zarxrax]

This sounds like a lot of work and time to invest, but I haven't seen anything yet to indicate that a logical and efficient recognition system can be built from this. Assuming that the entire purpose of all of this work is to discover that recognition system, isn't it also possible that no good recognition method exists, and this is all just a bunch of hard work from which nothing will ever materialize?
If it turns out that people may have to learn to recognize hundreds or thousands of cases to make this method viable, then will it even still have any theoretical benefit to learning this?

 

[Kirjava]

Since when was recognition a problem? You recog the 1LLL case like any other - make up your own system, it's not hard.

 

[Zarxrax]

So you basically just have to memorize like 1000+ cases?
My understanding from the beginning was that as it could be a 2-look system, there was going to be some way to massively cut down the amount of cases needed to be recognized during the first look?
How again is this supposed to be any better than other massive last layer systems that had fewer cases but never caught on?

 

[Kirjava]

So you basically just have to memorize like 1000+ cases?
My understanding from the beginning was that as it could be a 2-look system, there was going to be some way to massively cut down the amount of cases needed to be recognized during the first look?
How again is this supposed to be any better than other massive last layer systems that had fewer cases but never caught on?

None of this is anything to do with the system you use to recognise cases.

You seem to have trouble understanding this system.

It's ok though, I don't need your approval

 

[Zarxrax]

I'm just trying to understand exactly how this works and how it will be beneficial. I think I'm probably not the only one who doesn't really get it.

 

[uberCuber]

So you basically just have to memorize like 1000+ cases?

Just memorizing cases is much easier than memorizing a new alg for every case. 1000+ seems like a ridiculous number, but when every alg is easy and already familiar, (I'm assuming, I didn't look at every alg in the Misc section of the site), I don't see learning this as is as inherently difficult at all. Why? If I'm already really familiar with all algs involved, learning these cases is something I could easily do without a cube, which would tremendously speed up the process. If I wasn't busy memorizing parityOLL and stuff for 4x4, I'd start learning some of this myself right now.

 

[Kirjava]

I think I'd hold off on learning just now until I've started the next revision, am hoping it'll be quite friendly to learn this time.