2-Side PLL Recog +Hyperorientation= 1LLL Method

[jskyler91]

Hey Guys,

So I don't have great luck here with posting new ideas. I want to make it clear that I am not proposing this as new, here to argue or anything else, but I have never heard of anything like what I am about to describe so here you go. With that in mind i am just throwing an idea out there that I had a while back and for which I am currently working on. Its a 1 look last layer method that I am finding to be incredibly easy and very fast with some practice. The basic idea is that it utilizes 2 side PLL recognition to develop a numbering system (1-6) where each number corresponds to one of the six colors you see when doing 2 side PLL recog.

So for instance, 1 would be the FUL sticker, 2 the FU sticker, 3 the FUR sticker, 4 the RUL sticker, 5 the RU sticker and 6 the RUB sticker.

Then, during your OLL you learn where each of those stickers lie on each OLL (this is only 57 different schemes to learn and many are very obvious and easy). Here are a few examples:

What you would do is first recognize your OLL and then depending on the OLL you would look at the coresponding sticker in order of 1-6 according to that OLL's ordering and then, since you know 2-side PLL you would determine the exact state of your cube and execute as one long alg. This will of course take some serious practice, but learning 57 different schemes is very doable since many are easy, obvious and/or follow a pattern.

The nice thing is there is no need to memo any new algs or anything, you just need to learn 2 side PLL which is really useful by itself and then just learn each of your OLL's pattern for 1-6 and you can do 1LLL. I will be working on creating a chart of all of my OLL's and there schemes soon and will post it soon (see below for link). I have more than 57 algs for my OLL since I do it from different algles so I will be posting all of them that aren't just mirrors of others, but still even a 100 of these is really not bad at all. Its just learning the hyperorientation of 6 stickers. If you just learn to make patterns out of them it becomes really easy. I only have a few down, but its pretty darn cool.

The only real issue I see with this aside from learning 2-side PLL which only takes a few weeks is learning to visualize the cube and you might need to tilt the cube to find your specific sticker, but that can also be done while executing the OLL so ideally there would be no real let downs and you could recall while executing once you have the colors memoed.

The really nice thing here is that with only 6 stickers you can know your full LL from OLL and recog could start as early as last pair if you wanted it too lol, in fact you could make a similar recog system for last pair and OLL/PLL! So to sum up with only two more stickers than COLL you know your whole last layer including AUF, pretty nice IMO.

Another Few Exps-

This one is really easy to trace because its almost shaped perfectly i.e. the first three pieces are on the F Layer exactly where they should be

Edit: Done!!! took ages to do but here is the link for my algs and the position of the 12 sticker prior to OLL. You guys can choose your own 6 consecutive stickers to recog from; they can either be 1-5,4-9, 6-12,or 10-3. I made it so that the UFL sticker will be 1 and UF 2 etc. going counterclockwise. Let me know which ones you all choose, I will list mine here as I choose them.

https://docs.google.com/file/d/0B8dK8hmQEQ68YTRydFFhMVZ1SVE/edit

 

[Kirjava]

So for instance, 1 would be the FSL sticker, 2 the FSM sticker, 3 the FSR sticker, 4 the RSL sticker, 5 the RSM sticker and 6 the RSR sticker where the first letter refers to the face, the middle letter to the layer it is in (S for side) on and the final to the location of sticker as either on the left or the right sticker of that row of three stickers.

Why not use existing notation?

Then, during your OLL you learn where each of those stickers lie on each OLL (this is only 57 different schemes to learn and many are very obvious and easy).

Are many not awkward? When the six stickers are in random places, this could be very difficult.

What you would do is first recognize your OLL and then depending on the OLL you would look at the coresponding sticker in order of 1-6 according to that OLL's ordering and then, since you know 2-side PLL you would determine the exact state of your cube and execute as one long alg. This will of course take some serious practice, but learning 57 different schemes is very doable since many are easy, obvious and/or follow a pattern.

Might be useful for some select easy cases that you are demonstrating, but there will be huge variance with difficulty level as other cases will be horrific.

The nice thing is there is no need to memo any new algs or anything, you just need to learn 2 side PLL which is really useful by itself and then just learn each of your OLL's pattern for 1-6 and you can do 1LLL. I will be working on creating a chart of all of my OLL's and there schemes soon and will post it soon. I have more than 57 algs for my OLL since I do it from different algles so I will be posting all of them that aren't just mirrors of others, but still even a 100 of these is really not bad at all. Its just learning the hyperorientation of 6 stickers. If you just learn to make patterns out of them it becomes really easy. I only have a few down, but its pretty darn cool.

You need to make a translation matrix to allow this to be usable for actual 1LLL systems. Maybe you can make this work for my system since people keep *****ing about not having a recog system. (I don't personally think corner then edge recog is hard - this could end up being a viable alternative though.)

Cool stuff.

 

[applemobile]

Realistically with a 'compound alg' 1 look last layer like this, how much time would you be expecting to save?

 

[Kirjava]

You're trying to save time on PLL recognition by making it harder and adding it to OLL recognition.

Now that I think of it this may only be useful for pure 1LLL systems, which would require further abstraction to be usable.

 

[Litz]

That notation is a bit confusing, so you should probably change it to normal notation if you want more people to read it. The idea seems interesting, but tracking and recognizing 6 stickers isn't necessarily easy, depending on the case and the used algorithm. It would be nice to see the result though, if you intend to explore it. Also, as applemobile already asked, how much time do you really expect to save, since you're just moving recognition from PLL to OLL?

 

[NSKuber]

I'm kinda fast 2x2 solver(1-look), and it takes about 2-5 seconds to create case image after I've found what side I'll do. On the 3x3, I'll need to realise what stickers I'm looking for(depending on OLL), THEN at least 1-3 secs on creating image(6 pieces here, not 4 as on 2x2), and not to forget about OLL during this. Mindblowing. So I'm afraid you need LOTS, LOTS, LOTS OF PRACTISE to be at least as efficient as standart OLL/PLL. But I can be wrong, we'll see if someone try to use this system seriously.

 

[Dacuba]

Am I wrong, or is it as hard as recognizing a 1LLL case?
I once thought about making a table with my eprsonal OLLs and how they influence the permuation of the pieces. But until now I couldn't convince myself it would be worth the effort. But I think the recogniton for corners only should be very fast.

Maybe you can make this work for my system since people keep *****ing about not having a recog system

Just wanted to say how it makes me sad, that people fu'k about anything of your work. Like they don't realize how much effort you put in it. Just like bumping your thread with "when will you be finished?" "Hurry up" etc...

 

[aznanimedude]

so wait, by 1LLL you mean more of a, you take a look at your OLL and then know after the OLL what PLL you get, therefore you solve it in 1 look but 2 algorithms, rather than the other way a la ZZLL/ZBLL where it's 1 look -> 1 algorithm 1LLL, correct?

 

[A Leman]

Hello, I have a few questions if you don't mind.
1. What made you look into 1lll recognition? I'm just curious since ideas like this normally spur off of an interesting train of thought.
2. Have you tried recognizing 1lll with CLL and ELL recognition methods?
3. If so, could you tell me some advantages your system would have over that? I find that recognition rather easy even though I don’t have a use for it(and probably never will).
4. This one is based on my experimenting with ZBLL cases.
Are you possibly trying to learn to Algs based on a partial recognition system to increase you chance of skips, but not always get a skip when you use a special case?
5.Also,what are you going to use this recognition for?

Cool idea by the way.

 

[Escher]

Hm, would it be easier to recognise if you always solved the permutation of at least 2 corners?

As in, you recognise the corner case, AUF to solve (opp, adj and diag are the possible cases), and rotate to your desired angle? Step 2/3 can easily be merged, and step 1 should be very easy to train until it becomes natural. This is just speculation but I reckon 1LLL after that would be pretty easy to recognise...

 

[Kirjava]

Worth noting that you could use any two adjacent sides for the six stickers, minimizing awkward cases because you have four groups to choose from.

 

[mDiPalma]

what if you preoriented edges before finishing f2l (to reduce the number of cases), and then developed some sort of shorter (like 12-13 move) replacement algorithm for each compound-LL case?

 

[jskyler91]

so wait, by 1LLL you mean more of a, you take a look at your OLL and then know after the OLL what PLL you get, therefore you solve it in 1 look but 2 algorithms, rather than the other way a la ZZLL/ZBLL where it's 1 look -> 1 algorithm 1LLL, correct?

This is it exactly! I will add that exp to the OP lol

Realistically with a 'compound alg' 1 look last layer like this, how much time would you be expecting to save?

Not too much maybe a half second to a second, the thing I like most about this is that with only looking at 6 additional stickers you know OLL and your exact location of PLL plus AUF. All during OLL. Thats only two more stickers than for COLL lol and you can solve your whole last layer in one look, plus every bit of time helps right? And this won't be hard to learn, I already memoed like 6 schemes in a few days and I didn't have a good method for finding them at first, now I have a really easy one.

That notation is a bit confusing, so you should probably change it to normal notation if you want more people to read it. The idea seems interesting, but tracking and recognizing 6 stickers isn't necessarily easy, depending on the case and the used algorithm. It would be nice to see the result though, if you intend to explore it. Also, as applemobile already asked, how much time do you really expect to save, since you're just moving recognition from PLL to OLL?

I forgot standard notation lol, we should definitely use that!!!

Hello, I have a few questions if you don't mind.
1. What made you look into 1lll recognition? I'm just curious since ideas like this normally spur off of an interesting train of thought.
2. Have you tried recognizing 1lll with CLL and ELL recognition methods?
3. If so, could you tell me some advantages your system would have over that? I find that recognition rather easy even though I don’t have a use for it(and probably never will).
4. This one is based on my experimenting with ZBLL cases.
Are you possibly trying to learn to Algs based on a partial recognition system to increase you chance of skips, but not always get a skip when you use a special case?
5.Also,what are you going to use this recognition for?

Cool idea by the way.

Thanks, going in order:

1. Well I have been getting very frustrated with having to pause twice for recognition of OLL and then PLL and I figured hey why not try to make it faster? SO i learned 2- Side PLL and then I just realized that all I need to know to do a 1LLL is the 6 stickers and ya it came from that.
2.Ya, the problem is they require way too many stickers to look at a time. THis may involve a new way of looking at things, but quickly ordering 6 stickers, while not necesarilly easy can be learned easily and just takes some practice to get down. I already have the two exps down to a 2 second recog and I only have been using them for a few days. Thats 1 more second and I can execute the whole last layer without stopping at full speed. Pretty nice.
3. Again I think they involve more stickers right? The point of this is that you could learn to very quickly recog OLL and then scan appropriate pieces to know full LL in only 6 stickers.
4.I will probably use this system to avoid ENV and Y perms when they are coming by learning 57 new algs that solve those cases when they come.

Hm, would it be easier to recognise if you always solved the permutation of at least 2 corners?

As in, you recognise the corner case, AUF to solve (opp, adj and diag are the possible cases), and rotate to your desired angle? Step 2/3 can easily be merged, and step 1 should be very easy to train until it becomes natural. This is just speculation but I reckon 1LLL after that would be pretty easy to recognise...

That would make it even easier ya, but I figure with some practice recog will come and that would take extra time if I understand you right. THis method can be used with no additional algs, just a new PLL recog system. I mean how long can it take to scan 6 stickers? Then once you have the order down you do OLL and try to envision PLL while doing it, this gives you an additional second to recog and you likely will never need to stop.

Kir- Worth noting that you could use any two adjacent sides for the six stickers, minimizing awkward cases because you have four groups to choose from.

Good point I had not thought of that thanks!!!

what if you preoriented edges before finishing f2l (to reduce the number of cases), and then developed some sort of shorter (like 12-13 move) replacement algorithm for each compound-LL case?

Hmm, so many good ideas here lol. I already use edge control and try to disorient corners so my OLL cases are more friendly (I averaged all of the 0,1,2,3 and 4 corners up algs for time and found that 0 and 1 corners were the fastest overall, for me at least and with my algs of course), but not a bad idea at all lol!!!Maybe you could use this recog system for ZBLL as well, not sure how, but maybe. I will have to check that out, I don't really like trying to always have edges oriented though, takes too much effort and slows you down IMO, but I will look into it, I think maybe doing something with corners might be very feasible as well.

 

[jskyler91]

Video is up, its in the OP

 

[Kirjava]

Documentation of the method would be better than a 17 min video explaining what we already know

 

[jskyler91]

Documentation of the method would be better than a 17 min video explaining what we already know

Im working on choosing my 1-6's as I type, thanks again for pointing out that there were 4 to choose from that makes this really nice!!

To make things easier I will be using both LF as well as FR 2-side recog. This will make for much more eye friendly patterns on some of harder cases. I could use all 4 sides, but I feel like that would make it harder to visualize.

I'll post my results when they are done. I definitely think this is a very doable and easy way to have a 1LLL.

 

[jskyler91]

Does anyone know how and or where I might go to find all of the OLL's with just the OLL blackened and a way to view all of the sides at once? Imagine something with a 5x5 grid of squares but with the corner blocks missing and preferably computer generated format so its easy to fill in the blanks? Many thanks in advance, I am not a big computer person so this would help me immensely. I tried visualcube, but I can't figure out how to use it lol

 

[MWilson]

If you mean like

then copy "http://www.speedsolving.com/wiki/extensions/algdb/vcube/visualcube.php?fmt=png&bg=t&size=200&pzl=3&sch=nrgyob&view=plan&case=" into your browser URL bar and add the OLL alg to the end of it, like

"http://www.speedsolving.com/wiki/extensions/algdb/vcube/visualcube.php?fmt=png&bg=t&size=200&pzl=3&sch=nrgyob&view=plan&case=FRUR'U'F'"

for

 

[qqwref]

Yeah, I'd wanna see standard notation too (e.g. FR for an edge sticker, URF for a corner). Also it wasn't immediately clear from the topic title or introduction but this is a 1LLL recognition method, not a solving method like I expected.

I think the idea of looking for specific stickers based on the case could be a good one - maybe choosing the right group of stickers could make it very easy to recognize certain cases. This could probably also be used for C*LL too (only two pairs of stickers, and for each you need to check if they're the same, adjacent, or opposite) but I have a feeling it won't be as pleasant as existing recognition systems.

 

[jskyler91]

Yeah, I'd wanna see standard notation too (e.g. FR for an edge sticker, URF for a corner). Also it wasn't immediately clear from the topic title or introduction but this is a 1LLL recognition method, not a solving method like I expected.

I think the idea of looking for specific stickers based on the case could be a good one - maybe choosing the right group of stickers could make it very easy to recognize certain cases. This could probably also be used for C*LL too (only two pairs of stickers, and for each you need to check if they're the same, adjacent, or opposite) but I have a feeling it won't be as pleasant as existing recognition systems.

Ya, this is just a way to recog the entire last layer in 1 Look, so very literally a 1 Look Last Layer Method. I didn't know how else to describe it lol, i am very open to names if anyone has got them and ya I completely agree we should use normal notation, I just forgot it lol, if a mod wants to update feel free, if not I will when I post my patterns.

If you mean like

then copy "http://www.speedsolving.com/wiki/extensions/algdb/vcube/visualcube.php?fmt=png&bg=t&size=200&pzl=3&sch=nrgyob&view=plan&case=" into your browser URL bar and add the OLL alg to the end of it, like

"http://www.speedsolving.com/wiki/extensions/algdb/vcube/visualcube.php?fmt=png&bg=t&size=200&pzl=3&sch=nrgyob&view=plan&case=FRUR'U'F'"

for

And thanks that really clears it up for me, my only question is how could I add numbers to them? I want to put 1-12 on each of the non OLL colors so that people can choose their own patterns as well. This way I can just create a list of my algs and the image with where each of the 12 pieces is located before you perform the OLL. I hope that makes sense lol.

And actually is is possible to just show the OLL and leave the other 12 spots blank? That might be easier to work with.