2x2 L5C Method

[WoowyBaby]

This is the thread for the 2x2 method called L5C, or Last Five Corners. This is also called 1LLS, for One Look Last Slot.

The beginning of this discussion may have started here. There's probably somewhere else this was first thought of, and if so, tell me.

L5C is a method where you solve a permuted V, or 3/4 of a layer, first. Then you solve the rest of the cube (L5C) in one algorithm.
This method requires 486 total algs, subtracting CLL and TCLL gives a figure of 358 algs.
Current algsheet is found here.

Advantages:
-Making a Solved V requires significantly fewer moves than a layer or a face. Using a computer optimal solver, I have found that layers require 4.0 moves on average to solve optimally, and faces 3.0. I have not done solving for a Solved V, but it is around 2.2.
So, this means this method has a lower movecount and is easier to one-look.
-Algs can be really simple like R U R' or R U2 R' or R U R2 F R F'



I would like to add a naming convention to the algs. Each set has ~40 algs:
CLL TCLL+ TCLL-
L5C 1= L5C 1+ L5C 1-
L5C 2= L5C 2+ L5C 2-
L5C 3= L5C3+ L5C 3-

The number, 1, 2 or 3, tells you the orientation of the piece that belongs in the slot.
1 is oriented, 2 twisted clockwise, 3 twisted counterclockwise.
The symbol, =, +, or -, tells you the orientation of the piece that is IN the slot.
Again = is oriented, + twisted clockwise, - twisted counterclockwise.

Edit: Also there is subsets, like inside CLL there’s Sunes, Pi, etc, cases.
For L5C = sets, The names will be the same as CLL.
For L5C + and - sets, the names will be the same as TCLL


This method isn't really necessary but is cool.

 

[Gappo]

Well,actually I want to publish it later this week lol
But it's fine XD

 

[Gappo]

Some explanations about the algsheet:

1. Algs are not fully genned,and still not optimized
2. Full L5C algs will be published later this month (or even later this year...)
3. The algs showed above is version 1
version 2 will be the optimized version with some extra algs , and it'll be published around Q2 2019
4. I think we could publish a algsheet that everyone can edit (Not that one above lol)

 

[Skewbed]

Why not used conjugated CLL like the 42 method. It solves L5C with just CLL algorithms assumming there is an oriented corner.

 

[Sion]

Tbh, 2x2 is too easy for a method of this absurdity. Solving one layer of the 2x2 in a speedcubing setting isn't that difficult and is quite efficient regardless.

I wouldn't completely ignore this, though, since it might help make Corner first 3x3x3 methods more efficient.

 

[Duncan Bannon]

Tbh, 2x2 is too easy for a method of this absurdity. Solving one layer of the 2x2 in a speedcubing setting isn't that difficult and is quite efficient regardless.

I wouldn't completely ignore this, though, since it might help make Corner first 3x3x3 methods more efficient.

Kinda, but I disagree. Solving one layer is quite easy yes, but solving 1 less piece is always going to be better (and even if it isn't, you can just solve the layer). Will Callan is learning a subset of this I believe (LS, which is 3/4 of a face with the last corner being oriented). And he may continue to learn this, he already know probably around 300 2x2 algs, and to continue to get better than everyone else, continuing to learn algorithms is always going to need to happen.

Higher TPS and lower movecount is the only way to improve at 2x2. Will has most likely maxed out his TPS, but movecount can almost always be improved upon.

The 1-2 moves this alg set might save has a tremendous effect. At 8 TPS the difference between a 13 move and 12 move solution is .125 and the difference between a 13 move solution and a 11 move solution is .25 or 1/4 of a second (easily a WR difference). Given you won't always use this alg set.

At a comp you being able to have a 9 move solution because there is a pre solved (or close) V and nobody else can do that solution. Chris Olsen has a video about the creation of TCLL and he said that in that set he was able to do crazy cancellations with TCLL that nobody else saw. This is similar (in fact TCLL is a subset of L5C).

People said ZBLL was the craziest thing, but this has similar alg count and the movecount is much less, making it very much viable to learn (Jabari anyone?).

I apologise if this next paragraph is rude, I don't mean it that way. You said that 2x2 is too easy for this crazy of an alg set, but then said not to stop the generation because it could be used for corners first. Again no offense, but isn't that kinda crazy for a method that very very very few people use and is quite clearly not the fastest method out there???

 

[shadowslice e]

Kinda, but I disagree. Solving one layer is quite easy yes, but solving 1 less piece is always going to be better (and even if it isn't, you can just solve the layer). Will Callan is learning a subset of this I believe (LS, which is 3/4 of a face with the last corner being oriented). And he may continue to learn this, he already know probably around 300 2x2 algs, and to continue to get better than everyone else, continuing to learn algorithms is always going to need to happen.

Higher TPS and lower movecount is the only way to improve at 2x2. Will has most likely maxed out his TPS, but movecount can almost always be improved upon.

The 1-2 moves this alg set might save has a tremendous effect. At 8 TPS the difference between a 13 move and 12 move solution is .125 and the difference between a 13 move solution and a 11 move solution is .25 or 1/4 of a second (easily a WR difference). Given you won't always use this alg set.

At a comp you being able to have a 9 move solution because there is a pre solved (or close) V and nobody else can do that solution. Chris Olsen has a video about the creation of TCLL and he said that in that set he was able to do crazy cancellations with TCLL that nobody else saw. This is similar (in fact TCLL is a subset of L5C).

People said ZBLL was the craziest thing, but this has similar alg count and the movecount is much less, making it very much viable to learn (Jabari anyone?).

I apologise if this next paragraph is rude, I don't mean it that way. You said that 2x2 is too easy for this crazy of an alg set, but then said not to stop the generation because it could be used for corners first. Again no offense, but isn't that kinda crazy for a method that very very very few people use and is quite clearly not the fastest method out there???

Well it's not the best method out there. For instance EG is better and if you want to do this you could even reduce the alg count by using something like 22.

 

[Duncan Bannon]

Well it's not the best method out there. For instance EG is better

Yes, EG is currently (sorry HD) the best method out there in terms of speed. I'm 99.9% positive this method was "created" to be an addition to EG not a replacement. TCLL and LEG-1 were made to be used with EG. Chris Olsen even stated in the same TCLL video I mentioned earlier that TCLL was made to be used with EG and not a method by its self. And TCLL is a subset of L5C.

and if you want to do this you could even reduce the alg count by using something like 22.

I'm not very familiar with 22 and I know you very much are, but in doing so (limiting the algs) your also increasing the movecount by a margin that probably makes it equivalant to just using EG. And that is my point, when it comes down to 1-2 moves being the differnece between WR and not, 2x2'ers are willing to learn the algs. Sorry for the spelling errors too, my computor isn't working very well, so moving my cursor is hard.

Edit- Example solves coming soon showing my CLL and Eg-1 skills in movecount and time vs L5C

 

[shadowslice e]

I'm not very familiar with 22 and I know you very much are, but in doing so (limiting the algs) your also increasing the movecount by a margin that probably makes it equivalant to just using EG. And that is my point, when it comes down to 1-2 moves being the differnece between WR and not, 2x2'ers are willing to learn the algs.

Well if you use 22 with the same number of alg as L5C (say by incorporating EG-like techniques), then you end up with a method that's more efficient. I will grant however that L5C may still be more bang for it's buck though 22 is marginally better imo (and eventually/hopefully that may give it the edge long term).

 

[PapaSmurf]

Tbh, I think that L5C is the next step after teg-2 and leg etc. It will save those 2 or 3 moves that will bring down the times to allow insane averages. It is a lot of algs, but recog isn't a problem due to one looking, and 2x2 algs aren't that hard to learn.

 

[Duncan Bannon]

Well if you use 22 with the same number of alg as L5C (say by incorporating EG-like techniques), then you end up with a method that's more efficient. I will grant however that L5C may still be more bang for it's buck though 22 is marginally better imo (and eventually/hopefully that may give it the edge long term).

Wait. How would it be less moves? Please explain.

As promised- A L5C comparison VS EG (generating L5C just on the spot)

Spoiler: L5C vs EG

Scrambles from CS timer just used the given scramble and these are the results

Scramble - F' R2 F' R2 F R2 F' U' R'

EG-

x' y2
F R' F R - Layer
(U) R' U R' F R F' R U2 R' U R - CLL (cyotheking for these algs)
U

17 moves

Vs

L5C-

x' y
R - V
R' F2 U' L' U2 R F R' U F2 - L5C (not caring for finding good alg, and am picking the second lowest movecount solution because I'm just using cube explorer and setting up the case I see thus there is not pre AUF, so I'm being fair and adding that extra move (cause it is 3/4 that has pre AUF)

11 moves or 9 moves if you count the cancel. (using the 11 moves for the mean at the end)


Hopefully this is clear

Next scramble - U' R F R2 F2 R2 U F U

EG

y2
F' R U' R' - Layer
(U2) R' F2 R F' R' F2 R U' R' F R F' - CLL

17 moves

VS

L5C

y2
F' - V
R' U' L' U2 F' U L2 U' F U2 - L5C (again picking one move above optimal)

11 Moves


Next scramble - R2 F2 U' F R2 U' F2 U R'

EG

z
F R' U' R' - Face
U R' F R2 U R' F' R U2 R' - EG-1
U' - AUF

15 moves

VS

L5C

y'
F' - V
U2 F R U2 R2 U2 F L F L2 F' - L5C

12 moves

Not a perfect world because the algs from cyotheking aren't the most move efficient, but the numbers are still here.

Mean of EG- 16.3
Mean of L5C - 11.33

Lucky, perhaps, but these are the first 3 CS timer scrambles I got. At 8 Tps that is 2.04 vs 1.42

Thanks for your time reading this all.

 

[shadowslice e]

Wait. How would it be less moves? Please explain.

https://www.speedsolving.com/wiki/index.php/42
Have a look at the "As a 2x2x2 method" section and maybe some the advanced techniques.

Like I said above, it's a lot more complicated so maybe not really worth it compared to other things on 2x2 at least.

 

[LukasCubes]

How about a metwhod where the V isnt completely solved.... hmmm ~3690 algs?

 

[Melkor]

I don't know where the 486 comes from, L5C is 614 algs. Am I missing something?

 

[LukasCubes]

I don't know where the 486 comes from, L5C is 614 algs. Am I missing something?

idk where 486 comes from either.

 

[LukasCubes]

https://www.speedsolving.com/wiki/index.php/42
Have a look at the "As a 2x2x2 method" section and maybe some the advanced techniques.

Like I said above, it's a lot more complicated so maybe not really worth it compared to other things on 2x2 at least.

42 as a 2x2 method just seems like the 1LLS method tbh. Also what happened to ur yt channel? You aint upload in like 40 years lol.

 

[Melkor]

42 as a 2x2 method just seems like the 1LLS method tbh. Also what happened to ur yt channel? You aint upload in like 40 years lol.

Lol, yeah, we will forever only know how to recog the T cases of 42.