The New Method / Substep / Concept Idea Thread

[Shiv3r]

Not a method exactly but it's a concept so whatever

Basically an extension to the sim in cstimer that helps train lookahead more efficiently

So it comes preloaded with a method definition file that tells it what pieces need to be solved in what order and stuff, then when you start solving a pair (it'll detect this somehow i guess) it'll remove those stickers from the sim cube; when you make moves it'll give arrow cues (like in Posner cue paradigm) to direct your attention towards where new information is coming from (so for example if you do an R' an arrow will point to whatever was just on the B face to direct your attention there, since those pieces will be about to be revealed).

So the end result is it forces you not to pay attention to what you're currently solving and it trains you to direct your attention efficiently

Even better but more complex: an augmented reality type app that videos you solving, detects moves, and adds the arrows and hides pieces accordingly.

Hey anyway, I think there is something like that but only for corners.
Also, Teoidus, How far developed is your 2GR method? and can you make a video tutorial on it? 1:08 or so yesterday on 4x4(with Lewis method ofc), and I got 2 sub-1 times.
And also, I am starting to average ~

 

[Teoidus]

Hey anyway, I think there is something like that but only for corners.

Can you link me to it? I know there's the coracle lookahead drill but that's very different. It trains you on inferring a 3rd unknown corner sticker when only 2 stickers are visible (which is useful for lookahead but hardly what i'm suggesting here).

Also, Teoidus, How far developed is your 2GR method? and can you make a video tutorial on it?

It's still being developed. Progress is slow because I'm busy with other things but I'll release everything when I'm satisfied with all the tutorials and stuff.

1:08 or so yesterday on 4x4(with Lewis method ofc), and I got 2 sub-1 times.
And also, I am starting to average ~

What is the average movecount and how does it compare to Yau?

(Also what is Yau average movecount for each substep--cross+centers, 3-2-3, 3x3 stage?)

 

[Shiv3r]

Can you link me to it? I know there's the coracle lookahead drill but that's very different. It trains you on inferring a 3rd unknown corner sticker when only 2 stickers are visible (which is useful for lookahead but hardly what i'm suggesting here).


It's still being developed. Progress is slow because I'm busy with other things but I'll release everything when I'm satisfied with all the tutorials and stuff.




What is the average movecount and how does it compare to Yau?

(Also what is Yau average movecount for each substep--cross+centers, 3-2-3, 3x3 stage?)

Average movecount is slightly higher than Yau but is comparable to Meyer or Hoya. The lookahead trainer is exactly what I meant, i didnt look at teh link sorry.

 

[Teoidus]

Do you know where I might be able to find Yau statistics?

 

[Shiv3r]

Do you know where I might be able to find Yau statistics?

from what I have been told, Meyer's average movecount is 120-140, or "slightly Higher than Yau."

Movecount-wise, from my experience, Lewis method seems to be more move-efficient and lookahead friendly than Stadler. And just from doing solves with each method (I made sure I used roughly the same TPS), My method also seems to be more move-efficient than either K4 or Kenneth's big cubes method, which along with Stadler are the biggest 3 direct solving methods.

 

[Shiv3r]

Also, I just made this crappy movecount example solve (less efficient than I can do in speedsolves, It was super unlucky) with the method, just to show the basic Idea of the method. some comms I did super inefficiently.

 

[asacuber]

 

[Shiv3r]

Also, guys, I finished the FP pyraminx method algs, in full!(havent like tried speed-optimizing them, most are still just the optimal algs). Here they are.
for those of you who don't know what the FP method is, it is a method that was proposed over 2 years ago but there were never any algorithms generated, or if they were there are ZERO on the internet. I don't know why, because the algorithm count isn't too bad. There are 21 cases total, but 2 are ELL's, at least 4 of them are inverses of 4 other algs (I have still not figured out all the algs that are mirrors/inverses), and 5 of them are standard LBL algs(FP includes LBL as well.) so in reality, you only need to learn 21 - 5 - 4 - 2 = at most 10 new algorithms For a 2-look pyraminx method.
---------------------------
here are the steps:
-solve a face (like a layer but the edges don't have to be permuted correctly)
-in 1 algorithm, solve the rest.
--------------------------------------
a lot of the algorithms have not been made speed-optimal as they were genned using Jaap's digital pyraminx puzzle/solver. i am working on speed-optimizing them slowly.
Pros:
-2 looks for the whole method
-with minimal inspection, you can predict which of the FP sets you will get(you can narrow it down to at most 6 cases by just looking where the D edges get placed)
-VERY low algorithm count; if you know LBL and Keyhole(even though the 2 ELL algs are pretty easy), then you only have to learn 10 new algs.
-in inspection use less time trying to build a face vs. a layer, more time can be spent to plan further in the solve.
-Theoretically 1looking is easier
Cons:
-the algorithms currently are not very fast(this will change as they get improved)
-stell a few more algs than 1-flip or intuitive L4E
-not proven to be fast yet -- I haven't learned the whole thing yet, and I average like 12 secs.

 

[Sue Doenim]

Quick idea: would it be good to have alternative insertion algs for Oka, to help prevent bad cases?

 

[Shiv3r]

I

Quick idea: would it be good to have alternative insertion algs for Oka, to help prevent bad cases?

I have been thinking of genning algs for Oka+centers, you insert the edges while orienting centers

 

[Sue Doenim]

Where could I find an alg generator?

 

[Shiv3r]

https://www.jaapsch.net/puzzles/javascript/pyraminj.htm

 

[Teoidus]

35 STM 2GR

B D2 F' L F' L' F' R B D' F U2 B2 R U' L' F L' F' R'
// insp
U' r2' u' y' R U r2' // EOPair 6/6
y R2 U F2 // CPLine 3/9
R2 u2 U R U' r2 // Block 6/15
R U2 R2 U' R2 U' z' // F2L 6/21
U2 R U R2 U' R2 U' R2 U2 R2 U2 R' U R2 // 2GLL 14/35

 

[Y2k1]

35 STM 2GR

B D2 F' L F' L' F' R B D' F U2 B2 R U' L' F L' F' R'
// insp
U' r2' u' y' R U r2' // EOPair 6/6
y R2 U F2 // CPLine 3/9
R2 u2 U R U' r2 // Block 6/15
R U2 R2 U' R2 U' z' // F2L 6/21
U2 R U R2 U' R2 U' R2 U2 R2 U2 R' U R2 // 2GLL 14/35

So I've looked at all the stuff you posted for cpll, and I understand now how to find the key swap and read the 2gen corrections table, but I cant figure out how to read your F move charts. How do you track the F moves on the tables? This cp in inspection is really cool (if i can figure it all out)
Thanks

 

[Shiv3r]

Hey guys. There is a 4x4 method like Yau but for ZZ me and Neuro came up with. It is called NS4. You start 3x3 stage with EOXcross every time.
it also allows you to figure out if you have OLL parity even when finishing edgepairing, and by adding a few algs to EPLL you will never have to waste another look to resolve PLL parity.
Here are the steps:
-Solve L & R centers
-Solve the Line edges, place them at DL and DR, so a D/D' move would place them correctly. Also place an oriented cross dedge and an oriented E-slice edge at FL/FR.
-Solve last 4 centers
-Solve last cross dedge (solving another dedge with 2-at-a-time at this point is good), then Resolve the cross. Pair up the E-slice edge with a corner to make an F2L pair and place in one of the back slots oriented.
-Edgepairing+EO. You can either EOpair, or solve 2-at-a-time while orienting the edges afterwards, it's up to your taste. Resolve OLL parity now if you have it.
-ZZF2L. You should have full EOCross, and one F2L pair already solved in the back. there is a 50/50 chance it is in the correct spot already, but if it isnt, now place it in the right spot. You now have EOXcross, and continue with ZZF2L/Rotationless F2L or whatever you want to call it
-Last layer. I HEAVILY suggest using COLL-EPLL, because by adding ~ 4 algs and the 2 pure parities to the EPLL algset, you get EPLL+parity, which solves PLL parity at the same time as well. EPLL algs here.

The original thinking behind this method was this: CFOP has Yau/Hoya, Roux has Meyer, but what does ZZ have? a shittier form of redux+beginner edgepairing, Z4(which no one uses). but after playing with this, I have found that having a yau-like version of ZZ wasn't the only advantage of this method. You also get a really fast 3x3 stage, and with only a few extra EPLL+parity algs you have a purely 2-look last layer(CFOP LL on 4x4 varies and can go as high as 4 with double parity). And movecount isn't a disadvantage as well; with help from Teoidus, we found that the average movecount is about 128, which is only a little worse than Yau(which is 120-140 moves), and more efficient than Meyer(130-150).

Anyone think this is interesting?

 

[Teoidus]

So I've looked at all the stuff you posted for cpll, and I understand now how to find the key swap and read the 2gen corrections table, but I cant figure out how to read your F move charts. How do you track the F moves on the tables? This cp in inspection is really cool (if i can figure it all out)
Thanks

Each table has a cell which represents a 4-cycle of key swaps. To get from one state to another in each cell, you do an F move. (So if a cell contains swaps A B C D, and you have swap B, an F move will get you to C, another will get you to D, and another (or an F') will get you to A)

I have a much easier recognition method now but I haven't had time recently to make it look nice. Hopefully I'll get time eventually to do so

 

[Y2k1]

Each table has a cell which represents a 4-cycle of key swaps. To get from one state to another in each cell, you do an F move. (So if a cell contains swaps A B C D, and you have swap B, an F move will get you to C, another will get you to D, and another (or an F') will get you to A)

I have a much easier recognition method now but I haven't had time recently to make it look nice. Hopefully I'll get time eventually to do so

So on this chart how would say a UFL-DFR key swap be affected by an F? I still don't quite grasp how its layed out yet

 

[Teoidus]

So the ULF-DFR swap is in this set of 4:


To get what would happen to the swap after an F move, just look at the diagram to the right of the ULF-DFR swap (in this case that gives you URF-ULB). (To figure out what would happen after an F', look at the diagram to the "left", which would give you URB-DRB since the chart wraps around)

 

[Y2k1]

So the ULF-DFR swap is in this set of 4:
View attachment 7452

To get what would happen to the swap after an F move, just look at the diagram to the right of the ULF-DFR swap (in this case that gives you URF-ULB). (To figure out what would happen after an F', look at the diagram to the "left", which would give you URB-DRB since the chart wraps around)

Thanks a lot

 

[crafto22]

Okay can someone who understands this whole key swap CP 2GR stuff explain it to me pleaaaase? I'm very interested in this method due to its movecount but I just can't wrap my head around it. Anyone willing to dedicate just a bit of time to make a short in-depth tutorial would be much appreciated by many, I'm sure!

Sent from my Moto G (4) using Tapatalk