Hawaiian Kociemba 3x3x3 TUTORIAL - Method of Michael H.

hey guys ive almost completely transitioned over to this method because I can get insanely fast times with the petrus variant.
to get times about 8 seconds lower than my former main I do(im learning algs to make this faster):
1.petrus 2x2x2(plan during inspection)->2x2x3
2.solve final 2 pairs
3.2-look corners(OCLL+T and Y perm)
4.L5E EO(roux-like EO, but with only 5 edges)
5.insert the DF edge with M' U2 M' and EPLL

comparison:
average:31-32 secs
avg. with roux:38 secs

and this long sequence still allows me to get averages lower than my main(Roux)
This method is Really fast, i like it! Im learning OCLL and Im going to learn L5E/EHKPLL algorithms too. that will take a lot off my time.

Thank you Wirsandfirs3! youre awesome!
 
Shiv3r said:
hey guys ive almost completely transitioned over to this method because I can get insanely fast times with the petrus variant.
to get times about 8 seconds lower than my former main I do(im learning algs to make this faster):
1.petrus 2x2x2(plan during inspection)->2x2x3
2.solve final 2 pairs
3.2-look corners(OCLL+T and Y perm)
4.L5E EO(roux-like EO, but with only 5 edges)
5.insert the DF edge with M' U2 M' and EPLL

comparison:
average:31-32 secs
avg. with roux:38 secs

and this long sequence still allows me to get averages lower than my main(Roux)
This method is Really fast, i like it! Im learning OCLL and Im going to learn L5E/EHKPLL algorithms too. that will take a lot off my time.
Click to expand...

Congrats!
How far do you think your times can go down if you stick with the method variant?

I am currently trying to make a speedsolving method consisting of ZZ, Petrus, and little to no CFOP (Averaging around 37 moves).
 
Shiv3r said:
hey guys ive almost completely transitioned over to this method because I can get insanely fast times with the petrus variant.
to get times about 8 seconds lower than my former main I do(im learning algs to make this faster):
1.petrus 2x2x2(plan during inspection)->2x2x3
2.solve final 2 pairs
3.2-look corners(OCLL+T and Y perm)
4.L5E EO(roux-like EO, but with only 5 edges)
5.insert the DF edge with M' U2 M' and EPLL

comparison:
average:31-32 secs
avg. with roux:38 secs

and this long sequence still allows me to get averages lower than my main(Roux)
This method is Really fast, i like it! Im learning OCLL and Im going to learn L5E/EHKPLL algorithms too. that will take a lot off my time.

Thank you Wirsandfirs3! youre awesome!
Click to expand...
Glad this works for you :D
Further proof this is
a. A decent method
b. The Petrus variant is best for Speedsolving.
The EO with 5 edges is actually FreeFOP step 3. I have all the cases, if you want them. They are easy.
 
wir3sandfir3s said:
So most of your are probably familiar with Micheal Humuhumunukunukapua'a's debut video(I didn't copy and paste the name, I memorized it...). In this video, he does a few solves, claiming the method he uses is "Hawaiian Kociemba". The first solve was reconstructed on Reddit, and I have decided to try and decipher his method as no one knows for sure how it works. Keep in mind that this is mostly a theory, and that this could all be totally wrong, but if this is correct, then it is a pretty decent method.
EDIT: This is an amazingly efficient method, and great for both speed solving and FMC. Read some of the later replies.

1. Create EO arrow (only orient F2L pieces, arrow is 3 cross pieces. Missing piece is placed on right.)
2. F2L-1 edge: Use RUL moves, like in ZZ, to do F2L-1 edge. Use temporarily free R slice to your advantage.
3. HKOLL: This is mostly similar to normal OLL, except you also orient RD.
4. HKPLL: 149 cases, 21 PLL, 128 new.
Cube is solved.

1.EO "Arrow": If you are familiar with ZZ, this should be pretty easy. Orient all or just F2L edges, if you orient all you will get an easy corner OLL. The best choice is probably just F2L. This allows you to do F2L with just RUL moves like in ZZ. After edges are oriented, place in 3 D edges, and place the incorrect one at RD (or LD, if you want, but later algs will have to be mirrored).
2. F2L-1 edge: Pretty easy step, uses only RUL moves. The -1 edge is the missing D edge. BE SURE to use the free R slice to your advantage in the beginning, it will help your F2L.
3. HKOLL: Use 1 algorithm. Self explanatory. Similar to normal OLL.
4. HKPLL: Use 1/149 algorithms (21 PLL, 128 HKPLL) to permute everything and solve the cube :).

I am currently having trouble generating HKOLL algorithms, so I'll add them later.
HKPLL: https://www.dropbox.com/s/3dxhlcnskzj5wfc/HKPLL.LUR.txt?dl=0
EHKPLL: https://www.dropbox.com/s/1vhumt0nrz9q8tc/EHKPLL.LUR.txt?dl=0
L5E MU: https://www.dropbox.com/s/7mogt7yvgojnjye/L5E MU Cases.txt?dl=0
L5E RU: https://www.dropbox.com/s/88ktri1azi972v5/L5E RU Cases.txt?dl=0
Thanks to Cale S for generating these L5E algs.
Load this in Cube Explorer for visual (for now), I'll add pictures later. The HK algorithms only use RUL moves, and some of them are pretty long, but I'll fix them later. You may also have to AUF, especially for EHKPLL.
PLL: https://www.speedsolving.com/wiki/i...gDB?mode=view&view=default&puzzle=3&group=PLL
2-gen OLL: https://www.speedsolving.com/wiki/index.php/OLL#OLL_24 (only use the 2-gen ones, r moves are fine. Can also use the COLL Algs where all you have to do is twist the corners.)
L5E Spreadsheet: https://docs.google.com/spreadsheets...it?usp=sharing
Thanks to Alex Maass for making this.

HKPLL can also be solved using commutators. You can also try COLL and then just do a conjugated EPLL, but you would have to orient all edges in the beginning.
Scramble: R2-B2-F-L'-B2-U'-B2-R-D2-R2-F-R2-U-F-D2-B-U
z y' //
D' L' U x // Cross (without the green white edge piece
R' U' R U' R' U' R // Green-Orange Pair
L U2 L' // Blue-Orange Pair
L' U L U' L' U L // Green-Red Pair
R U' R' // Green-Red Pair Insert
r U R' U R U2 r' U // Basically OLL
R' U' R U R U R U' R' U' // Basically PLL

This is the actual reconstruction for the first solve by vikktorz, exactly as he posted it.

Scramble: R2 B L2 B' R' U B R' D' B R' U' R' D B' D U' L' D'
y R' U' R U' R' U R //F2L-1
R U' R' U' R U R' y' //F2L-2
R U2 R2' F R F' R U2 R' //HKOLL
U' R U' R' D R U R' D' R U R' D R U' R' D' //PLL
Second solve reconstruction by JTWong71.

Scramble: U F U D L' U B' U R R L U' R U L' B U' R2 U2 R2
y r x' U2 U' // Stuff?
u' R U' R //Stuff?
F R U R' U' R U R' U' F' // OLL?
U2 R' D' R U R' D R // PLL?
Third solve reconstruction by TheFearlessPro.

Giant List: https://docs.google.com/document/d/1l7a6vJc_eflZn1LDt9GAq4lPGQiD8HEm7oJFe2CmhfA/mobilebasic
Thanks to JTwong71 for making this.

Pretty cool method, the reconstruction is 41 moves... seems efficient, too.
EDIT: Actually, every solve so far is 45 moves or less... This is a really efficient method. Check other solves in the replies, and post your here.
Here are a bunch of variations to decrease the number of algorithms used or to make the method more efficient. I'll add them as I think of them.
This will probably be THE MOST complex variation but can be VERY fast and significantly improve recognition. You will want to plan it in inspection. During the explanation, pretend the E slice is divided in half - left and right.
1. Solve a 2x2x3 block (preferably on the left) while leaving 2 edges disoriented. Be sure to track these edges.
2. Now it gets pretty tricky, and reliant on look ahead. Pre mute the bottom corner on the right, they do not need to be oriented. Pair the front one with one of the disoriented edges.
3. Figure out which adjacent corners needs to be swapped. Put them on the left, while making sure the other disoriented edge is at UR or UL. If corners need to swapped diagonally, do this step but use a conjugate to move the corners adjacent to each other.
4. Do either F' U F (if the edge is at UR) or F' U' F (if the edge is at UL) to reduce the cube to an entirely 2-gen state. Yes, all R U moves the entire way (if you want, I wouldn't recommend it as some 2-gem EHKPLLs are pretty long). Solve the other 2 pairs on the left (cube is in F2L-1 state) and use 1/7 Algs to orient corners whilst keeping permutation, then 1/7 EHKPLLs.
This sounds like a lot, but it can be executed pretty quickly. The reason for doing all this is when you finish F2L-1, you will only have to orient corners (only use 2 gen algorithms) and the corners will be solved. You will then be left with a EHKPLL, which there are only 7 of and are pretty easy to recognize and execute. This makes the last layer only use 14 algorithms, beating ZZ-R, DRASTICALLY improving recognition, and takes the cake for fewest number of algs for 2LLL.
Just basically orient all edges in the beginning and use COLL when you get to the last layer to be left with an easy EHKPLL. That makes the last layer only require 49 algorithms.
Use CLL to finish the corners, then do a D move to place the incomplete D edge in the back and do L5E.
Thanks to Alex Maass for creating this variant.
1. Make a 2x2x3 block.
2. Solve other 2 pairs, not RD.
3. COLL or CLL.
4. L5E, or orient edges as you would in FreeFOP or Roux then do EHKPLL.

Thanks to everyone who helped with solve reconstructions and algs. :D
Click to expand...
Its Fake Solves. The solves are from Rowe
 
Shiv3r said:
hey guys ive almost completely transitioned over to this method because I can get insanely fast times with the petrus variant.
to get times about 8 seconds lower than my former main I do(im learning algs to make this faster):
1.petrus 2x2x2(plan during inspection)->2x2x3
2.solve final 2 pairs
3.2-look corners(OCLL+T and Y perm)
4.L5E EO(roux-like EO, but with only 5 edges)
5.insert the DF edge with M' U2 M' and EPLL

comparison:
average:31-32 secs
avg. with roux:38 secs

and this long sequence still allows me to get averages lower than my main(Roux)
This method is Really fast, i like it! Im learning OCLL and Im going to learn L5E/EHKPLL algorithms too. that will take a lot off my time.

Thank you Wirsandfirs3! youre awesome!
Click to expand...

What you described is almost exactly the T-CELL method, except I think in that method, you leave out a corner out of F2L as well, and do TCLL to solve the remaining 5 corners. My point is, that method isn't exactly new or revolutionary.

In fact, I think I can improve it a bit. So I'm fairly certain that doing L5E the way your doing, and doing LSE Roux style is about the same amount of moves, and IMO, LSE Roux style has better lookahead, so the method could be revised as:

1. Build 2 2x2x1s (At least 1 planned in inspection)
2. Solve final 2 pairs
3. CMLL
4. LSE

At this point, this method is Roux with a different way of building the blocks. And in my opinion, building one block at a time is better anyway, due to no rotations and the moveset quickly being limited to [R,r,M,U].

But if you're better with the method that you proposed, good for you :).
 
IwantCubingTips said:
Its Fake Solves. The solves are from Rowe
Click to expand...

I don't think that's really relevant to this. Sure, the solves were faked, but the point is that it gave people the (not so new) idea of leaving a part out of F2L to solve while solving LL.

Slightly irrelevant thought, I don't think it's good to put all of these very diverse methods under the umbrella term "Hawaiian Kociemba", considering how different some of the variations are.
 
You must log in or register to reply here.

Similar threads

A
NEW Guide to Choosing a Speedcubing Method (Speedcubing.org)
5 6 7
Replies
126
Views
17,677
PetraPine
PetraPine
TheSlykrCubr
ZZ-CT vs Hawaiian Kociemba
Replies
4
Views
916
Username: Username:
Username: Username:
S
Hawaiian Kociemba
Replies
8
Views
1,739
Kaneki Uchiha
Kaneki Uchiha
Username: Username:
Quest to become sub-10 Hawaiian Kociemba (ditching CFOP, HK is better & we have pineapples)
2
Replies
21
Views
4,406
Username: Username:
Username: Username:
Cubingcubecuber
  • Poll Poll
Quest For Sub 10 Hawaiian Kociemba | New Directions
6 7 8
Replies
154
Views
25,672
zed zed
zed zed
Back
Top