ZZ speedcubing method

[ostracod]

HUZZAH.

I'm working on my own collection of F2LL algorithms. You can find them here:

http://web.mac.com/teisenmann/Tripod/winter.html

The page includes a case identification system which is useful for memorization. Each F2LL case is given a letter, much like PLL cases. I tried my best to make the letters resemble their cases... Remember that I had to use ALL 26 letters and a symbol (&), so there are a few cases that are a stretch (like S and H).

All of the algorithms are optimal. The average move count is 8.07 (face metric). I plan to add notes for each case soon!

 

[Neroflux]

rar, i hate the diagrams.

 

[ostracod]

In Soviet Russia, diagram hate YOU!!

If you know how to read a standard LL picture, you can understand these as well. Just remember that the slot is at FR, and the blue thing is the corner-edge pair which goes in the slot.

 

[Neroflux]

no really, i can read them, just with a little more difficulty. hey no need to be so pissed, im just complaining.

 

[ostracod]

I'm not pissed, it's a Russian Reversal.

http://en.wikipedia.org/wiki/Yakov_Smirnoff

 

[ostracod]

Ok, I added some notes for each algorithm. Some of it is like poetry. XD This one is my favorite:

"N case (9 moves):
U' R' F2 R F2 U L' U L
U c/e is sledge-hammered in with double F turns. This results in the displacement of the LF c/e, which is repositioned with a slight toggle."

Here's the link again:

http://web.mac.com/teisenmann/Tripod/winter.html

These kinds of descriptions help me memorize the algorithms much better. I hope they do the same for others learning this variation!

 

[Matthew]

Example ZZ solve by me. Anyone who's good at ZZ, please help!

Scramble (Yellow is D face and Blue is F face.)
B2 D2 L D B2 L2 R' U R U' B2 R2 D2 R' U2 F2 D R2 L2 U2 F B R' F B2

EOline
B U R D L' D' F' B2

F2L
U' L' R2 U2 L U L R' U2 L R' U R U2 R U' R' U R' U' R U' R' U R

My linear solution:

EOLine:
U' L' B' F' R F U' R2 D (9 moves but we have a CE pair connected)

F2L:
u' L2 u R' U R U L' R2 U R U2 R' U2 R2 U' R2 U' R U' L U' L' U' L U2 L' (27)

I build in another way EOLine so I have a little more difficult f2l cases - your cases are more pleasant

 

[ostracod]

I was surprised to find a lack of ZZ method videos on YouTube, so I made a tutorial... I put a lot of emphasis on EOline (it is the most unique feature, after all!). I hope people find this remotely interesting/helpful.

Part 1:

http://www.youtube.com/watch?v=mO0IeE40wfc

Part 2:

http://www.youtube.com/watch?v=MjKyZg_PJms

 

[ostracod]

I thought I might as well...

BUMP

This topic. I have a lot of stuff to say still!

As you might have learned by reading this thread, I am trying out the Winter variation of the ZZ method (which involves orienting all corners during the last c/e placement). I have just finished learning ALL 27 algorithms on my page. This has taken me under 2 weeks... I don't think that many other people have memorized full WV, so here are some tips:

-The method is QUITE useless until you've memorized ALL of the algorithms. Unlike solving LL cases, WV cases cannot be done in parts; it's all or nothing. So deciding to learn WV is a bit of a leap of faith, because you won't make progress until you've finished.

-You have to be REALLY good at recognizing and executing mirror cases/algorithms. 3/4 of the time you will get a case which is a mirror of what you've learned! What I'm doing now to practice mirrors is I'm making 27 flash cards, one for each case. Each card has a picture of a case drawn in heavy sharpie, and on the back I've drawn the MIRROR by seeing where the marker leaked through. This way, I can practice all 4 mirrors by flipping the notecard along the X and Y axes.

-Use lots of tricks for memorization! I find it extremely helpful to use my lettering system for both recognition and identification (I can remember what the "A" algorithm is, and I can easily translate a case into "A"). Finding similarities between algorithms helps. Lastly, the most useful thing to do is to follow the F2L pieces during the execution of each algorithm. I intentionally picked most of the algorithms on the page so that once can easily see where c/e pairs are moving.

-PRACTICE PRACTICE PRACTICE. For learning algorithms, I will repeat a brand new one at least 10 times, then I'll repeat somewhat new ones a few times. A few times a day, I will go from A through & and execute each algorithm, just to make sure I remember all of them.

So far, this variation seems to be rewarding. When I do a WV algorithm, I think to myself, "Wow, I totally just placed that c/e pair and oriented all the corners! That was fast." XP

Again, this method saves about 3 moves during solving. I don't know what the move count is for COLL, but the Winter variation has a legitimate reason for having a lower move count: LL algorithms tend to work by breaking up the F2L; however, with WV, the F2L is ALREADY broken up, so you save those few breakup moves!

I would only recommend this method to people who are comfortable with mirror cases/algs (LOTS of them). If you can do mirrors, then you have to learn 27 WV algs + 13 PLL algs (which many people already know). You will theoretically obtain faster solve times WITHOUT learning bajillions of algorithms.

Now here's my question for ZZ solvers:

Should I learn all the algorithms for the R U R' case ("Easy join up")? Blah, another user on this forum, has called this the Summer variation. I have already generated all 27 of the algorithms, and unfortunately, the average move count is more than 9. This is higher than for WV, which is about 8. The average move count leads me to believe that it's not worth the effort to learn the Summer variation... Is this a correct assumption, though??

And also, the origin of this variation is STILL unknown!! PLEASE, if somebody knows, tell me! Thanks!

 

[Werner]

I would say sebastian felix is the "creator" of this method
But i would not bet my life on it

http://pagesperso-orange.fr/absolutemind/f2ll-angl.htm

Also there is this page

http://ten10.myweb.hinet.net/f2ll.html

Ostracod
Ever thought about learning
COLL 40
WV 27 +
EJF2L 17

whenever you got a pair you would use WV
If you got a ugly pair you would use EJF2l or COLL

When i get fast i think i will start using this ZZ method for pure OH cubing.
I dont think it would be that hard to learn those 87 algorithms (I already know full fridrich )

 

[ostracod]

Werner,

I'm pretty sure I've heard of that combination of algs before, but I don't think it's worth learning WV with COLL and EJF2L... Perhaps EJF2L would be worth it, but whenever I have a corner+edge next to each other, I separate them and either join them or make an R U R' case (which would lead to WV or... SV). COLL would just be a hassle, because it would involve learning a second LL method in addition to OLL + PLL. But if you had the will power to learn all of that, then you'd be pretty fast. ;D Then again, 1 look LL might be better at that point...

I saw the first page before (perhaps I'll email the owner to find out more info)... But the second page I have not. The little animation things look helpful.

 

[Werner]

Btw have anyone done any studies to the PLL and OLL skip probability using
WV and EJF2l, with and without ZZ.
I would also like to see some studies about skips done with ZB

My guess is that in ZZ only half of the standar F2l cases are possible
edges can be flipped correctly or incorrectly.
question is how many of these ends with a R U' R insertion normally ?
My guess would be that the OLL skip probability is very high 50%+ using
ZZ + WV only on the normal cases. With setup moves it is 100% but if this is efficent or not can be debated.
LL skips in both ZZ and ZB would be interesting to know.
If you theoretically got 100% oll skips you then would get a LL skip every 72 solve ( which imo is extremely good, think about competitions)

Just me releasing my thoughts. If you learn COLL algorithms that does affect the edges you would be able to do a 1look ll pretty quickly. COLL + H Z or U
4sec LL anyone ? And a skip every 12th solve seems very tempting...

 

[Lucas Garron]

Btw have anyone done any studies to the PLL and OLL skip probability using
WV and EJF2l, with and without ZZ.

PLL skip: 1/72
OLL skip: Nonsense

 

[ostracod]

According to my calculations, even with setup moves (for R U R' case, doing R U2 R' instead), WV is still more efficient by a fraction of a move. So you WOULD get a 100% OLL skip, if you use WV 100% of the time (since it is more efficient than direct placement). If you really wanted, you could learn the Summer variation as well for R U R' cases, but as I said before, the benefits are slim.

 

[ostracod]

We're closer to finding the origin of the Winter variation... I sent Sebastien Felix an email asking him about where he found the variation. He sent me this reply back:

Hi, i first see the Winter Variation(what I called F2LL because it was easy to remember and we immediatly see that it happens at the end of F2L and that it helps to simplify the LL) on the international yahoo speedcubing group. In the file section you could find a "winter variation" document, originally designed for petrus after step 3 is completed and shared by "Cubacca1972".
I feel good to see you enjoying F2LL, it can leads to very short LL's. But like multislotting, its much more usefull for OH.

I hope that i answered you correctly, have fun

sebastien

Perhaps after some searching I may be able to contact this person?

 

[Zeroknight]

Awesome. So move count for EBWP (EOLine/Cross, Blockbuilding, Winter Variant, and PLL) is how much again?

 

[ostracod]

It should be 3 less than ZZ-VH w/ PLL. I'll do a solve to see what I can get...

FULL FACE TURN METRIC (ZZ-WV) with arbitrary scramble:

EOLine: 6 moves (2 bad edges, place both line pieces)
Right block: 10 moves (average 2x2x1 build, then c/e pair already made)
Part of left block: 7 moves (not good..)
Formation of final c/e: 5 moves (again, not lucky)
Winter variation: 9 moves (W phase, not an easy one)
PLL: 12 moves (G-perm, not the easiest PLL)

So I got... 49. This was a somewhat unlucky solve. Here's another one.

EOLine: 8 moves (TONS of bad edges. However, I spotted a premade c/e pair, which I preserved.)
Half of right block: 3 (using the salvaged c/e pair...)
Left block: 11 (while making the other block, a c/e pair was made, which I used)
Final c/e pair formation: 9 (the pieces were together, so I had to shimmy and shake)
Winter Variation: 6 (Z phase, a nice short one )
PLL: 13 (F perm... long)

Sum: 50

...I'm think I'm just incompetent at low move count solves. XD However, I'm pretty sure that WV is cutting down my average move count by a small amount.

However, I think a problem might be that I'm wasting moves by trying to make a c/e pair on U. For the second scramble, perhaps MGLS would have been more efficient, since the corner and edge were in an "EJ" pair. However, if the c/e can easily be made on U, WV is a fast technique.

 

[Musturd]

Does anyone have algorithms for solving the last F2L pair and permuting the edges at the same time? (or is this not a good idea for some reason?)
Then all you would have to do after EOLine + F2L is COLL* and you're done.

 

[ostracod]

Ah, this has been done before. It's similar to ZZ-d, except at least somewhat possible, since there are only 2 dozen permutations. Here's the thread:

http://www.speedsolving.com/forum/showthread.php?t=8871

It's not a very good idea, because recognition is a pain (but perhaps it is possible to do quickly if you work hard?) However, unlike WV, there are only a handful of algorithms to learn.

PS:
http://www.youtube.com/watch?v=UEnLpHP7_4A
- Another resource for ZZ beginners trying to do EO.

 

[Musturd]

That thread is for permuting the corners, though, right?
Has no one tried to do edges?
IMO edges should be easier to recognize.