ZZ speedcubing method

[*LukeMayn*]

Musturd: I discussed that, you would eng up having to do a CLL last.
Which is similar to doing OLL after PLL :/
PS I wish I kept on doing ZZ... I had a 26 sec avg :facepalm:

 

[Athefre]

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

There is a Cubacca1972 here, the member still posts and comes up with neat ideas.

 

[Neroflux]

zz is cool, i did it for a week and got 20.xx avg.

i do it eo cross style though... T.T

 

[ostracod]

To Mustard: Ah, sorry, I didn't see that you were talking about edge permutation. This is similar to ZZ-b, except if you permuted ALL edges (instead of 2 opposite ones), you would have fewer cases. I think your idea could work well.

To Athefre (and others): I did PM Cubacca1972, and he kindly sent me a reply. He turns out to be the creator of WV. I finally found out that the name "Winter variation" comes from his last name, which is "Winter". The term "F2LL" didn't come from him, I think from Sebastian. I'm happy to know these things. ;D

 

[Jhong253]

Still don't understand EOLine or whatever it is...

 

[a small kitten]

It's half a cross on the D layer with all edges oriented.

 

[Jhong253]

I got that part, but still don't understand how to make it and stuff...

 

[ostracod]

Try searching for "ZZ tutorial" on youtube to find some of my videos. I give 2 explanations of EOLine.

 

[ostracod]

Ok, so I've been using full Winter variation for about a week now, and I have a few thoughts on it.

Recognition for WV is NOT slow. Even with all the mirrors and whatnot, I can execute the algorithms without slowing down; it's all a matter of learning not specifically what sides to turn, but how to move corner/edge pairs around for each case (thinking about pieces in a relative position to the slot). In most cases, I save a few moves during the solve. I did an average of 3, and I got 36.13 seconds, which is my best yet (and also got a new PR, 34.38 seconds ). Because WV has "only" 27 algorithms, it's not too much to learn, and the benefits outdo the work involved.

However, there IS one weakness, which is when you have... the corner next to the edge, but misoriented/placed. In this case, you would have to separate the corner + edge, then REjoin them correctly, then use WV...... So here's where I recommend what some other people have said in the forum:

AFTER you learn WV, learn EJF2L (aka: CLS MGLS).

I think this order is important, since there are considerably more algorithms for EJF2L (56, compare with 27 for WV), and EJF2L is only faster than WV probably less than 50% of the time (either when the edge is placed, or when you have an "EJ" pair; WV should faster for the other cases, algorithms ARE 1.5 moves shorter on average...)

So it seems that the Winter variation and MGLS really complement each other, because each is fast where the other is weak. I think I'll start learning EJF2L pretty SUNE (HAHA. Sune. Soon. I've made that joke before. I still find it hillllarrrious). I'll learn maybe 1 algorithm per day, since I'm not on vacation now.

Here's my little flowchart to a super ZZ method. Note that each step in the process doesn't abandon any previously learned knowledge (it's all cumulative). By the end, the solver will know a little over 100 algorithms.
1. Learn basic ZZ-VH, with OLL + PLL (3 algorithms)
2. Learn 1 look OLL (7 algorithms)
3. Learn 1 look PLL (13 algorithms)
4. Learn the Winter Variation (27 algorithms)
5. Learn EJF2L (56 algorithms)

 

[a small kitten]

With a little over 100 algs you might as well do ZZLL. Right?

 

[ostracod]

To a Small Kitten: Ahh... You mean permute 2 opposite edges instead of corner orientation, then do 1 look LL? My guess is that ZZLL is more efficient. Let's see...

It's very fuzzy to see exactly how many moves each technique takes during F2L, since some require setup moves. But I'll try my best.

Roughly 50% of the time, edge and corner are not adjacent, or only corner is in last slot. In these cases, you would probably use WV. ZZ-b setup is relatively easy, but WV requires a few extra moves if there is an R U R' case, which we'll say happens... Perhaps 30% of the time, so we'll add 1 move on average to WV for extra move setup. Once a desirable case has been reached, ZZ-b requires an average of 5 moves to insert the c/e pair, and WV requires 8. But remember that the 1 extra setup move increases this number to 9.
ZZ-b: 5
ZZ-WV: 9

The other rough 50% of the time, you will get an EJ pair, or the edge will be in the last slot. This is where EJF2L is more efficient than WV. Assume that EJF2L has an average of 2 moves for setup (placing final edge if it is not placed, or just turning the top face to place the final corner over the final edge...), then 9.5 moves for algorithms. ZZ-b would need more setup moves... perhaps an average of 4? Then the algorithms still take 5 moves.
ZZ-b: 9
ZZ-EJF2L: 11.5

So (theoretically) on average, ZZ-b will use 3.25 fewer moves than ZZ-WV/EJF2L for the F2L.

However, then we have our last layer.

PLL optimally takes 11.2 moves. I am unable to find a calculated average for ZZLL... However, scrolling down a list of optimal averages, and seeing a couple dozen move counts, it seems that the numbers range from 10 to 15. This leads me to believe that it takes around 12.5 moves to do ZZLL.

So (probably) on average, ZZ-b will use 1.25 moves MORE than ZZ-WV/EJF2L for the LL.

And so, if we combine F2L averages and LL averages, we find that ZZ-b roughhhhly probbbably uses 2 moves FEWER than ZZ-WV/EJF2L.

So which one is better? I assume that ZZ-WV/EJF2L and ZZ-b have a similar number of algorithms (someone correct me if I'm wrong), and it seems that ZZ-b is more move efficient by a couple moves. We must also consider recognition as a factor... recognizing how to permute 2 opposite edges before finishing F2L is easier than orienting the corners, but I get the impression that ZZLL recognition is a lot more tricky than PLL.......

SO, which one is better? I still can't say for sure, but ZZ-b wins by a tiny bit.

If you can find any flaws in my reasoning, please point it out!

 

[*LukeMayn*]

ok, I'm back to a ~25ish avg!!!
Now just have to memo winter variation, so far I know 3 xD

 

[Joseph Gibney]

I really think that ZZ-b is the most promising variant as a major speedcubing method. Personally I think that WV and CLS would be more useful as a way to skip OLL (sorry Lucas ) when not using phasing.
I think for most people, ZZF2L and OLL/PLL is a fine method until at least sub 15, and requires very few algorithms. COLL is a nice way to possibly speed up a bit, as is WV and CLS.
If you manage to learn full ZZ-b, then you actually know a large portion of ZBLL, and could eventually switch to that. At that point, now that you're not phasing, WV and CLS would be a nice way to force into PLL.

Of course ZZ-d would be an awesome method, with perhaps ~47 moves and very fast turn rate, however it seems to be out of our grasp. I have put forth some effort into trying to reduce to <R,U> during F2L, but so far all attempts have been only partial, and required very long recognition. If anyone has any insight into this, let me know! If this could be done, F2L would be a bit harder (as you can't form both sides at the same time as easily), but the LL would be amazing: 2GLL is one look with relatively few and fast algorithms. Combined with corner orientation control (WV, CLS) you could often force EPLL, and even last layer skips would not be very rare. AMAZING!!

EOline is a bit difficult for many to master, but I found it to be alright with a bit of practice. For most I'd imagine the <R,U,L> F2L to be a bit of a challenge, since CFOP users need to use a different mind set. You must solve 2x2x1 blocks, not just pairs. Multislotting might be more feasible with ZZ with fewer possible cases (not sure about this). Also, some people rely on one hand heavily during F2L, for ZZ you must use both hands nearly equally. People who use many cube rotations during F2L might find RUL F2L slower than normal F2L. This style actually fits me well, I hardly rotate at all.

I'm currently a bit under 15 using ZZ with OLL/PLL, but I haven't been practicing it that much (for comparison I'm a bit under 13 with CFOP 3LLL, thus with ZZ the advantage for me is having 1 look OLL). I think EOline is still holding me back, as is the F2L. While the RUL fits me well, I still haven't gotten completely used to it. I plan to at least learn COLL and 2GLL soon, and eventually ZZ-b, although it would be great if zz-d was possible. I already know some WV, and learning WV+CLS would be nice, but not a major goal yet. Working on EOline + F2L will help me the most right now by far.

I kinda still want to finish up CFOP by learning OLL (and the two PLLs I don't know.... ahaha), would probably get me sub 12 with some practice, but I dunno, what do you guys think? I could probably get it done in less than a day if I actually tried...

For anyone trying ZZ, I would suggest focusing on your EOline + F2L, but of course these LL ideas are neat and GOOD LUCK!!

 

[ostracod]

I think it's pointless to learn BOTH ZZ-b and ZZ-WV/EJF2L, unless you want to see how well they work in comparison. When solving the slot, there is no reason to believe that one would be better to use than the other (unless you're low move count solving, in which case you could try each variation...) Plus, each variation has its OWN set of 100 algorithms. You'd be learning 200+. At that point, you'd be better off learning 1 look LL.

I agree, ZZ-d would be very fast, if there were a way to find the "missing link". But really, there ISN'T a good way to find the missing link. By definition, you must find the permutations of 6 corners (and fix them) in order to put the cube into <R,U>. This obviously makes for difficult recognition... But perhaps someone can find an effective way to recognize ZZ-d cases.

And I think your idea of multislotting is a good one... It would definitely be easier, and it would probably make solving considerably faster. However, I'm not sure if you can use any special techniques (WV, EJF2L, etc) if you multislot. I will look into that for sure.

 

[Escher]

Ostracod, multislotting is something that you use when solving fridrich pairs. You basically try and make the next pair as easy as you can while solving one pair. You can end up with f2ls of very few moves. Winter Variation is actually a part of 'multislotting' - the aim of multislotting is to influence the next step, just as WV does.
applying multislotting knowledge to ZZf2l would be fantastic - not only would the f2l be RUL, it would also be as short or shorter than the normal f2l (as i understand it, the zzf2l is slightly longer than the fridrich equivalent?).
Multislotting is something that most people do intuitively, like knowing the different effects on the cube between R U' R' and R U2 R'.
So yes, actually WV and EJf2l fit nicely into multislotting.

 

[Lord Voldemort]

I got sub 22 with Fridrich, and a sub 20 3 of 5, so I'm focusing more on ZZ again. I used it in the weekly comp, btw. Around 35 seconds, but with almost no practice on ZZ in the last 2 weeks.

I dion't like Winter Variation... a lot of it seems sort of obvious, and it's perhaps faster to use R U' R'/R U2 R' insertions to perhaps force a Sune/Anti-Sune as much as possible. I might just learn it, but I don't know.

 

[*LukeMayn*]

Lord:
It's shite easy to learn... I could learn the whole thing in 1 day if I could be bothered sadly I can't
so I'll stick with 1 or 2 a day
EDIT: woah, I just realized that there is a 1/72 chance of LL skip with WV
If anyone gets good with this, I bet the WR single would be easy EOline, nice F2L, LL skip

 

[*LukeMayn*]

Okay, I'm bored...
So I'll make a ZZ progreesion chart!

Aiming for: 1min
EOline - 12 secs
F2L - 28 secs
OLL - 10 secs
2-look PLL - 10 secs
Work on your EOLine and F2L the most But also don't for get OLL, PLL... should be easy to minus 10 secs from here

Aiming for: 50 secs
EOline - 10 secs
F2L - 25 secs
OLL - 7 secs
2 look PLL - 8 secs
Work on F2L the most, also practise EOline. Practise OLL a bit too. I would say mainly recognition and TPS

Aiming for: 45 secs
EOline - 9 secs
F2L - 22 secs
OLL - 6 secs
PLL - 8 secs
As always practise EOline and F2L the most. Learn block-building and try making both side at the same time. Just get TPS higher for OLL, PLL

Aiming for:40 secs
EOline - 8 secs
F2L - 20 secs
OLL - 5 secs secs
2 look PLL - 7 secs
EOline and F2L... Practically do what you did in the last step

Aiming for:35 secs
EOline - 7 secs
F2L - 18 secs
OLL - 4 secs
PLL - 6 secs
please note, full PLL, you have to get round to it some time... And you can learn the whole thing in 3 weeks, 1 a day...

Aiming for:30 secs
EOline - 6 secs
F2L - 16 secs
OLL - 3.5 secs
PLL - 4.5 secs
Now it's time to work hard on F2L and Eoline, the TPS will come naturally from here, make sure you are doing optimal moves .

Aiming for:25secs
EOline - 4.5 secs
F2L - 13 secs
OLL - 3 secs
PLL - 4.5 secs
REALLY work on PLL's especially recognition. OLL's should be nice and quick now! to get to the next step do MASSIVE work on F2L, here are some tips:
Optimal turns
Block building
TPS
Recognition
and whatever YOU think you have to practice.

Aiming for: 22 secs
EOline 4.5 secs
F2L11 secs
OLL3 secs
PLL3.5 secs
You may want to start to learn WV now. like the 0 corners case and the 3 and 2 corner cases. Practise EOline recognition, improve move count.

Aiming for: 20 secs
EOline - 4 secs
F2L minus last slot - 8.5 secs ( yes a big leap, practise hard, it will come)
Winter variation - 4 secs
PLL - 3.5 secs
whew, at the 20 mark finally, look foward so your sub 20 avgs. they're great Try PLL time attacks and aim for sub 1 min. more work on F2L and uh one thing...
ZOMG WV learn the cases, they're super easy to memo and very quick.

Aiming for: 17 secs
EOline - 3.5 secs
F2L minus last slot - 7.5 secs
Winter variation - 3 secs
PLL - 3 secs
You should be getting regular sub 20 avg's. Pratise your WV speed and recog, work on EOline and predicting how the F2L will turn out and how you'll approach it. Maybe start to learn CLS.

Aiming for: 15 secs
EOline - 3 secs
F2L minus last slot - 7 secs
Winter variation+CLS - 2 secs
PLL - 3 secs

And from here it's up to you

Aiming for: 14 secs
EOline - 3 secs
F2L minus last slot - 6.5 secs
Winter variation+CLS - 2 secs
PLL - 2.5 secs

Aiming for: 13 secs
EOline - 2.5 secs
F2L minus last slot - 6 secs
Winter variation - 2 secs
PLL - 2.5 secs

Aiming for: 12 secs
EOline - 2.25 secs
F2L minus last slot - 6 secs
Winter variation - 1.75 secs
PLL - 2 secs

Aiming for: 11 secs
EOline - 2 secs
F2L minus last slot - 5.5secs
Winter variation+CLS - 1.5 secs
PLL - 2 secs

Aiming for: 10 secs
EOline - 1.75 secs
F2L minus last slot - 5 secs
Winter variation+CLS - 1.25 secs
PLL - 2 secs

Okay, so I got a bit unrealistc at the end, but I think it would be about right...

 

[Johannes91]

Aiming for: 10 secs
EOline - 1.75 secs
F2L minus last slot - 5 secs
Winter variation - 1.25 secs
PLL - 2 secs

Okay, so I got a bit unrealistc at the end, but I think it would be about right...

- Optimal EOLine is about 6.13 moves (less with color neutrality), so 1.75s is easy.

- For F2L + OCLL, I just tried an average of fairly fast linear solves.

27, (23), 33, 24, 30, 29, (40*), 32, 25, 31, 32 => 29.22 FTM avg
* whole LL

I haven't practised this much at all, didn't use any fancy tricks, and don't know optimal algs for all WV cases, so about 25 moves should be possible. But even with 30 moves, 6 seconds is clearly doable.

- And PLL, the fast cubers are probably doing it sub-2 already.

So, sub-10 does sound quite realistic to me.

 

[ostracod]

I should theoretically be at 20 seconds, since I know all of the Winter variation, but I get times around 35... I think I need to push my TPS a lot. XP