[Help Thread] ZZ and ZB Discussion

[Petro Leum]

MdiPalma, do you use ZZ-rainbow as your main method now? how is the cp recognition?

 

[aznanimedude]

seems he's gotten used to it?

 

[mDiPalma]

oh yeah, for twohanded, i switched to zz-d (porkynator variation) but i don't think he'd mind if we called it "ZZ Rainbow"

i average around 20 seconds with it, but that's cuz im restricted to building the left block before the right block in order to have reasonably fast cp recognition. i'm eventually gonna switch to L/R neutral cuz it really feels bad to waste an uber-easy xeoline on the right simply because i'm used to solving the left block first.

what i learned was (and i know its uber obvious) that L' U R U' L switches the positions of UBR and URF, and so you just want to have all corners in a relatively solved permutation state or relatively "swapped" permutation state. that helped me learn to see what alg to use and when. i could do a quick video tomorrow on recognition if you really need it. but im no expert by any means.

the cp recognition is moderate. I'd say it takes like 3 seconds total to identify the cp and apply the algo. plus you have to solve the two DR corners first. eventually, i hope to be able to expand my cp recognition repertoire to have the DR corners diagonal from one another. and then move on from that. obviously, i believe that cp recognition should eventually be determinable simply by lookahead during the left block.

but yeah, i'm gonna keep up with zz-d. it has a lot of potential from what i can see. I will also be learning 2gll as I go. i'll keep you posted. feel free to join me! this should be really fun!

 

[Petro Leum]

oh yeah, for twohanded, i switched to zz-d (porkynator variation) but i don't think he'd mind if we called it "ZZ Rainbow"

i average around 20 seconds with it, but that's cuz im restricted to building the left block before the right block in order to have reasonably fast cp recognition. i'm eventually gonna switch to L/R neutral cuz it really feels bad to waste an uber-easy xeoline on the right simply because i'm used to solving the left block first.

what i learned was (and i know its uber obvious) that L' U R U' L switches the positions of UBR and URF, and so you just want to have all corners in a relatively solved permutation state or relatively "swapped" permutation state. that helped me learn to see what alg to use and when. i could do a quick video tomorrow on recognition if you really need it. but im no expert by any means.

the cp recognition is moderate. I'd say it takes like 3 seconds total to identify the cp and apply the algo. plus you have to solve the two DR corners first. eventually, i hope to be able to expand my cp recognition repertoire to have the DR corners diagonal from one another. and then move on from that. obviously, i believe that cp recognition should eventually be determinable simply by lookahead during the left block.

but yeah, i'm gonna keep up with zz-d. it has a lot of potential from what i can see. I will also be learning 2gll as I go. i'll keep you posted. feel free to join me! this should be really fun!

nah, i dont need it ;D
that was my thought so. sometimes i just have a glance at the stickers and see immediately what the cp is (most of the time when ist already solved), but it doesnt work always - maybe its connected with co - which stickers are really "visible".

I guess with enough practice and experience one will be able to see the "corner-phase" just during insertion of first blocks last pair. then you dont have to isolate corners anymore, but would need some more algorithms to exploit this efficiently.

So isolation of the corners is not really reccomendable, since its fairly hard to do (with sorting out the other face's pieces), wastes moves, hinders your xeoline usage and doesnt allow you to at least TRY to recognise cp without isolation.
It would also be beneficial to get somewhat colorneutral for second block/ll as you can choose whether to do U +R or R+U.... if you know what i mean ;D
I am also looking into a blockbuilding idea which could reduce 2GLL algs required by quite a bit... it doesnt seem practical though, since i often end up with 2 flipped corners -.-'


Oh, and i might also switch/ work out a slightly different procedure, but only once a sub15 race i am participating in has finished.

 

[porkynator]

Nice to see you're becoming fast with this method! Do you isolate the corners after EOLine or do you build the left block first?
As to me, I think I'll use normal ZZ for now, and see if it's worth switching when I finish learning 2GLL (I know T and U at the moment, but I can't speedsolve half of the U cases).

 

[Petro Leum]

ive primarily tested it with one handed, and i prefer isolating a face directly after EOLine, as it gives you a purely 2gen first block.

However, ive also experimented with replacign the Line wit hthe Isolation, means: (you start an x2 away from your normal solving position)
1. EO
2.isolate DLF and DLB one the L face together with only pieces of U and L
3. Blockbuild a right 2x2x3 using R,r,U,M without smashing the Edge Orientation (so after a r/r' or M/M' only a U2, then slice back)
4.CP
5. z Rotation and either do the original Left Block or the 1x2x2 that would now be on the R Face
6. Enjoy the LL Skip you will probably get.

its kinda not ZZ-Like anymore and the "blockbuilding" is very restricted and most times leads to "bring DF and DB opposite positions on U > M2/r2 to place line >continue with normal right block
anyway, you can have really bad cases at the isolation and i think it would be worty to implement some of it in preinspection... i personally find it harder than cp recognition and EOLine alone.

Oh, and i just write down the cases for cp - meaning that to recognise fast, i Auf to bring UBL in its correct spot, then i look at UFL and UFR and identify Face colours. i just made a handwritten chart, i can post if you want.
I find this way of recognition faster then detecting the Corner Cycle - its just like CLL recognition, only that you have to look for your B colour instead of pairs of stickers.

 

[mDiPalma]

i think it's actually more move and time efficient to do EOLINE + ANY BLOCK + CP + OTHER BLOCK + 2GLL

this is because you can run xeoline or pair preservation to give yourself a fast first two steps.

this weekend i'm going to learn the cp cases with the two D layer corners placed at UFL and ULB (and obviously the reflections). knowing this orientation as well as porkynator's original suggestion means that the "corner placement" step will take only 1.4 moves on average and only 2 moves maximum. (<<those numbers might be completely wrong.)

these 1.4 moves can easily be look-aheaded during the first block, leaving the solver a cancelled niklas, a <12 (?) move 2gen second block, and a 12 move 2GLL away from solved, which is very nice position to be in after 7-8 seconds of solving.

 

[Petro Leum]

i think it's actually more move and time efficient to do EOLINE + ANY BLOCK + CP + OTHER BLOCK + 2GLL

this is because you can run xeoline or pair preservation to give yourself a fast first two steps.

this weekend i'm going to learn the cp cases with the two D layer corners placed at UFL and ULB (and obviously the reflections). knowing this orientation as well as porkynator's original suggestion means that the "corner placement" step will take only 1.4 moves on average and only 2 moves maximum. (<<those numbers might be completely wrong.)

these 1.4 moves can easily be look-aheaded during the first block, leaving the solver a cancelled niklas, a <12 (?) move 2gen second block, and a 12 move 2GLL away from solved, which is very nice position to be in after 7-8 seconds of solving.

of course, but not really.

it would probably be more fficient, however your first block wouldnt be 2gen... well thats an idea i liked because i love OH but w/e.

now, what exactly should change during cp when you have placed the d layer corners in U? isnt it just the same only with harder recognition?

and wouldnt it be best to not only be L/R neutral but also Second block neutral (meaning being able to play U layer corners in R, CP, solve 1x2x2 on U and 2GLL on R)?

Problem is, after all i dont see the huge advantage with 2GLL. cp is is still taking to much think time and too many moves to compete with COLL/EPLL or OCLL/PLL or even Phasing/ZZLL.
Fact is the "enhanced ergonomics" of a 2GLL are only a matter when solving OH. consider that.

i only see a cance for zz-d in general if we can cut down cp recognition to"nothing" and still be able to save 5-10 moves compared to "ZZ light"

PS: since OP is most probably dead or at least the superlative of inactive, is there any way someone else could "take over the thread"?

 

[Zeotor]

I have some questions.

For ZZ, how many different kinds of realistic-to-use color neutrality are there? I know that two are mentioned in this post. (I'm fairly sure that complete color neutrality with ZZ is not realistic.)

Say someone were to start using ZZ as their primary speedcubing method and had no experience with ZZ. Should that person try to be color neutral in one of those ways? If so, which one? I know that there have been discussions on this forum about whether to be color neutral or not. However, those discussions probably focused on the concept as it applied to methods such as CFOP. With ZZ, it's different (right?).

 

[porkynator]

I am F/B and U/D color neutral (which means I can buil the line on either yellow or white face and hold the cube with either orange or red in front of me); I think this is very easy to do.
"y axis" color neutrality seems harder, but still possible to use efficently. Maybe "x axis" color neutrality is almost the same.
More color neutrality for EOLine seems to difficult to me, but maybe someone will be abel to do it in the future... the real problem is the inspection.

Pseudoblocks are really cool, I think you will be interested: it means you build ANY 3x2x1 block on right and left (so it doesn't have to match the colours of the EOLine, it may be off by a R/R'/R2 (L/L'/L2) move), and then you have a different LL recognition system. I don't use it for speedsolves, but it's a fun method to play with, and I think it has speedsolving potential.

Real color neutrality with CFOP is very easy, compared to ZZ, petrus or Roux, in my opinion.

 

[Athefre]

Pseudoblocks are really cool, I think you will be interested: it means you build ANY 3x2x1 block on right and left (so it doesn't have to match the colours of the EOLine, it may be off by a R/R'/R2 (L/L'/L2) move), and then you have a different LL recognition system. I don't use it for speedsolves, but it's a fun method to play with, and I think it has speedsolving potential.

Definitely. Pseudo blocks work really well in ZZ. In other methods you aren't given enough opportunity to build and, in the case of Roux, misoriented edges cause problems in later steps. But with ZZ, most of the solve gives you chances to take advantage of the frequent free or easy pairs. And LL recognition is no longer a problem with the choice between NMCOLL and NMLL.

 

[Noahaha]

Real color neutrality with CFOP is very easy, compared to ZZ, petrus or Roux, in my opinion.

Not true for Petrus, since you basically just start with whatever is already partially solved. Non-CN Petrus is very difficult IMO.

 

[aznanimedude]

I have some questions.

For ZZ, how many different kinds of realistic-to-use color neutrality are there? I know that two are mentioned in this post. (I'm fairly sure that complete color neutrality with ZZ is not realistic.)

Say someone were to start using ZZ as their primary speedcubing method and had no experience with ZZ. Should that person try to be color neutral in one of those ways? If so, which one? I know that there have been discussions on this forum about whether to be color neutral or not. However, those discussions probably focused on the concept as it applied to methods such as CFOP. With ZZ, it's different (right?).

i'm F/B color neutral as in i'll start it with either blue or green in front, not U/D neutral though, still always white bottom. complete color neutrality at least for me would be difficult because i can't imagine a way to count the number of bad edges and formulate a good eoline in 15 second inspection. Y-axis neutrality i think is feasible though, so that you can opt out in hopes that on the other axis there's an easier EOL.

so i guess to answer your question, if i were to go for one i'd go for y-axis neutrality so that one color (and of course it's opposite) are always in the U/D face and be color neutral in which color would be in the F face.

 

[mDiPalma]

i am sub 17 with: WR WO YR YO
i am sub 15 with: WG WB YG YB
it's good to be able to solve a y rotation out of your standard orientation. this allows you to essentially "reroll" your bad edges without hurting recognition too badly.

 

[Zeotor]

I'm only using ZZ this month to see if I enjoy using it. If I do, it will become my new main method for 3x3. I'll be y-axis color neutral.

How should I practice using y-axis color neutrality right now? Should I inspect both orientations, decide which one is better, then plan the EO+Line? Or should I just switch between orientations for each solve?

And LL recognition is no longer a problem with the choice between NMCOLL and NMLL.

Is NMCLL used first and then something like EPLL is used to finish the last layer?

 

[Athefre]

Is NMCLL used first and then something like EPLL is used to finish the last layer?

Yes. Same cases for EPLL of course.

 

[Zaterlord]

I have some questions.

For ZZ, how many different kinds of realistic-to-use color neutrality are there? I know that two are mentioned in this post. (I'm fairly sure that complete color neutrality with ZZ is not realistic.)

Say someone were to start using ZZ as their primary speedcubing method and had no experience with ZZ. Should that person try to be color neutral in one of those ways? If so, which one? I know that there have been discussions on this forum about whether to be color neutral or not. However, those discussions probably focused on the concept as it applied to methods such as CFOP. With ZZ, it's different (right?).

At first when I tried ZZ I only solved with blue in front and yellow on top, but then I saw a post of somebody saying that he didn't know any ZZ user that wasn't colour neutral. I thought he was talking about full CN and since then I solve with colour neutrality.
But I'm just averaging 35 seconds with ZZ and I use to long inspection so maybe it is a bad choice.
(ZZ is not my main method)

 

[mDiPalma]

NEVER plan out EOLINE during inspection
ONLY plan out smartEO
BLOCKBUILD DURING line for XEOLINE or XXEOLINE

 

[Pyjam]

Your theory.

 

[ThomasJE]

I'm thinking of switching to ZZ, but I have a few questions:

- I've known F2L for 8 months, and my PB's are in my sig (I usually average 22, with most PLL). So, would I find it difficult to switch (with different F2L, EO(+)Line)? Also, is it worth switching?
- After EO, you only need to use R, U and L moves. So, if I did the left block, and did the 1x2x2 block in the right, leaving one F2L slot, would I only get an R U R' insertion?
- If so, would you not need to learn Summer variation, and then learn WV, leaving you with a PLL?
- For you ZZ 4x4 solvers, do you use [WIKI]Z4[/WIKI]?

Thanks in advance.