• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Is ZZ an objectively worse method than Roux and CFOP?

adimare

Member
Joined
Oct 29, 2009
Messages
381
Location
Costa Rica
WCA
2011MARE02
Uh could you tell me please what do you average with zz, and with cfop
~12 sec with ZZ. It's hard to tell what I average with CFOP because I stopped using it years ago. I tried it a few days ago because I got the urge to stop sucking at 4x4 and was able to get consistent times with what I get with ZZ after a few solves even though I... wait, why am I answering this? What does this have to do with anything of what I said? Did you figure out you didn't have any good responses and decided to go full ad hominem? Is your plan to just go "oh well you're not sub-X yet so your opinion doesn't count"? Because I got bad news for you, I could be a honey badger that somehow learned how to type but is fully incapable of solving a Rubik's cube, you'd still have to disregard that and respond to my arguments.

I'm sorry to disagree, but lookhead is a lot easier in ZZ because the edges are oriented, and solving a 2x3 block required significantly less moves on average than solving two pairs (and you don't have to care about edge orientation and cube rotation). And as usual, you may influence the last layer case while solving the second block.
Solving 5 pieces requires significantly less moves on average than solving 4? Can you prove that?

And I'm sorry, but pieces being oriented doesn't help lookahead that much (it's cool when you're solving the last pair and can't see the edge, but you can safely assume it's in the back and can pair it and insert it without having to figure out its orientation, but that's about it). On the other hand, having edges you're solving stuck in the bottom layer does affect lookahead negatively, so does the fact that after EO line the edge you're looking to solve can be in one of 10 locations vs one of 8 locations in CFOP, or that when you create an initial pair/block to solve a 2x2x1 the remaining edge is sometimes stuck with the pieces you just paired up (a problem that can be solved by either separating them, or planning how you pair up pieces more carefully to avoid it from happening, both of which take more time than the mindless stuff you can do for CFOP pairs and both of which force you to pay more attention to the block you're solving at the moment instead of paying attention to the next thing you're gonna solve). In summary, worse lookahead.
 

Pyjam

Premium Member
Joined
Oct 8, 2010
Messages
2,142
Location
La Baule, France
Lookahead is better because any edge on top that is not yellow is an edge for the F2L block, and it's very easy to see how to build a 2x2 block with what's visible on top, and while doing it it's very easy to spot where are the two remaining pieces to complete the 2x3 block.

The "ZZ Example Solve" Game!" thread contains many examples with very short blocks, and the theoretical superiority of blockbuilding in regards to the movecount has been proven.

That said, the transition from EO-line to left block isn't always easy, especially if the EO-line was hard and the DL edge has landed on the right side.

I would say, if there's an obvious disadvantage about ZZ is that you can barely take advantage of a prebuild block resulting from the scramble. Let's say, there's two good pairs already made, or even a 2x2 block. Any good CFOP user will likely see an XX-Cross during inspection, while a ZZ user will stupidly do an EO-line and break the two pairs. OTOH, if all edges are oriented after the scramble, any CFOP user will take advantage of it without even notice it, solving the F2L without any rotation and skipping EO during OLL, or even get an OLL-skip with (or w/o) Winter Variation.

That's why I think Leor is superior to ZZ. That, plus Leor has less L-moves, and it's often possible to combine EO-Middle and a 2x2 block on the right.
 
Last edited:

Petro Leum

Member
Joined
May 3, 2012
Messages
800
Location
Germany
WCA
2012KALH01
Oh my, this thread is a gem. I don't know how I missed it until now.

As someone who has used CFOP until about sub15, and now averages ~9.25 with ZZ ( ~12 onehanded), let me give you my perception of the pros and cons of the method, when compared with CFOP:

Cons:

-While EOLine by itself is just as easy as Cross, and only like 1 move longer, it is significantly harder to plan anything after EOLine (tracking pieces of a 1x2x2 through seemingly arbitrary moves that solve EO is harder than straight blockbuilding for XCross) HOWEVER: EOCross is extremely easy to plan every solve and only adds 1-2 moves on top.

-L2/R2 moves are bad, and blockbuilding is hard and awkward. While a bit overstated, this is true (though only for two-handed) and it is why I nowadays believe EOCross is superior to EOLine, only for two-handed. You sacrifice some movecount for better inspection, easy lookahead and crazy TPS spam in F2L.

-RUL moveset in F2L is bad due to frequent regrips and slow LU. This is true, but can be a) mostly alleviated by solving the left block first (it has much fewer moves then and you get an all-RU-block afterwards or b) completelely avoided by using z-Rotations. for example, one could sovle the right block first, rotate z and solve another right block, getting to LL on what was originally the left side. Alternatively, one could just solve EOLine on left and solve the left block first with RUD, then rotate z' and solve the right block with RU. Since I struggle recognizing ZBLL on different colours and never learned to start a solve from Z, I myself often solve the right block first if it's easy and then rotate z to have a completely RU left block.

-The last slot cases are long and awkward. Yeah, since you have to stay 2gen, they are longer in movecount, but because CFOP ones often include rotations, they are often just as fast to execute in reality. 2 or 3 of these cases really suck though, without question. This is also why I advocate not solving 1x2x2s on each side and then doing essentially 2 LS-inserts, but rather finishing one side completely with blockbuilding before starting the other one. This reduces regrips and movecount.

-ZZ is not very flexible for use with other puzzles/events. I cannot understate this. ZZ is absolutely, completely, utterly useless for 4x4, and I would advise noone who wants to get good at 4x4 to use it. I myself have tried many ZZ variants for 4x4 and none of them come close to Yau+CFOP, even though I don't even know full OLL. It is not useful for Megaminx, neither for 2x2 (even though COLLs work as CLLs, they are not very nice to execute, and you might struggle with colour neutrality when coming from ZZ). ZZ can work well on uneven big cubes like 5x5 and 7x7, but that's about it

Pros:

-Extremely Easy lookahead. Because you always solve in the same one or two orientations, you do not have to be aware of where your pairs go - the same colours go to the same place everytime. There are also just plain fewer F2L cases because the edges are already oriented - this makes keyhole-cases which solve two slots at once extremely easy - you do not have to be aware of how the edge is oriented, there is always only one possible case.

-Nice execution due to no rotations and a lot of 2-gen. While rotations aren't that bad for twohanded, it's still noticeable and the long streaks of 2-gen solving are a dream. For onehanded, this alone makes ZZ much, much better than CFOP could ever hope to be.

-Many possibilities for last layer. Every alg you learn for last layer is insanely useful. From OLL/PLL over COLL/EPLL, winter variation, 2gen reduction, phasing + ZZLL all the way to full ZBLL, everything is worth learning and isn't just one random trick you might use in one of a hundred solves. While with CFOP, you can learn all these things and they might prove useful, it is hard to have a reliable consistent LL system. Full ZBLL will not be used on every CFOP solve, making the method itself inconsistent. I have learned all of ZBLL and just having 1 look and 1 alg on every solve is great and has dropped my times significantly. This also makes ZZ very interesting and anti-boredom. There's honestly so much to learn.

-Efficiency. If you do EOCross and ZBLL you are still at around 50 movecount, lower than the average CFOP solve. If you do proper Blockbuilding and ZBLL, staying under 50 is very easy. This again is great for onehanded.

So, to conclude: in my opinion, ZZ has a lot of potential. I'm not sure if it is better than CFOP for twohanded, but it's surely just as good. It doesn't compare to the god method Roux, though. For onehanded, ZZ is vastly superior to CFOP in absolutely every aspect. How some people think CFOP could be better than ZZ for onehanded eludes me. I used to think it's better than Roux as well, but I misjudged just how fast people could get at table abusing M slices. So I would still crown Roux the overall King Method for 3x3. (If table abuse were prohibited, though..... ;))
 

adimare

Member
Joined
Oct 29, 2009
Messages
381
Location
Costa Rica
WCA
2011MARE02
-Extremely Easy lookahead. Because you always solve in the same one or two orientations, you do not have to be aware of where your pairs go - the same colours go to the same place everytime.
I've been using ZZ for years now, yet whenever I do a random CFOP solve I seem to have no issues with this. Maybe a beginner in ZZ will spend less time figuring out where a pair goes than a beginner in CFOP does, but after a bit of practice I'm pretty sure the difference will either be negligible or non-existent.

There are also just plain fewer F2L cases because the edges are already oriented
That's debatable. Yes, objectively there are half as many cases for F2L pairs in ZZ, however:

1) You're ignoring the fact that in ZZ you're not just solving F2L pairs, you're building 3x2x1 blocks, which, if you're doing it right, require you to first connect corners with cross edges half the time, something you never have to do in CFOP F2L. That doubles the amount of corner+edge pairing cases you have to deal with in ZZ and puts it on par with CFOP.

2) At least to me, these two cases, for instance, are the same:
visualcube.php
visualcube.php


And even though the second one won't ever show up in ZZ, the fact that it does show up on CFOP doesn't really hinder lookahead. You don't recognize it based on the exact colors you're looking at, you recognize it based on the relationship between the position of the corner's white sticker with the edge.

ZZ vs CFOP lookahead could probably be its own topic with pros and cons for each one. Not having to deal with edges in the bottom layer at all is a huge pro for CFOP and a major con for ZZ lookahead during F2L, for instance.
 

Pyjam

Premium Member
Joined
Oct 8, 2010
Messages
2,142
Location
La Baule, France
For your 1st case, I know 6 different algs depending on the position of blocks on the top layer in order to influence the last layer case.

R' D' R U' R' D R U R U' R'
R U2' R' U R U R' U R U' R'
R U2' R' U L U' R U R' L'
U' R' U2 R' U R' U' R U2 R
U R U' R' U' R' F' R U R U' R' F
R U R' U2 R U' R' U R U' R'
 

GenTheSnail

Member
Joined
Mar 18, 2016
Messages
2,249
Location
Illinois, USA
WCA
2016GEEN01
YouTube
Visit Channel
[ZZ] is not useful for Megaminx
I agree with most of your post but this gets me...
It can't work on F2L, but I use ZZ for my S2L and it's great. I'm essentially sub1 on megaminx with my ZZ adaption. I would consider that not amazing but definitely pretty good. And I'm still improving.

Also it's pretty good for Feet.
But yeah, nothing else.
 

PapaSmurf

Member
Joined
Jan 4, 2018
Messages
1,103
WCA
2016TUDO02
YouTube
Visit Channel
You can get it to work on 4x4. I have a 4x4 method (4z4) which works pretty well. Also, one other thing: I do agree that it is definitely as good as CFOP with 2H (I’d be tempted to say that it’s better, due to lower movecount, better LL), but I’d disagree with Petro Leum about eocross. I’d say that 3/4 eocross is better, as it still allows the block building. And to all those CFOPers: yeah, L2/R2 sucks, but so does RUyLU. Also, on the topic of is eo worth it, ZZ averages much fewer moves than CFOP with fewer algs, so I’d say so (going on wiki stats). You only need 46 algs to be around 52 moves, compared to 78 for 55 with CFOP. If you say “alg number doesn’t matter”, you get 45 with 493 zbll, and 45 with either full zb (which is where I’d say CFOP is at its full) or 1lll, which are both a lot more algs than 493. Then you go to 1llsll, which is a lot more possible in zz than cfop (even if it’s impossible practically). Basically, zz is superior in terms of movecount, last layer and is (IMO) equal in movegroup. It isn’t as good in EoLine to F2L transition, but that can be practiced.
 

Petro Leum

Member
Joined
May 3, 2012
Messages
800
Location
Germany
WCA
2012KALH01
I've been using ZZ for years now, yet whenever I do a random CFOP solve I seem to have no issues with this. Maybe a beginner in ZZ will spend less time figuring out where a pair goes than a beginner in CFOP does, but after a bit of practice I'm pretty sure the difference will either be negligible or non-existent.

Hmm.... I have been using ZZ for 6 years and I always struggle with this when I solve with CFOP. I can do it, no problem, but it does cost me cognitive capacity that could be used for looking ahead. I'm not sure whether this doesn't make a difference for advanced solvers too.

I've been using ZZ for years now, yet whenever I do a random CFOP solve I seem to have no issues with this. Maybe a beginner in ZZ will spend less time figuring out where a pair goes than a beginner in CFOP does, but after a bit of practice I'm pretty sure the difference will either be negligible or non-existent.


That's debatable. Yes, objectively there are half as many cases for F2L pairs in ZZ, however:

1) You're ignoring the fact that in ZZ you're not just solving F2L pairs, you're building 3x2x1 blocks, which, if you're doing it right, require you to first connect corners with cross edges half the time, something you never have to do in CFOP F2L. That doubles the amount of corner+edge pairing cases you have to deal with in ZZ and puts it on par with CFOP.

2) At least to me, these two cases, for instance, are the same:
visualcube.php
visualcube.php


And even though the second one won't ever show up in ZZ, the fact that it does show up on CFOP doesn't really hinder lookahead. You don't recognize it based on the exact colors you're looking at, you recognize it based on the relationship between the position of the corner's white sticker with the edge.

ZZ vs CFOP lookahead could probably be its own topic with pros and cons for each one. Not having to deal with edges in the bottom layer at all is a huge pro for CFOP and a major con for ZZ lookahead during F2L, for instance.

You are technically right in that it is the same number of cases, but if you say that, you ignore that there IS a difference in that you have to recognize the edge orientation (albeit indirectly, by checking where the pair is located relatively to the slot and then rotating/not rotating) which is something you just don't do in ZZ. Also keep in mind while these two cases appear equal from a CFOP solvers perspective, so do similar cases where you either pair a cross edge or a pair edge with a corner.

That said, I still believe EOCross for twohanded is superior, which is basically just easy mode CFOP F2L with no doubt.

I agree with most of your post but this gets me...
It can't work on F2L, but I use ZZ for my S2L and it's great. I'm essentially sub1 on megaminx with my ZZ adaption. I would consider that not amazing but definitely pretty good. And I'm still improving.

Also it's pretty good for Feet.
But yeah, nothing else.

Hmm, I don't know how that works for megaminx. Can you show me an example/ is there a tutorial of sorts?

And yeah, I totally forgot about feet, which might be ZZ's greatest potential. Shame so few people actually use the method!
 

Petro Leum

Member
Joined
May 3, 2012
Messages
800
Location
Germany
WCA
2012KALH01
...I’d say that 3/4 eocross is better, as it still allows the block building. And to all those CFOPers: yeah, L2/R2 sucks, but so does RUyLU. ...

True, and I believe this is what I automatically do to preserve good movecount in actual solves. For me it's mostly a matter of inspection approach. I do always plan full eocross - but I don't always execute it for the benefit of abusing nice block building cases.

And about CFOPers: Keep in mind many of the good CFOPers actually only du RUy.... which is ugly, but surely higher TPS than RUyLU.
 

CarterK

Member
Joined
Oct 19, 2017
Messages
517
WCA
2015KUCA01
YouTube
Visit Channel
Good CFOPers use all moves when applicable, where they are good. Stuff like F' U F and f R f' really isn't that bad. You can also be super efficient(and avoid rotations!) in F2L with some really cool stuff that you can't do in ZZ, such as S R S' and R B' U' B R'.
 

PapaSmurf

Member
Joined
Jan 4, 2018
Messages
1,103
WCA
2016TUDO02
YouTube
Visit Channel
That's because it's harder to directly compare. With CFOP and ZZ, you can compare cross to eoline, f2l to f2l and ll to ll. With Roux, I guess you have to compare the advantages and disadvantages of the whole method than the steps. So, roux has a lot of gen regrips, so whole cube to RrUM to whole cube (potentially) to MU. ZZ has whole cube to RUL to whole cube (potentially). The obvious advantage of roux is that you need less algs. So even if you learn 2 cmlls for each case+eolr, you'll be learning about 1/4 of the algs as full zbll for the same (or maybe even less by about 2 moves) than ZZ-A. You have bling spots in both, so for roux you have (in second block) DB, DR and BR, while with zz you have DL, BL, DR and BR (although doing some form of eocross will remove one or two of these). I'd argue (from these points) that ZZ is as viable and is on par with roux, and CFOP from the above points.
 

CarterK

Member
Joined
Oct 19, 2017
Messages
517
WCA
2015KUCA01
YouTube
Visit Channel
That's because it's harder to directly compare. With CFOP and ZZ, you can compare cross to eoline, f2l to f2l and ll to ll. With Roux, I guess you have to compare the advantages and disadvantages of the whole method than the steps. So, roux has a lot of gen regrips, so whole cube to RrUM to whole cube (potentially) to MU. ZZ has whole cube to RUL to whole cube (potentially). The obvious advantage of roux is that you need less algs. So even if you learn 2 cmlls for each case+eolr, you'll be learning about 1/4 of the algs as full zbll for the same (or maybe even less by about 2 moves) than ZZ-A. You have bling spots in both, so for roux you have (in second block) DB, DR and BR, while with zz you have DL, BL, DR and BR (although doing some form of eocross will remove one or two of these). I'd argue (from these points) that ZZ is as viable and is on par with roux, and CFOP from the above points.
You also forgot that Roux is 10+ moves more efficient than ZZ and CFOP.
 

GenTheSnail

Member
Joined
Mar 18, 2016
Messages
2,249
Location
Illinois, USA
WCA
2016GEEN01
YouTube
Visit Channel
Hmm, I don't know how that works for megaminx. Can you show me an example/ is there a tutorial of sorts?

And yeah, I totally forgot about feet, which might be ZZ's greatest potential. Shame so few people actually use the method!
I came up with my method (ZZ-Spike) almost two years ago, and I made an example solve video year later.
https://www.speedsolving.com/forum/threads/zz-spike-walk-through-solve-zz-on-megaminx.65029 (this is the same link as the one in my sig)
I never made a formal written tutorial and in the video I completely skip over S2L but it's the same as ZZ F2L but with an extra side. I still use that method and have gone officially sub 1--I also have a handful of sub 50s and a PB 53 ao5.
And yeah, I know that just because a method can achieve fast results doesn't mean that it's a fast method, but so far I haven't hit any barriers that indicate that it's a slower method than others so far.

Sure you can do ZBLL, but that slows it down.
?

At first sure, but ZBLL is ultimately faster than OCLL/PLL or COLL/EPLL. I'm always like yes! a ZBLL I don't have to do another alg afterwards!
Unless I'm still learning the set and am like crud what is this case.
 
Top