Merits of CFOP

[sneaklyfox]

it's 洪.

Too bad. I was hoping it would be the Confucius 孔 (mine). I don't often meet too many other descendants though KongShou on the forums is apparently a distant relative. Sorry for being OT. Had to ask.

 

[Kirjava]

there is slightly more intuition which slows the solve down

intuition becomes algorithmic, I disagree with your statement

 

[CheesecakeCuber]

intuition becomes algorithmic, I disagree with your statement

I agree with kirjava. Haven't you heard about muscle-memory and closing your eyes to solve F2L pairs?

 

[KongShou]

Too bad. I was hoping it would be the Confucius 孔 (mine). I don't often meet too many other descendants though KongShou on the forums is apparently a distant relative. Sorry for being OT. Had to ask.

a very distant one, i have to add. since you seem to come from guangzhou or hongkong. and spell your name like hung. I do it the mandarin way, kong(孔) , also my dad come from HeiLongJiang(黑龙江) they're like at the opposite end of china lol.

anyway back on topic.

i agree with kirjava

my f2l is basically algorithmic, even tho i never learnt any f2l algs.

 

[ravenguild08]

Heise is very good for FMC but is very very hard to understand and get used to using it sucks for speed.
But with FMC you use whatever would get you you the best result.

Yeah, in FMC you just sort of use whatever works. Method neutrality is ideal.

This post was written in the frame of speedsolving, but I also wasn't aiming to claim that CFOP is superior to the other methods. And well, it's not superior anyway.
Tim Major and TDM say that rotations are a big weakness for CFOP, and... I guess I agree? I don't know why I didn't write about it originally.

And like everyone else, I also agree with Kirjava. Back in 2007, I actually figured out slotting by myself and discovered nearly all the cases through intuition and experimentation. These days, all of them have turned into pure muscle memory.


Also, KongShou and sneaklyfox, I'm of Taiwanese descent, so I don't think I'm too closely related either...

 

[Dapianokid]

I discovered slotting and thought it was the secret to fast cubing Then I learned X-cross!
I do have to admit that F2L pairs becomes algorithmic, although in slow turn solves it is to a much lesser degree than in fast turn solves. It (CFOP) is a 7 look method, as the F2L slots each take one look, and in some cases you can skip ahead.find shortcuts where you can solve two slots simultaneously. (cross takes one look and X-cross can too; LL takes two looks for advanced cubers and 2look-OLL can even be achieved with 1 look with a little practice)

I'm finding cube rotations to be a big problem as I get significantly faster. When I focus on speed, I really like to move fast without it necessarily being super-efficient. I usually try to either keep the cube fluidly rotating and take advantage of every face and move without regard to finger tricks during a slow solve during F2L, or I move really fast and solve the closest pair I can find while looking ahead at the next pair using R and U turns only, then rotating the cube and solving the next pairs. This doesn't work well when the pieces I want are stuck in other slots, so I have to either do a really slow B turn or do at least one cube rotation.

Petrus takes a lot of cube rotations during the first two steps, however. More than CFOP, I can promise you that. then, you get a few seconds of solving 5 pieces + a center to finish up F2L and then you have LL which can be solved however you want (I used to use Petrus LL, it works really well for beginners)...
I think Roux is honestly the fastest method. Few cube rotations, more efficient than ZZ and CFOP (unless you're a blockbuilding god and like FreeFOP), and it's easy to understand. I suck at it, though.

 

[kcl]

I think with time Roux could possibly become King. However, I think CFOP and Roux will remain about the same speed. More people in the future will just become fast with roux.

 

[Chree]

I think with time Roux could possibly become King. However, I think CFOP and Roux will remain about the same speed. More people in the future will just become fast with roux.

I dunno... with time? Maybe. But it's gonna be a loooooooong time. Like the article says, the momentum of CFOP is pretty monstrous. The Roux tutorials are needles in haystacks. Beginners will most like stumble across CFOP first.

And this may just be me, but so many M slices and Rw's, though ergonomic, is not intuitive. CFOP is. Newbies will be flocking to it for a long, long, long time.

 

[kcl]

I dunno... with time? Maybe. But it's gonna be a loooooooong time. Like the article says, the momentum of CFOP is pretty monstrous. The Roux tutorials are needles in haystacks. Beginners will most like stumble across CFOP first.

And this may just be me, but so many M slices and Rw's, though ergonomic, is not intuitive. CFOP is. Newbies will be flocking to it for a long, long, long time.

Look at Alexander Lau.. It took him 2 years.

 

[TheOneOnTheLeft]

And this may just be me, but so many M slices and Rw's, though ergonomic, is not intuitive. CFOP is.

On top of this, M slices are a pain on a standard Rubik's brand, so it's harder for beginners to get into Roux until they've upgraded to a better cube - CFOP doesn't have this issue.

 

[bundat]

I find it hilarious that a "Rubik's storebought" was mentioned RIGHT AFTER a mention of Alexander Lau.

 

[Tim Major]

Whilst you do your first pair you can look ahead to your next couple of pairs, because you solve pairs the same 70% of the time. In Roux, there aren't "2nd block" cases, so whilst you can plan out your first block, and fast Roux solvers can plan some of their second block, there is slightly more intuition which slows the solve down. Generally, the more brain dead a method is, the faster it is, as you can just turn.

i agree with kirjava

my f2l is basically algorithmic, even tho i never learnt any f2l algs.

I basically said that F2L is superior in terms of TPS/lookahead, because all cases become algorithmic, removing the intuition, speeding the solves up. Since I average low 20 on proper Roux (sub 20 if I do it inefficiently) and 13~ with CFOP, I can't give an expert opinion. However, I feel like F2L cases become algorithmic over time removing intuition speeding up solves, whereas the 2nd block in Roux which I don't plan entirely in inspection, is slowed down as you need to think rather than just spam cases you've done thousand times.

So what I mean is I agree with "my f2l is basically algorithmic, even tho i never learnt any f2l algs", I think Kirjava was saying something different. I stated that F2L = algorithmic and Roux 2nd block = intuition, but if I understand Kirjava correctly then he is saying that Roux 2nd block becomes algorithmic too.

@CheesecakeCuber: I have taught many people CFOP, and given tips and my biggest tip is generally "learn F2L pairs intuitively enough that they become algorithms you can do with your eyes closed, so I certainly have heard of the "closing your eyes to solve F2L pairs. I have been cubing since 2009 and I've helped/tried to help a lot of people with this technique.

 

[GuRoux]

I basically said that F2L is superior in terms of TPS/lookahead, because all cases become algorithmic, removing the intuition, speeding the solves up. Since I average low 20 on proper Roux (sub 20 if I do it inefficiently) and 13~ with CFOP, I can't give an expert opinion. However, I feel like F2L cases become algorithmic over time removing intuition speeding up solves, whereas the 2nd block in Roux which I don't plan entirely in inspection, is slowed down as you need to think rather than just spam cases you've done thousand times.

I feel roux is better than F2L when it comes to lookahead and maybe TPS for the first two blocks. First off, there are fewer pieces to keep track of in the second block compared to 4 F2L pairs. It should be easier to predict where the pieces will end up and make decisions when there are only a few pieces to work with isolated to one side of the cube. After a while of doing roux, you almost never have to look at the bottom or the back to find pieces. There is no need for cube rotations. And is shouldn't take long for first two block building to become as algorithmic as cfop.

 

[Tim Major]

And is shouldn't take long for first two block building to become as algorithmic as cfop.

This is the only part I took out of your post that changes my knowledge/arguments. All CFOP users under 15 seconds could likely do 2 F2L pairs blindfolded with a quick glance (IE, once the cross is done, glance at 4 pieces and solve both pairs). Can you glance at all pieces of the 2nd block and then do it like it's a CLL, as in, recognition then fast turning algorithm? If so, then that would change my argument


However;

I feel roux is better than F2L when it comes to lookahead and maybe TPS for the first two blocks.

This is an opinion, which is formed by the fact that Roux is your main method. I turn much much much faster in F2L than in F2B, and I can lookahead much easier with little to no pauses. Of course you would be better at lookahead/TPS for your main method.

 

[JasonK]

All CFOP users under 15 seconds could likely do 2 F2L pairs blindfolded with a quick glance (IE, once the cross is done, glance at 4 pieces and solve both pairs).

Maybe it's just me, but I avg 13 and I definitely can't do this with a "quick glance".

 

[bundat]

I actually analyzed statistics relevant to this in my other topic "How do top-cubers lookahead", and found that:

Alex Lau averages 9+ tps for second block.
(based from reconstructions' stats, e.g. sub-2 9.29 tps second block in the Danish challenge vid)

Faz and Mats are also around the same.

Faz's 4.79 unofficial has almost 10 tps for his F2L.

Mats 5.55 WR's F2L is almost 10 tps too.

I analyzed these because I didn't think these were speeds where you could even lookahead properly.

But at least Alex proves that Roux's "intuitive/lookahead step" (2nd block, also L6E) can be as fast as CFOP's (F2L).

Algorithmic or intuitive, building the 2nd block in 17 moves in 1.83 seconds (in the Danish challenge) at least proves that needing "slightly more intuition" doesn't seem to slow the solve down.

Also interesting to note, his CMLL average tps is 4-7 in many solves.
Also, in Faz's 4.79, his LL is only 6 tps (despite having an almost 10 tps F2L)
Also, in Mats' 5.55, his LL (simple CCW EPLL) is below 7 tps (but almost 10 for F2L).

And so it actually seems that "recog+drilled alg" steps are actually potentially slower than "lookahead+intuition" steps, due to the recog pause. (except for rare exceptions, such as Alex's 5.11, where his CMLL is 11.11 tps... I think it was either much faster, like his 14 tps perm vid, or that he was able to predict the CMLL).

All CFOP users under 15 seconds could likely do 2 F2L pairs blindfolded with a quick glance (IE, once the cross is done, glance at 4 pieces and solve both pairs).

Maybe it's just me, but I avg 13 and I definitely can't do this with a "quick glance".

Have a look at Rowan's posts in my thread that I mentioned.

He goes into detail on how being able to "predict" the entire F2L, and how such lookahead supposedly totally separates 12-13 second cubers from consistent sub-10/9/etc cubers.

 

[JasonK]

Have a look at Rowan's posts in my thread that I mentioned.

He goes into detail on how being able to "predict" the entire F2L, and how such lookahead supposedly totally separates 12-13 second cubers from consistent sub-10/9/etc cubers.

I know all about that. I wasn't asking how to do it, I was just challenging Tim's idea that anyone who's sub-15 could do it.

 

[ottozing]

Maybe it's just me, but I avg 13 and I definitely can't do this with a "quick glance".

I average mid 9 and honestly almost never do 2 F2L pairs with a quick glance (I also find it kinda hard to do when I'm purely trying to do this exercise). My thought process during a CFOP solve is.

Inspection: Find the cross and the first pair (If I can't see the first pair, I'll modify my solution so it had more rotations and wide U moves so I have more chences to find my first pair).
After my cross: Assuming I plan the first pair, solve the pair while tracking an F2L corner and maybe an edge depending on if I find it quick enough. If not, I'll just solve the first pair I see in front of me. If I only see a corner, I rotate and look for the edge.
After my first pair (This thought process repeats until I'm up to LSLL): Find a corner (Generally I would have tracked the next corner anyway so this isn't too hard) and either find the corresponding edge with either instant recognition or through look ahead (More often than not it will be the former, but doing the latter isn't that uncommon, especially during the final few pairs as finding pieces gets easier).
During the last pair: Determine my LL EO type (0 right, 2 right, or 4 right) and influence it how I see fit. If I know I'll get an EO skip and I can reduce the last pair to R U' R', I'll do WV (I'll occasionally do this for R' U R and L' U L, but not that much).
During LL: Recognize the OLL and the CP (Unless it's a sune case which is the only CLL set that actually takes some effort to recognize for me, or if I know from the little CP info I have from first glance that my PLL won't be diag and/or can't be an OLLCP case I know), then solve the OLL with whatever CP influence I see fit.
Then PLL is pretty obvious.

Maybe I'm just a nub compared to most sub 15 solvers............

 

[stoic]

My thought process during a CFOP solve is

Thanks for posting this, I found it fascinating.

 

[bundat]

Wow, that's pretty impressive that you got that far with "corner bias" type of lookahead.
Do you know your tps during F2L?