[Help Thread] 4x4 Discussion

[CornerCutter]

Thank you for the info guys! I will practice more. I think part of the problem is my cube because I can solve the 3x3 in 20sec but the 4x4 takes about 40sec.

 

[dskids]

Hi guys,

I'm currently averaging about 2:00 on 4x4 with Yau (cross edges on left, normal centers, 3-2-3). I want to get to sub 1:30. I'm thinking about learning half centers method, but before I spend a bunch of time learning any new techniques I'm just wondering:

What are appropriate splits for a 1:30 solve? Need to figure out where my main deficiencies are. Thanks!

 

[pipkiksass]

Hi guys,

I'm currently averaging about 2:00 on 4x4 with Yau (cross edges on left, normal centers, 3-2-3). I want to get to sub 1:30. I'm thinking about learning half centers method, but before I spend a bunch of time learning any new techniques I'm just wondering:

What are appropriate splits for a 1:30 solve? Need to figure out where my main deficiencies are. Thanks!

At this stage, I really think it's more a question of practice. I use half centres, and love em, but otherwise same as you and I'm averaging about 1:30.

What do you average in 3x3? IMHO split times depend on 3x3 at least to some extent. For example, I average about 19 in 3x3, and my 3x3 stage on 4x4 takes about 30s on average, including parities. So basically I'm averaging 60 seconds for "Yauduction", then a further 30 for 3x3 stage.

Couldn't say what an idealised split is for F2C, cross edges, and L4C, but if you let me know your 3x3 times I'll find out. There's a cubing world vid from a few years back which mentions sub 1 splits, and I could extrapolate them up.


Sent from my iPhone using Tapatalk

 

[dskids]

At this stage, I really think it's more a question of practice. I use half centres, and love em, but otherwise same as you and I'm averaging about 1:30.

What do you average in 3x3? IMHO split times depend on 3x3 at least to some extent. For example, I average about 19 in 3x3, and my 3x3 stage on 4x4 takes about 30s on average, including parities. So basically I'm averaging 60 seconds for "Yauduction", then a further 30 for 3x3 stage.

Couldn't say what an idealised split is for F2C, cross edges, and L4C, but if you let me know your 3x3 times I'll find out. There's a cubing world vid from a few years back which mentions sub 1 splits, and I could extrapolate them up.


Sent from my iPhone using Tapatalk

Yeah, if you don't mind looking it up, I'd really appreciate the info! Currently averaging about 28 sec on 3x3. I spent some time doing slow solves last night to get a feel for half centers and I'm really intrigued, thinking about putting in the time to get faster with it. I also just started doing cross edges on the M-slice with cross on left. I have a feeling I'm spending a disproportionate amount of time on that step currently.

 

[One Wheel]

I also just started doing cross edges on the M-slice with cross on left. I have a feeling I'm spending a disproportionate amount of time on that step currently.

That is probably the hardest step because of how many places there are too look for pieces. I think I've gotten good enough at yau reduction that my worst step in 4x4 is 3x3, but first three cross edges together with parity pretty much make the difference between a 1:40 solve and a 2:05 solve. The best I can suggest is to start looking for your cross edges when you are making your second center, but I'm curious if anybody else has other suggestions. I'm sure it's possible to plan f2c and maybe f3e in inspection, but I'm usually doing well to plan my first center and maybe a little of the second.

 

[mark49152]

A few years ago I analysed 20 reconstructions of fast Yau solves, sub 30. Average splits were as follows.

9.8% F2C
11.5% F3E
17.2% L4C
4.0% cross
23.4% edge pairing
34.2% 3x3 stage

 

[pipkiksass]

A few years ago I analysed 20 reconstructions of fast Yau solves, sub 30. Average splits were as follows.

9.8% F2C
11.5% F3E
17.2% L4C
4.0% cross
23.4% edge pairing
34.2% 3x3 stage

Thanks Mark.

So @dskids, at 2 mins you're looking about 12s for F2C, 14 for F3E, 20 for L4C, 5 for last cross edge, 28 for L8E, 41 for 3x3.

For me, it's closer to 9/10/15/3/21/30.

I'll time some splits later, interested to see where I'm slow. I'd imagine most is transitions, as I'm pretty sure I'm well ahead on edge pairing and 3x3, even with parity.


Sent from my iPhone using Tapatalk

 

[pipkiksass]

Timed some splits, which gave some VERY interesting results.

Apologies for the double-post, but this is too much for an edit!

Spoiler: analysis, for anyone who cares

As a % of the average, that gives the following (Mark's idealised splits in brackets):

F2C - 11.36 (9.8)
F3E - 17.95 (11.5)
L4C - 16.14 (17.2)
LCE - 5.45 (4.0)
L8E - 23.86 (23.4)
3x3 - 26.82 (34.2)

So, looking at the AVERAGES, my F2C and F3E are a little bit slow. I knew this was the case, I was aware that they are by far my weakest step. My L4C is ahead of where it needs to be. My last cross edge is a bit sucky. My edge pairing is right on where it needs to be. My 3x3 stage is WELL ahead of where it should be. In other words everything else is actually behind, but being helped along by my solid 3x3 stage.

Looking at the individual times gives a little more info. My first 2 centres SUCK. I should be doing this step a second faster, on average. Something to work on, but not as much as...

F3E - even my best time (14s) is 16% of the solve, vs. an idealised 13. On average, I'm nearly 3 seconds behind on this step. I knew it was bad, but didn't realise HOW bad. My worst solve took me 31 seconds to get to this stage, vs. an idealised 18.7. OUCH!!!

Last 4 centres I'm well ahead of where I should be. I use half centres, and the 19 really hurt my mean here. Without the 19, this would have been 13.xx, and around 15%, vs. idealised 17.2.

Last cross edge sucks. I don't know any 'tricks' for this, and haven't invested any time playing with it. Any tips for practicing it would be appreciated, but it's a bit of a "meh" step, as it's so brief.

Edge pairing, I'm right where I thought I was. I had a couple of really bad cases, which hurt the mean once again. On a good solve, I'm looking at 16/17 seconds, which is around 18%, vs. 23.4 idealised.

So it looks like my 3x3 stage is completely carrying my 4x4 times! Pretty ludicrously consistent, apart from the 29 (being double parity). Without parity I'm sure these would have been 22/23 across the board.

So (no real news to me) it looks like I need to practice F2C and cross edges. From that stage onward, I think I'm 'ahead of the curve', but my solid centres, pairing, and 3x3 are definitely compensating for a really crappy start to my solves.

I'd highly recommend doing this - took a little while, as I don't have a camera setup, and had to improvise, but it gives some good info about your solves. I haven't recorded myself doing 4x4 before, or seen myself solve, and there's a lot of room for improvement!!! :-D

 

[One Wheel]

A few years ago I analysed 20 reconstructions of fast Yau solves, sub 30. Average splits were as follows.

9.8% F2C
11.5% F3E
17.2% L4C
4.0% cross
23.4% edge pairing
34.2% 3x3 stage

Is this skewed toward easy solves/ solves without parity? You say they were sub - 30 solves, would it give noticeably different results if you analyzed several random solves by a few people who are generally fast, or did you take that into account?

 

[mark49152]

Is this skewed toward easy solves/ solves without parity? You say they were sub - 30 solves, would it give noticeably different results if you analyzed several random solves by a few people who are generally fast, or did you take that into account?

I wanted idealised splits to aim for, rather than typical splits. I worked from reconstructions not videos. Where the reconstructions gave coarser time splits, like a single time for all of redux, I used ETM to estimate how those phases further broke down. Some of the solves did have parities but I'm not sure I noted all of those.

I chose to focus on faster solves because I assumed that they would give results closer to the ideal, due to fewer anomalies where say a bad step wastes a couple of seconds, and also due to fewer pauses so the extrapolation from ETM to time might be a bit more reliable. But it's true there's a bias and there's also a likelihood of some positive anomalies where say a step was particularly easy.

There's a lot of approximation and assumption in there.

 

[Wannabefeliks]

hi I got into 4x4 a few weeks ago and currently average around 2:50 but my times are fairly erratic. I use Yau method, half centers for the L4C and 3-2-3 edge pairing. Any tips to help me get faster would be appreciated.

 

[pipkiksass]

hi I got into 4x4 a few weeks ago and currently average around 2:50 but my times are fairly erratic. I use Yau method, half centers for the L4C and 3-2-3 edge pairing. Any tips to help me get faster would be appreciated.

Just practice. My times plummeted within a couple of weeks, and started getting more consistent once I started to deal with bad edge pairing cases more consistently.

Make sure you're being efficient with L4C - you can practice them by scrambling cross on left with just U, and R layer turns.

Turn slowly during F2C and cross edges, and try to be smooth and look ahead.

Join a forum race for 3x3 stage on 4x4 - I'm currently in the race to sub -25, but with my Yuxin Blue 4x4. This helps with general TPS and turning style on 4x4.

But mostly, just practice. Don't time all your solves, you learn more from slow practice.


Sent from my iPhone using Tapatalk

 

[Noblesuntzu]

I got a 4x4 Rubik's cube for Christmas, and I've already solved it. I suck at cubing, so I used the basic middle piece's first, then 3x3 basics. I'm kind of confused though because the orientation of the first layer corners is incorrect, which I thought impossible. Is this paradox? I severely hope not, too many turns lol.

Webcam-toy-photo2

• Noblesuntzu
• Dec 26, 2016

Lol forgot to add the picture

 

[GenTheSnail]

Your picture does not provide enough information to determine if the cube has parity. Please solve more of the cube, and then take another picture.
But, from the picture, it looks like you have build part of a face, but not a layer. Simply put the corners in correctly. There is no corner orientation parity.

Additionally, there are 2 parities to the 4x4x4 Cube, OLL and PLL parity. These parities are not usually noticed and *do not* affect the first layer. [*yeah they effect the whole cube, but not really*]
Algs to solve either parity:

If you have an uneven number of flipped edge(s), then use this algorithm to flip one of those edges:
r' U2 l F2 l' F2 r2 U2 r U2 r' U2 F2 r2 F2 - where each r/l is a slice move

If you have an uneven number of dedge and corner cycles, use this algorithm to swap two of those pieces:
Rw2 F2 U2 r2 undo moves U2 F2 Rw2 - same here for that slice move

 

[sqAree]

Tbh it looks like you built the centers wrong.

 

[CLL Smooth]

Webcam-toy-photo2

• Noblesuntzu
• Dec 26, 2016

Lol forgot to add the picture

Looks like a paradox to me. Or a mean joke.

 

[Rcuber123]

Or he solved the centers wrong.
Swap a pair of opposite centers

 

[asacuber]

Do Rw2 D2 U2 Rw2 from any angle

 

[Dom]

Do Rw2 D2 U2 Rw2 from any angle

When I'm doing a Hoya solve, I find it very useful to turn the top face towards me when I do this alg. That way, it won't mess up my cross.

 

[Noblesuntzu]

Yeah I was being an idiot lol. Thanks for the help guys