[Help Thread] 4x4 Discussion

[uberCuber]

Does anybody have a good trick for "O-perm" on 4x4? Aka PLL parity into Z-perm. Or adjacent parity U U-perm U' which also isn't very fast.

Here's an O-perm alg, which is basically a cancellation with doing PLL parity in the middle of an MU Z-perm alg (using lowercase letters as slices):

M2 U M2 U M' U2 l2 U2 r2 Uw2 r2 u2 M' - counterclockwise
M2 U' M2 U' M' U2 l2 U2 r2 Uw2 r2 u2 M' - clockwise

I don't actually use these, though, because I can't do them as fast as RU Z-perm + PLL parity. It's that u2 before the final M' that kills me :/

 

[speedcuber50]

Have they changed the Dayan + mf8 4x4 since Feliks did his record? If so, why didn't PUZL mention that anywhere?

 

[Username]

Have they changed the Dayan + mf8 4x4 since Feliks did his record? If so, why didn't PUZL mention that anywhere?

Why does it matter? You keep asking what feliks did..

 

[EMI]

Here's an O-perm alg, which is basically a cancellation with doing PLL parity in the middle of an MU Z-perm alg (using lowercase letters as slices):

M2 U M2 U M' U2 l2 U2 r2 Uw2 r2 u2 M' - counterclockwise
M2 U' M2 U' M' U2 l2 U2 r2 Uw2 r2 u2 M' - clockwise

I don't actually use these, though, because I can't do them as fast as RU Z-perm + PLL parity. It's that u2 before the final M' that kills me :/

Thanks. I'll try these out.

 

[windhero]

Afaik Felix doesnt care about hardware like most cubers. In some interview I remember him saying that when he solves the 3x3 at home he just solves whatever cube comes into hand; not just a single specific main cube. Therefore what cube he uses is not necessarily the cube that is the best on the market. Unmodded SS 4x4s arent really that great but they are still much better than anything else out there. A Florian/konsta modded SS 4x4 is in a completely different league compared to any modded or un-modded 4x4 on the market.

 

[Username]

Afaik Felix doesnt care about hardware like most cubers. In some interview I remember him saying that when he solves the 3x3 at home he just solves whatever cube comes into hand; not just a single specific main cube. Therefore what cube he uses is not necessarily the cube that is the best on the market. Unmodded SS 4x4s arent really that great but they are still much better than anything else out there. A Florian/konsta modded SS 4x4 is in a completely different league compared to any modded or un-modded 4x4 on the market.

Yet the WR is set with an unmodded SS, aswell as the previous WR's by Feliks

 

[yoshinator]

Yet the WR is set with an unmodded SS, aswell as the previous WR's by Feliks

And yet that means absolutely nothing. The quality of the puzzle is my affected by who uses it, or whether it was used to break any records.

 

[Username]

And yet that means absolutely nothing. The quality of the puzzle is my affected by who uses it, or whether it was used to break any records.

That's not what I meant, I meant that unmodded cubes have potential to get fast times, just like modded ones.

 

[windhero]

I still believe that Felix would get a bump up in times with a modded SS 4x4 with custom stickers. Brighter shades and a puzzle that doesnt lock up + feels more like a 3x3 cant really make the times worse with my logic.

Btw I'm making a video of the SS 4x4 v5 and on how konstamodding affects it and how it compares to a florian modded one. The material is already filmed so I should be done with it today'ish.

 

[Username]

I still believe that Felix would get a bump up in times with a modded SS 4x4 with custom stickers. Brighter shades and a puzzle that doesnt lock up + feels more like a 3x3 cant really make the times worse with my logic.

Btw I'm making a video of the SS 4x4 v5 and on how konstamodding affects it and how it compares to a florian modded one. The material is already filmed so I should be done with it today'ish.

This is what I'm thinking too.

I can't wait to see your video


One question; I have a konsta modded SSv4, which I really like. People tell me that doing konstamod only is stupid and it needs the Florian mod to be good. Do you think it would make it any better? I'm afraid of overmodding the pieces, so that's why I haven't done it earlier

 

[windhero]

This is what I'm thinking too.

I can't wait to see your video


One question; I have a konsta modded SSv4, which I really like. People tell me that doing konstamod only is stupid and it needs the Florian mod to be good. Do you think it would make it any better? I'm afraid of overmodding the pieces, so that's why I haven't done it earlier

It's not stupid. I'll spoil a bit of my video with some stats I got. I only did Ao12's so the numbers arent well founded but it could be worse;

With a v5 straight out of the box I managed to get an 1:27.74 with a std deviation of 6.84 seconds Ao12
With a konstamodded v5 (also core lubed with lubix-like diff oil, pieces still with stock lube) I got 1:24.37 with a std. deviation of 4.81 seconds.

What we care about here is the std deviation as it basically tells us how stable our solves are or how many lockups or unexpected bumps we get. The higher the std deviation, the likelier it is that the puzzle locks up a lot.

Konsta modding made my v5 stabler by 42% based on these statistics. It also made my Ao12 3 seconds faster but that doesnt really matter.

I then tested a konsta modded non florian modded v5 vs full konsta/florian modded SS 4x4 v4 (Since konsta/florian modded puzzles are basically the same it doesnt matter that these puzzles are different versions).

I did and Ao12 with the 3x3 stage only as that is what the florian mod affects the most.

Florian modded: session avg: 26.39 (σ = 2.26)

Non-florian modded:
session avg: 26.40 (σ = 2.95)

The florian modded one performed 30% more stable than the non-florian modded one. Funny thing is my Ao12 with florian modded was 26.39 and with the unmodded one 26.40. So the difference in speed was 0.01, but the difference in how that average was formed was huge. I got much slower AND faster times on the puzzle that didnt have the mods (this is probably because I used the same 12 scrambles and tried it out with the modded one first, so I "knew the scrambles" when I tried with the unmodded, even though I did mix the scrambles up and there is no way I could memorize 12 scrambles this fast)


Conclusion; Both mods are equally necessary. The other helps with inner layer turns and the other with outer layer turns. For a perfect cube you need to do both.

 

[EMI]

You didn't compare to a well broken-in one. I have one and it doesn't lock up much neither. the WR cubes were not unlubed "out of the box" cubes.

btw it's the cuber not the cube

 

[jayefbe]

I don't think an ao12 is a large enough sample size to make any real conclusions. I could do two ao12s on the same cube, end up with different times and different standard deviations, and make similar conclusions that would ultimately be meaningless. To parse out differences as subtle as what one would find between a modded and unmodded cube would probably require more like 100 solves on each. Not exactly feasible for a short session, I know.

Also, while I understand why you used the same scrambles, it's skewing your results pretty significantly. I've done the same scrambles twice before, not immediately after but during the same session, and it results in a significant decrease in time because I can recall what's coming.

 

[windhero]

I don't think an ao12 is a large enough sample size to make any real conclusions. I could do two ao12s on the same cube, end up with different times and different standard deviations, and make similar conclusions that would ultimately be meaningless. To parse out differences as subtle as what one would find between a modded and unmodded cube would probably require more like 100 solves on each. Not exactly feasible for a short session, I know.

Also, while I understand why you used the same scrambles, it's skewing your results pretty significantly. I've done the same scrambles twice before, not immediately after but during the same session, and it results in a significant decrease in time because I can recall what's coming.

Unfortunately I didnt have the time to make 100 solves and therefore settled with the ao12. There was a clear difference with that amount of solves alone, which is a rough indicator of the efficacy of the mod. I used the same scrambles only with the 3x3 phase, for the full yau-duction I just hand-scrambled all the solves. The real result of the ao100 would could either be of a smaller or a bigger difference, however I dont see it possible for them to be identical. 30% and 42% differences are huge, even a 15% difference would make it worth it to me and most likely many others. I still have an untouched SS 4x4 but it is white and I my recognition with it sucks, so I dont think it would be a fair test if I tried with it. The std deviation could change only to major pauses because of bad recognition.

Also for the florian test I used the modded one with half brights for the primavista scrambles - that basically gives the advantage to the unmodded one (even though I did the 12 scrambles in a completely different order than I did with the first test). However the florian modded had still a much lower std deviation which further proves the point that without the mod the cube locks up and the solves arent as stable.


Also what comes to the well broken in argument, I'd say breaking in can equal a konstamod after quite a while. The konstamod would then play as a shortcut compared to the well broken in cube, I wouldnt say the broken in cube is any better than a konstamodded cube. However breaking in will never make a florian mod effect or atleast not even near to the same extent. I have had a well broken in SS v4 and that is what I base my opinion on.

 

[David1994]

I currently use redux on 4x4 and just learnt how to do 3-2-2-2-3 pairing.
I havent practiced it much but average around the 2 minute mark with it so far.....Should I continue using redux until I get faster or switch to yau method?

Probably an annoying question,but yeah.

 

[BaMiao]

I currently use redux on 4x4 and just learnt how to do 3-2-2-2-3 pairing.
I havent practiced it much but average around the 2 minute mark with it so far.....Should I continue using redux until I get faster or switch to yau method?

Probably an annoying question,but yeah.

I say go ahead and learn yau.

 

[PeelingStickers]

stil using redux ftw >.>

Yeah learn yau.

I need to practice yau.

 

[David1994]

Sorry to keep asking but I have a quick question:

I am colour neutral on 3x3 but I have to think a bit more on 4x4(in particular with yau). anyway do you think I should just become dual CN(i.e yellow and white centers)?
I know feliks is colour neutral but just does white/yellow centers with yau. Thanks

 

[AGazzaz]

Hello,
I Have got a 4x4 cube for 4 days and I can solve it intuitively without any algorithms until I get to the parity point and I can not figure it out, Shall I start learning algorithms now or is it possible to figure the rest out without any help?

Thanks.

 

[kunparekh18]

Hello,
I Have got a 4x4 cube for 4 days and I can solve it intuitively without any algorithms until I get to the parity point and I can not figure it out, Shall I start learning algorithms now or is it possible to figure the rest out without any help?

Thanks.

I would suggest learning an 2 algs - one for OLL parity and one for PLL parity.

OLL Parity - http://www.youtube.com/watch?v=bgRrzlJUat0

PLL Parity - http://www.youtube.com/watch?v=i_eza0B-uFM