Snigel's progression thread (dabbling in bigger cubes)

[Snigel]

I've now spent almost two weeks doing slow solves and thinking more about what I'm doing. I think my F2L is getting better, but not as quickly as I thought. I've also realised that the cross is a major time sink, because I can get close to 20 rather often with a good (easy) cross, but struggle to get under 25 with a bad cross. There should of course be a difference, but not five seconds! I also seem to cube much less when I don't time my solves.

I did around 100 solves yesterday and today, so here are my current PBs:

2023-06-30: Single 17.75, ao5 23.88, ao12 25.16, ao100 28.66, ao500 30.10
2023-07-12: Single 17.75, ao5 23.34, ao12 25.16, ao100 27.08, ao500 29.46

 

[Snigel]

I took a longer break from timed solves as I was on vacation and mostly cubed on flights, busses and trains. I do feel that F2L is getting more efficient, and if I could execute my slow F2L at normal speed, my times should drop by a lot. I guess that's true for most people, but I think I could chop of 6-7 seconds just by being familiar enough with F2L that I can start finding the next pair while solving the current. As I've mentioned earlier, my cross is bad, as bad scrambles take 4-5 seconds longer than good ones, which really shouldn't be the case.

So, current plan for the coming few weeks:

1. Do more timed solves while trying to maintain efficient F2L. An example of this would be to stop rotating so I have the two right slots unsolved. Another would be to randomly cube rotate to find pieces. Consciously trying to find the next pair while solving the current one is also a priority, as this can sometimes take seconds.
2. Do some focused cross work, making sure I can plan out crosses in inspection even for bad cases. As I'm colour neutral, these cases aren't even that bad compared to what you can get if you only do white cross, so this shouldn't be too hard.

I already did some timed solves today and yesterday, so here are the current records:

2023-07-12: Single 17.75, ao5 23.34, ao12 25.16, ao100 27.08, ao500 29.46
2023-07-22: Single 17.75, ao5 22.12, ao12 24.22, ao100 26.61, ao500 28.17

 

[Snigel]

Just doing more timed solves seems to work quite well, so that's what I've done. I've had some long, good streaks, pushing down the longer averages by a lot. Also new single PB and ao12 PB. I'll also switch to ao1000 for the last PB shown.

2023-07-22: Single 17.75, ao5 22.12, ao12 24.22, ao100 26.61, ao500 28.17
2023-07-26: Single 16.86, ao5 22.12, ao12 23.64, ao100 25.75, ao1000 27.95

 

[Snigel]

I think my goal of a sub-20 ao12 is within reach now, with some luck. I don't think it will happen for at least a few weeks yet, but it could! I have had several ao12 with more than one solve under 20 recently, but then the average is spoiled with a few times in the 26-30 range. Still, times are steadily improving!

2023-07-26: Single 16.86, ao5 22.12, ao12 23.64, ao100 25.75, ao1000 27.95
2023-08-03: Single 16.86, ao5 21.12, ao12 22.66, ao100 24.55, ao1000 27.12

 

[Snigel]

No major breakthroughs, but times are gradually improving. First sub-20 ao5, for example, and a new single PB.

The biggest problem seems to be finding the pieces. If I find them quickly, either because I manage to track them or because I spot them fast, I get times below 20 rather often. But then there are solves where I have to really look for pieces doing F2L, somitmes more than once, and it's not unusualy that this takes 10 seconds extra! For example, in this week's forum competition, I got an average 3 seconds above my current ao1000, which is really bad.

Current records:

2023-08-03: Single 16.86, ao5 21.12, ao12 22.66, ao100 24.55, ao1000 27.12
2023-08-10: Single 15.91, ao5 19.64, ao12 21.18, ao100 23.69, ao1000 25.47

 

[TheSpeedSkewber]

I've always wondered - is it pronounced Sni-gull or S-nigh-jul?

 

[Snigel]

I've always wondered - is it pronounced Sni-gull or S-nigh-jul?

[sni:gel], so a front, closed, long vowel! Or you can click the speaker symbol here: https://svenska.se/tre/?sok=snigel. It means "Snail" in Swedish.

 

[Snigel]

I haven't cubed that much recently, but times are still improving slowly. I really think I'm not doing this in an optimal way to get faster (I should go slower and not time solves), but as I think that this will enable me to reach my goal in the fairly near future, I'll try to get a sub-20 ao12 and then consider how to proceed.

If I were to comment on something noteworthy, it would be that getting sub-20 is no longer spectacular. In the previous 100 solves, I had 11 under 20 seconds, but I also had 14 solves over 27 seconds. In the previous 1000 solves, I had 58 solves under 20 seconds and 306 solves over 27 seconds. For an ao12 under 20, I don't think the biggest hurdle is to improve the good solves, but to avoid the bad ones. Sure, I might be able to get a sub-20 ao12 with one 27-second solve, but definitely not with two!

Records as of today, compared to last update:

2023-08-10: Single 15.91, ao5 19.64, ao12 21.18, ao100 23.69, ao1000 25.47
2023-08-22: Single 15.83, ao5 19.64, ao12 21.18, ao100 23.69, ao1000 25.20

 

[Snigel]

I just had an unbelievably smooth solve, resulting in a new personal record with exactly two seconds! And no skip involved either. The funny thing is that I was listening to an audio book while solving and I have basically no memory of the solve solve itself. I only remember the PLL.

Updated records as of today:

2023-08-22: Single 15.83, ao5 19.64, ao12 21.18, ao100 23.69, ao1000 25.20
2023-08-28: Single 13.83, ao5 19.64, ao12 21.11, ao100 22.83, ao1000 24.52

 

[Snigel]

Two notable improvements: I've become a lot more consistent. I still get 25+ occasionally, but it's not unusual to avoid that for long periods of time. In addition, my average times when things are going okay (not failing to find pieces, messing up OLL or things like that) have started creeping down towards 20, but more often in the 21-22 range. This is very clear from the significant ao1000 improvement! I was also very, very close to achieving my goal of a sub-20 ao12, but messed up the last solve. Any day now, though.

Updated records as of today:

2023-08-28: Single 13.83, ao5 19.64, ao12 21.11, ao100 22.83, ao1000 24.52
2023-09-12: Single 13.83, ao5 18.89, ao12 20.16, ao100 22.13, ao1000 23.07

 

[Snigel]

I just got an ao12 of 20.00! This means that I'm as fast with CFOP as I ever was with Petrus, but it also means I didn't get sub-20 with the smallest possible margin. Again.

 

[Snigel]

I finally did it! After being tantalisingly close to getting a sub-20 ao12 for at least a month, I got an ao12 of 19.30 This means that I'm now definitely faster than I've ever been with Petrus.

Updated records as of today:

2023-09-12: Single 13.83, ao5 18.89, ao12 20.16, ao100 22.13, ao1000 23.07
2023-10-02: Single 13.72, ao5 18.18, ao12 19.30, ao100 21.31, ao1000 22.21

 

[Snigel]

Now that I’ve reached my first goal, it’s time to set a new one. I don’t want to set a time-related goal because that encourages me to do timed solves. I haven’t been cubing that much recently, so most of the time, I only do timed solves. That’s not good at all for making long-term progress; I have even forgotten some of the F2L cases I used to know because I don’t practise them enough. Here’s what I’ll do:

• I will work my way through the basic F2L cases I don’t know. Most of F2L is completely intuitive now, and I only have alg-like solutions for the cases when the two pieces are connected in the top layer. Learning the cases where both are already in the correct slot but incorrectly oriented and where only one of them is would help.
• I will focus on looking further into the solve during inspection by giving myself more time. This is especially true for good scrambles, where I tend to start early if I see a good cross, even if just a few seconds more might have allowed me to at least look for the first F2L pair, too.
• I will only do timed solves after I’ve spent a decent chunk of time on the above two activities, and I’ll cap it at 50 solves/day to ensure I don’t overdo it. I will try to record more of those solves for critique and feedback purposes.

I’ll try to stick to this for the rest of the month at least, and then we’ll see!

 

[Snigel]

I've put regular 3x3 on the back burner for now, and I'll focus on OP/M2 for BLD instead. My first goal is to solve a cube using M2/OP with notes and while looking. This shouldn't take very long, and the only reason I haven't succeeded yet is that I'm not used to the lettering scheme and that I make mistakes with some set-up moves.

 

[Snigel]

I did my first successful blindfolded solve using this new method today! It felt much less of an achievement than when I did it the first time ages ago, probably because I wasn't even sure it was possible back then, whereas this time it was more about learning a lettering scheme and some algs.

Some thoughts:

• I struggle with some aspects of memroisation, especially when it comes to pegging letters to words. The first problem is what language to use. I speak Swedish, English and Chinese daily, and think in all three languages. At first, I thought it would be a good idea to stick to one language, especially if I use the phonetic values of the letters rather than the spelling. Some letters have quite different sounds and it's confusing to switch. However, sticking to one language is also frustrating because there might be a perfect word for a specific letter combination in another language. I'm not sure how big of an issue it is to mix languages, but my guess is that it's fine. Eventually, I will be able to convert sticker pairs directly to words without passing through letters anyway. Intuitively connecting a letter pair with a concept seems more important than being consistent with the letter/sound systems.
• It's really bad to change cube orientation. I used to always have white top, blue front, but when I realised that nobody else does this earlier this year, I switched to white top, green front. I'm entirely used to this for normal solving and scrambling by now, but for blindfolded it really messes up tracing. I keep thinking that green means B, blue means F, and I also switch L and R. This means that tracing takes forever, especially for corners.
• The setup moves for M2/OP are so much easier than the old method I used. With the old method, memorisation was fairly easy and most attempts failed because of mistakes in execution, especially with parity, which was a nightmare. The way you solved parity was to bring two corners and two edges to the top layer while observing certain restrictions on what turns were allowed, set up a T-perm (or Y or something else that switches two edges and corners), execute the alg and then reverse. The problem was that in some situations, this required fairly long setup moves and (almost) always messed something up. Solving with parity was ten times harder than without parity. With this new method, execution is a breeze. I can do most of the setup moves intuitively already without having to think too much. Parity does not increase difficulty significantly.

Here's what I'll do next:

• Work on consistency: I still make mistakes often, both in tracing and execution.
• Sort out mnemonics: I need to come up with a more systematic way to deal with the conversion from letter to sound.

The next goal that incorporates both of the above:

• Three successful solves in a row, with each solve taking less than 7 minutes.

 

[Discomantis]

Have you ever been to a comp?

 

[Snigel]

Have you ever been to a comp?

Nope! Maybe I should so I can meet people who know what this is all about. I have a few friends who know how to solve a cube, but that's about it. I'm not too interested in the competition part, though, but as I understand it, that's not the main point with competitions for most people anyway.

 

[Snigel]

Some progress with BLD: I'm getting a bit more consistent and a lot more confident. I still make mistakes, usually with memorisation or tracing, and both are still frustratingly slow. I almost always remember what to do, though, so recall is not an issue. Most of the time is spent making sure I have the right cycles and that I haven't missed an edge.

Here are all timed solves to date (9):

Spoiler: Timed solves

No.​	Time​	Scramble​
1​	DNF(7:29.44)​	U2 R2 D' L' F2 L2 F2 R2 B D2 F D2 R2 F U' L' R D' F' L2 F​
2​	9:06.06​	B' U L' D2 B' L2 F R' F L2 B2 L2 U2 D' L2 F2 U L2 D R2 B2​
3​	9:43.95​	B2 D2 R2 D' F U2 L' U R D2 F2 R2 U2 D2 B2 R F2 B2 R​
4​	DNF(13:14.71)​	L2 B' U2 B L2 B' L2 D2 F U2 R2 F U' F2 R D2 B' U L' D'​
5​	DNF(11:08.42)​	F D2 F R2 B L2 B D2 L2 F R2 F' D L' R B' U L' U B' D2​
6​	DNF(9:38.60)​	U D R L F' D' L D2 L B' U D R2 F2 R2 F2 D B2 R2 D2 L2​
7​	10:28.67​	R' F' D' R' B2 D B U' B2 U2 B2 D2 L' B2 R F2 R' B2 R2 U​
8​	DNF(9:18.68)​	R' B' D' R2 D' R2 L F' D U2 F' L2 B' D2 B' L2 B R2 F D2​
9​	9:23.05​	D' L2 U2 B2 L2 R2 D' B2 L2 D2 B D U2 F' L' D2 B2 R2 D' B2​

 

[Deleted member 36445]

Niceeee

Any method development lately dude?

They were brilliant!

 

[Snigel]

Niceeee

Any method development lately dude?

They were brilliant!

No, none at all, unfortunately. I have been cubing on my commute and occasionally at home, but I've been too busy with life and work to do something serious (including BLD). I don't think I've done a timed solve in a month. I borrowed a smart cube from a friend to see if I could get average splits for normal 3x3, but he forgot to give me the springs and nuts to change the core tensions, with the result that the cube is completely useless (it's so tight it turns worse than any other cube I own that's purchased this decade).

Glad you popped in, though, it makes me motivated to see if I can get some cubing done during the weekend!