Snigel's progression thread (dabbling in bigger cubes)

[Snigel]

Another week of doing almost exclusively BLD! I just barely reached my goal of a sighted ao5 under a minute (59.88). This took a lot more grinding than I expected, especially the last five seconds or so. If nothing goes wrong, I can expect times around 1:00 to 1:10, which is good enough for now, I think.

Sighted M2/OP as of today

• Single: 46.73
• Ao5: 1: 59.88
• Ao12: 1:04.87
• Ao100: 1:09.40

I've experimented with different cubes over the week, too. I think Tornado v3 is my favourite for BLD so far. The slice moves are not great, but okay. Tengyun v2 has great slice moves, but I hate doing dozens of Y-perm on that cube; it seems to catch at least half the time. I also tried the old and trusted R3SM 2020 Maglev, which also works. Increasing the axis distance on the Tornado seems to be best for now.

I also spent more time updating my letter pair memo images to the point where I'm happy with most of them. I still haven't done any full solutions blindfolded, but I have been doing edges and corners separately. My tracing still sucks, and about half of all attempts are DNFs, mostly because I flipped an edge in tracing or messed something up. I still find the I/S W/C switches hard to keep track of.

For fun, I tried using a memory palace to just see if I could memorise three cubes (no tracing, just remembering the letter sequences). It turned out to be trivially easy. I remember all three cubes the day after without reviewing too. Memorisation is clearly not the issue here, tracing is!

Blindfolded M2/OP as of today

Edges only:

• Single: 1:41.06 (1:04.73 memo)
• Ao5: 2:19.67
• Ao12: DNF

Corners only:

• Single: 1:06.22 (38.12 memo)
• Ao5: 1:56.83
• Ao12: DNF

Full solve:

• None yet!

Goals for the coming week:

• Blindfolded ao12 edges only without any DNFs
• Blindfolded ao12 corners only without any DNFs
• Identify errors in all timed solves for one week (reconstruct, identify what went wrong)

 

[Snigel]

More BLD! I had three goals this week:

• Blindfolded ao12 edges only without any DNFs
  Completed! Average: 2:52.70
• Blindfolded ao12 corners only without any DNFs
  Completed! Average 1:43.72
• Identify errors in all timed solves for one week (reconstruct, identify what went wrong)
  Completed!

Regarding the first two, I took my time with some solves to avoid making mistakes, but that's okay as it was the whole point offsetting this goal. I now know that if I'm careful and nothing weird happens (like dropping the cube), I can do blindfolded solves. Great!

Regarding the goal of identifying errors, here are the results:

Blindfolded, corners only

44/52 successful solves. Errors:

1. Tracing, incorrect cycle change
2. Execution, don’t know
3. Execution, dropped the cube
4. Tracing, wrong piece
5. Tracing, wrong piece
6. Tracing, twisted corner
7. Tracing, wrong piece
8. Tracing, twisted corner

By far the most common problem here is simply making basic mistakes when tracing, either by continuing with the wrong corner or the wrong sticker on the right corner. These mistakes should go away with time.

Blindfolded, edges only

47/65 successful solves. Errors:

1. Tracing, flipped piece
2. Tracing, wrong piece
3. Tracing, flipped piece
4. Tracing, flipped piece
5. Execution, skipped pair
6. Tracing, incorrect cycle change
7. Tracing, flipped piece (S/I)
8. Tracing, wrong piece
9. Execution, don’t know
10. Tracing, incorrect cycle change
11. Tracing, missed piece
12. Execution, undo setup move
13. Tracing, missed piece
14. Execution, messed up alg (I)
15. Tracing, missed piece
16. Tracing, flipped piece (I(S)
17. Execution, forgot piece
18. Tracing, missed piece

Almost all errors are related to tracing, so similar to corners. I find edges much harder, first because I sometimes miss pieces (which I can avoid by counting, but this takes time) second because I mess up the /IS and C/W switches when in second position sometimes. I know how it works, but I just plain forget sometimes. Still, most errors are the same type of basic incorrect tracing as with edges.

Blindfolded M2/OP as of today

Edges only:

• Single: 1:24.80 (58.95 memo)
• Ao5: 1:58.70
• Ao12: 2:52.70

Corners only:

• Single: 45.16 (26.45 memo)
• Ao5: 1:27.82
• Ao12: 1:42.72

Full solve:

• None yet!

This means that it's time to turn to full solves! I'm not entirely sure what to expect, but my theoretical best single should be around 2:15, ao5 around 3:30 and ao12 around 4:40, or a bit higher because of parity algs, the general increase in complexity and the lower chance of being lucky.

On the other hand, I also get better at both tracing and execution, so let's see if I can beat the above!

Goals for the coming week

Full solves (M2/OP)

• Single: 2:15
• Ao5: 3:30
• Ao12: 4:40

This seems slightly unrealistic to me, especially the ao12, but I'll give it a go!

I will likely complete my old goal to solve 3/3 in under 7:00, too, which should be easy at this point.

 

[bldcuber314]

You should REALLY start doing full solves, you seem to be doing fine with success rate on edges only. There's only like 8 more letters to memo for corners, which will not make solving much harder. And then you will be getting 3BLD successes.

 

[Snigel]

You should REALLY start doing full solves, you seem to be doing fine with success rate on edges only. There's only like 8 more letters to memo for corners, which will not make solving much harder. And then you will be getting 3BLD successes.

I'll give it a go! Corners seem trivially easy compared to the edges, so I'm pretty optimistic. The reason I didn't do it earlier is that a week ago, I failed maybe 2/3 of edge attempts because of random tracing errors, so doing full solves only made it harder to figure out where I had gone wrong.

 

[Snigel]

I did 100 solves this week, but a fair number were aborted early or obvious failures. I also tried about a dozen where I didn't care about accuracy just to see what would happen. I had 52/100 successful solves overall, and these are my first even with M2/OP.

My goals for this week

Full solves:

• Single: 2:15
• Ao5: 3:30
• Ao12: 4:40

Blindfolded M2/OP as of today

Full solves:

• Single: 2:39.53
• Ao5: 3:26.26
• Ao12: 4:06.73

In general, I'm quite happy with my progression this week. I still feel that tracing and memo are the biggest hurdles, especially figuring out if I'm done tracing or not. It sometimes takes me 30 seconds just to make sure I didn't forget a corner or an edge. For corners, I can usually feel if I have the right number of pairs, but for edges, I have to go through the memo sequence again and count to make sure.

Just to put this in numbers, execution is usually between 60-90 seconds, but memo is usually between 120-180 seconds, roughly twice as long.

Goals for the coming week

I'm not sure what kind of goal I want to set for coming week(s). Should I aim for accuracy, i.e. slow solves which I will nail (almost) every time? Or should I keep trying to increase speed, with less focus on accuracy? Any opinions about what would be most helpful in the long run? Maybe @bldcuber314 or @abunickabhi might have some advice?

 

[Curunír]

I did 100 solves this week, but a fair number were aborted early or obvious failures. I also tried about a dozen where I didn't care about accuracy just to see what would happen. I had 52/100 successful solves overall, and these are my first even with M2/OP.

My goals for this week

Full solves:

• Single: 2:15
• Ao5: 3:30
• Ao12: 4:40

Blindfolded M2/OP as of today

Full solves:

• Single: 2:39.53
• Ao5: 3:26.26
• Ao12: 4:06.73

In general, I'm quite happy with my progression this week. I still feel that tracing and memo are the biggest hurdles, especially figuring out if I'm done tracing or not. It sometimes takes me 30 seconds just to make sure I didn't forget a corner or an edge. For corners, I can usually feel if I have the right number of pairs, but for edges, I have to go through the memo sequence again and count to make sure.

Just to put this in numbers, execution is usually between 60-90 seconds, but memo is usually between 120-180 seconds, roughly twice as long.

Goals for the coming week

I'm not sure what kind of goal I want to set for coming week(s). Should I aim for accuracy, i.e. slow solves which I will nail (almost) every time? Or should I keep trying to increase speed, with less focus on accuracy? Any opinions about what would be most helpful in the long run? Maybe @bldcuber314 or @abunickabhi might have some advice?

Hey @Snigel here's my tip

I would recommend trying to get your accuracy above about 70% then rush until your accuracy dips below about 40%. Then try to keep that pace well getting your accuracy back. Then repeat until your happy with you times.


This was the strategy I was using back when I was averaging 3:00ish and I got my average to sub-50 in under 2 months.





Disclaimer
I am not saying you will necessarily progress this fast.
how fast you progress depends on a lot of things like how much you practise, what method you're using and how fast you can learn new things in general

 

[abunickabhi]

I did 100 solves this week, but a fair number were aborted early or obvious failures. I also tried about a dozen where I didn't care about accuracy just to see what would happen. I had 52/100 successful solves overall, and these are my first even with M2/OP.

My goals for this week

Full solves:

• Single: 2:15
• Ao5: 3:30
• Ao12: 4:40

Blindfolded M2/OP as of today

Full solves:

• Single: 2:39.53
• Ao5: 3:26.26
• Ao12: 4:06.73

In general, I'm quite happy with my progression this week. I still feel that tracing and memo are the biggest hurdles, especially figuring out if I'm done tracing or not. It sometimes takes me 30 seconds just to make sure I didn't forget a corner or an edge. For corners, I can usually feel if I have the right number of pairs, but for edges, I have to go through the memo sequence again and count to make sure.

Just to put this in numbers, execution is usually between 60-90 seconds, but memo is usually between 120-180 seconds, roughly twice as long.

Goals for the coming week

I'm not sure what kind of goal I want to set for coming week(s). Should I aim for accuracy, i.e. slow solves which I will nail (almost) every time? Or should I keep trying to increase speed, with less focus on accuracy? Any opinions about what would be most helpful in the long run? Maybe @bldcuber314 or @abunickabhi might have some advice?

Good goals!

You can try to do slow blind solves,

and fast sighted solves using the BLD execution method.

 

[bldcuber314]

I did 100 solves this week, but a fair number were aborted early or obvious failures. I also tried about a dozen where I didn't care about accuracy just to see what would happen. I had 52/100 successful solves overall, and these are my first even with M2/OP.

My goals for this week

Full solves:

• Single: 2:15
• Ao5: 3:30
• Ao12: 4:40

Blindfolded M2/OP as of today

Full solves:

• Single: 2:39.53
• Ao5: 3:26.26
• Ao12: 4:06.73

In general, I'm quite happy with my progression this week. I still feel that tracing and memo are the biggest hurdles, especially figuring out if I'm done tracing or not. It sometimes takes me 30 seconds just to make sure I didn't forget a corner or an edge. For corners, I can usually feel if I have the right number of pairs, but for edges, I have to go through the memo sequence again and count to make sure.

Just to put this in numbers, execution is usually between 60-90 seconds, but memo is usually between 120-180 seconds, roughly twice as long.

Goals for the coming week

I'm not sure what kind of goal I want to set for coming week(s). Should I aim for accuracy, i.e. slow solves which I will nail (almost) every time? Or should I keep trying to increase speed, with less focus on accuracy? Any opinions about what would be most helpful in the long run? Maybe @bldcuber314 or @abunickabhi might have some advice?

Great progress! For improving your execution, really the main thing at this point is grinding until you know all the setup moves really well and fast TPS. You should have focus on the next letter so that you don't have pauses. As for memo, make sure you use letter pairs(this is the mistake I made until I was averaging like sub 2:30). A lot of people recommend letter pair sheets, although I find it easier to just come up with them on the fly, I can normally do it instantly. Also see how many times you review and try to reduce this as much as you can without dramatically reducing your success rate(I accidentally started one-passing a few weeks ago and my success rate went down to 10%. Yes my BLD progression so far has been full of mistakes.) You should also stick with a certain orientation(unless your buffer is flipped/twisted) so that you can trace faster. I used to use fingers to check if I had covered all of memo, but now I find quickly checking through is quicker, you should try this. Hope this helps!

 

[Snigel]

I would recommend trying to get your accuracy above about 70% then rush until your accuracy dips below about 40%. Then try to keep that pace well getting your accuracy back. Then repeat until your happy with you times.


This was the strategy I was using back when I was averaging 3:00ish and I got my average to sub-50 in under 2 months.

This is almost what I did. When I went for the ao12, my accuracy was over 90%, but after finishing that, I aimed at ao5 and then single solve, and then accuracy didn't matter as much. I think this is a good way to think about it, so maybe alternating between aiming for single solves and ao12 (or more) will have the desired effect. I've used this strategy elsewhere, such as when learning to type fast, and it worked well there. Thank you for the advice!

Good goals!

You can try to do slow blind solves,

and fast sighted solves using the BLD execution method.

Yes, I've done quite a lot of those. I'm pretty stable at around 50-60 seconds. I still make mistakes when trying to go fast, but very rarely if I take an extra 5-10 seconds and don't push it. My problem is tracing and memoing, though.

Great progress! For improving your execution, really the main thing at this point is grinding until you know all the setup moves really well and fast TPS. You should have focus on the next letter so that you don't have pauses. As for memo, make sure you use letter pairs(this is the mistake I made until I was averaging like sub 2:30). A lot of people recommend letter pair sheets, although I find it easier to just come up with them on the fly, I can normally do it instantly. Also see how many times you review and try to reduce this as much as you can without dramatically reducing your success rate(I accidentally started one-passing a few weeks ago and my success rate went down to 10%. Yes my BLD progression so far has been full of mistakes.) You should also stick with a certain orientation(unless your buffer is flipped/twisted) so that you can trace faster. I used to use fingers to check if I had covered all of memo, but now I find quickly checking through is quicker, you should try this. Hope this helps!

I don't think execution is the problem; I almost always mess up because of tracing errors. What do you think is "normal" when it comes to the percentage of time taken up by memo vs. tracing? Right now, it's about 2/3 memo and 1/3 execution for me, on average.

The way I memo right now is using predefined images for all letter pairs. If I find something intuitive, that's okay, but I know the pair-to-image associations well enough that I normally recall them immediately. It would be interesting to see how fast I can memo just the letters, as this would indicate how much of my slowness comes from translating cubies to letters and how much comes from memorising those letters. My guess is that the former is the problem, which has become clear when reading the helpful comments here.

Two things, though:

1. I currently memo edge flips and corner twists using the basic method. For edges, I'm not sure if visual memorisation would be better, as one single image also seems fast and easy to remember. For corners, though, I would greatly speed up execution by using corner twisting algs instead of shuffling through the buffer. Sometimes you get 2-3 twisted corners and it takes 4-6 Y-perms plus set-up moves. That's a lot of turning!
2. I currently only rely on visual mnemonics. I've heard that people use phonetic mnemonics for edges, but I think this doesn't work with execution times as long as mine (around a minute). Do you or anyone else here have any insights to share regarding this? In my experience, phonetic mnemonics don't last very long, so if you solve edges in 15 seconds and repeat the phonetic memo in your head while solving it, that might work (it's probably much faster), but if solving edges takes 40 seconds and I'm not good enough to keep an auditory trace alive in my working memory while also solving the cube, visual memorisation just seems better

General conclusion: I think I'll aim to improve my ao12 and thereby accuracy for a week, then switch to mo3 or even single solves after that, focusing on speed. Thank you all for your advice and encouragement!

 

[abunickabhi]

More BLD! I had three goals this week:

• Blindfolded ao12 edges only without any DNFs
  Completed! Average: 2:52.70
• Blindfolded ao12 corners only without any DNFs
  Completed! Average 1:43.72
• Identify errors in all timed solves for one week (reconstruct, identify what went wrong)
  Completed!

Regarding the first two, I took my time with some solves to avoid making mistakes, but that's okay as it was the whole point offsetting this goal. I now know that if I'm careful and nothing weird happens (like dropping the cube), I can do blindfolded solves. Great!

Regarding the goal of identifying errors, here are the results:

Blindfolded, corners only

44/52 successful solves. Errors:

1. Tracing, incorrect cycle change
2. Execution, don’t know
3. Execution, dropped the cube
4. Tracing, wrong piece
5. Tracing, wrong piece
6. Tracing, twisted corner
7. Tracing, wrong piece
8. Tracing, twisted corner

By far the most common problem here is simply making basic mistakes when tracing, either by continuing with the wrong corner or the wrong sticker on the right corner. These mistakes should go away with time.

Blindfolded, edges only

47/65 successful solves. Errors:

1. Tracing, flipped piece
2. Tracing, wrong piece
3. Tracing, flipped piece
4. Tracing, flipped piece
5. Execution, skipped pair
6. Tracing, incorrect cycle change
7. Tracing, flipped piece (S/I)
8. Tracing, wrong piece
9. Execution, don’t know
10. Tracing, incorrect cycle change
11. Tracing, missed piece
12. Execution, undo setup move
13. Tracing, missed piece
14. Execution, messed up alg (I)
15. Tracing, missed piece
16. Tracing, flipped piece (I(S)
17. Execution, forgot piece
18. Tracing, missed piece

Almost all errors are related to tracing, so similar to corners. I find edges much harder, first because I sometimes miss pieces (which I can avoid by counting, but this takes time) second because I mess up the /IS and C/W switches when in second position sometimes. I know how it works, but I just plain forget sometimes. Still, most errors are the same type of basic incorrect tracing as with edges.

Blindfolded M2/OP as of today

Edges only:

• Single: 1:24.80 (58.95 memo)
• Ao5: 1:58.70
• Ao12: 2:52.70

Corners only:

• Single: 45.16 (26.45 memo)
• Ao5: 1:27.82
• Ao12: 1:42.72

Full solve:

• None yet!

This means that it's time to turn to full solves! I'm not entirely sure what to expect, but my theoretical best single should be around 2:15, ao5 around 3:30 and ao12 around 4:40, or a bit higher because of parity algs, the general increase in complexity and the lower chance of being lucky.

On the other hand, I also get better at both tracing and execution, so let's see if I can beat the above!

Goals for the coming week

Full solves (M2/OP)

• Single: 2:15
• Ao5: 3:30
• Ao12: 4:40

This seems slightly unrealistic to me, especially the ao12, but I'll give it a go!

I will likely complete my old goal to solve 3/3 in under 7:00, too, which should be easy at this point.

Good analysis of the DNFs.

Good luck improving on memo and tracing. Sighted solves using BLD method help iron out these mistakes faster.

 

[Snigel]

I'm not sure if this counts as speedsolving, but while travelling in China, I picked up Moyu Meilong cubes from 2x2 up to 7x7 for cheap. I also bought a Qyi 10x10 just because. I solved all of them (the first time for me), and it was disappointingly easy. The 10x10 was mostly a chore, to be honest. I timed the solves, but most of the time with the 10x10, I was aligning edges and praying that the cube wouldn't explode. I should have gotten a Moyu instead, but it wasn't available when I had the opportunity to buy it.

Now I'm trying to figure out how to best write the Chinese title of my website (揭密中文) using the 10x10 and the 7x7. This is not cubing related as such, but I thought someone might think it interesting. I realised that with three colours, it's possible to cram strokes/components more closely together and still be able to make them out. Here is my current best attempt, although I will fiddle around with it some more before actually putting it on the cubes:

 

[Snigel]

I haven't done any BLD in weeks, but I have dabbled with bigger cubes (as mentioned earlier). I timed my first 100 solves on both 4x4 and 5x5 so that I have some baseline for further exploration. I kind of enjoy both of these more than I thought!

4x4 Yau (after 100 solves):

• Single: 1:35.322
• Ao12: 1:52.71

5x5 Yau (after 100 solves):

• Single: 2:40.60
• Ao12: 4:33.32

I also made one attempt at 6x6, but the cube (MeiLong v2) makes the experience so bad that I will probably never solve that cube again. It took around 20 minutes. I also did one more 7x7, and that cube is much better (also MeiLong v2, so I'm starting to think that their odd-number cubes are better). It still took 25 minutes or so, so I probably won't be putting much effort into that either.