Older cubers discussions

[Jason Green]

I wouldn't either, not just to be strict but more about the nature of BLD. When it comes to memo, anything can be a trigger. Many many times, memo comes around when you open your eyes even without looking anything. But it's just me, too.

That's a good point. I was trying to imagine this happening and was thinking I would feel like I "might" have saw the edge even if I didn't. Maybe this is what I was thinking without knowing it, that opening my eyes could trigger my memory. @chtiger I agree I'd have a PB*. Best advice is just beat it soon and then the worries are over.

I'm glad @Jason Green chimed-in since he doesn't seem to believe in plateaus yet. He rolled past sub 20 and keeps on going.

I have a lot of random thoughts about timers and myself, I'm not sure how useful any of it is but I'll take this as enough excuse to ramble. These days I time a lot of my solves (primarily Stackmat) with less untimed practice, but I don't log/track most of them. I don't personally see much value in tracking an Ao1000 or even Ao100 at my level. I've improved since the oldest times in a big session and I've been watching the trend happen, it tells me nothing new. I also think it would discourage me from working on new things, "I'm not doing that new alg, I don't want the hit to my global average!"

So I basically have two modes of timed solves: sessions where I'm working on things and sessions where I'm actively out to break PRs and evaluate progress. I seem to be in a routine where it's 2-5 weeks between the serious timed sessions with focused practice and spam solving in between. I've developed a decent instinct for when to start timing/logging, several of my PB singles were the first solve after getting the timer out. Sometimes you can just feel it's a good day and maybe a good idea to see if the clock agrees. I can't command it though, like you buckling down in the clutch, so far it happens when it decides to.

For the majority-- the "casual timed solves"-- I ask myself "why" after a slow solve. Slow because I paused five seconds recalling a new alg is ok. Slow because white looked abysmal and I did a rare yellow cross I don't practice, that's ok. Slow because I spent extra time thinking about a better way to deal with a particular F2L pair, perfectly fine. Slow because I'm thinking "I need to be fast" and I out-turn my lookahead, not okay. Improving means ditching some instincts and establishing new ones and that's a "slower before faster" game. So I try (emphasize try) not to get too frustrated with the inevitable regressions as long as it's not a chronic problem that shouldn't be happening and isn't part of a new learning curve. Many parallels with how I've practiced guitar for decades: thinking is slow and I spend more time practicing slower than I do full speed in order to eventually bypass the thinking.

Thanks for the compliment! Although I have to say I feel like I'm plateued now. I've been right around 18 seconds for quite a while now. Of course I practice different things now, something I think a lot of you have always done more than me. The PB single I recently got is a time I will get a few of here and there, it was just very fortunate to get at a comp!

I'm sure many are familiar with the Bruce Lee quote. I never heard this until a cubing post somewhere (probably here). "There are no limits. There are only plateaus, and you must not stay there, you must go beyond them."

I am similar in not caring much about big averages. I haven't even done a good average of 100 or 50 in ages. This is a conscious decision not to make cubing too much work largely. I can get very bogged down with details like this and I'm afraid it would make me burn out, or feel like I "have to" keep doing it to keep cubing. Although the graphs some of you make with thousands of solves are cool, I know mine would basically look the same with maybe slightly different slopes here and there.

It reminds me of my first cell phone. I was about a year out of college. My phone had a call timer that was cumulative minutes and seconds. I think I had 200 minutes a month, but of course every call you made was rounded UP to the next minute. I didn't go to the extent of logging every call, instead I made a spreadsheet that tracked the call timer minutes per month versus the billed minutes. Longer calls would make the ratio closer to 1:1, short calls would tend to affect it more, but I figured on average my call habits would be consistent. The billed minutes were always around 1.7 times what my call timer was. I tracked it for several months and could know by looking at my timer if I was close to going over minutes (not even sure if I could easily check my usage with the provider, but you know I did not want to rely on there possibly delayed data anyway!). In the end it was a fun exercise to play with, but honestly I could have just watched the billed amount for a while and probably gotten a good mental approximation of the same information without crunching numbers. And that's the approach I chose for cubing. After a year I was about X, I've improved about Y%, practicing about Z minutes a day on average. Good enough for me. Not trying to say don't track stats, just saying do whatever makes it the most fun for you.

Wow that was a ramble!

 

[Mike Hughey]

Thanks for all the responses. I like Jason's suggestions about throwing away timer results when using new algs. It might give me a better feel for how I'm really doing.

I've always tended to be on the other side of this - tracking and timing far too much. That probably comes from my getting hooked on the weekly online competition back when I first started on here in 2007. Since I "compete" in most events every week, I don't give myself a chance to go 2 to 5 weeks between timed sessions. Probably part of what has stunted my advancing. But I enjoy the online competitions too much, so I doubt I'll quit. I could still switch to not doing any timed sessions other than the competitions, though, and that would still mostly work for me. I've sort of been doing that with 3x3x3 for the past few weeks; I think I'll keep doing it - I like what I'm seeing on these competition averages!

 

[mark49152]

@Mike Hughey : Yes I also throw away some bad times. I time every session - it's habit, I wouldn't want to miss a PB, and I rely on the timer for scrambles. But I don't want to be deterred from learning and progressing so if I see a chance to use a new alg or something else clever like a fancy parity trick or edge control, etc., I go for it and throw the time away if it all goes wrong. Actually I kind of write the solve off at the moment I decide to go for the new alg, rather than afterwards when I see the time, but if it gives a good result I'll keep the time anyway .

 

[muchacho]

I care too much about timing every solve, I'm pretty sure I should spend more time learning something new and practice it in slow solves... but probably won't do it until I'm sure I've stopped improving.

Mo100 PB: 18.265 (old was 18.307 from June)

 

[Lid]

Got my first magnetic cubes today, instant PB a12 on 4x4 (WuQue) (solves 2-13)
1:02.074, 58.819, 1:09.083, 57.126, 1:05.130, 1:07.080, 1:00.222, 1:02.385, (1:19.951), (56.828), 58.043, 59.357
= 1:01.932 (also last 5 = 59.928 avg)

Testing the 3x3 later

 

[xyzzy]

But I don't want to be deterred from learning and progressing so if I see a chance to use a new alg or something else clever like a fancy parity trick or edge control, etc., I go for it and throw the time away if it all goes wrong.

Sometimes, what I do when I encounter a situation like "I should know this alg, but recalling would take too long" is to just solve it the old/inefficient way, drill the new alg a couple of times, then get back to timed solves. (Doesn't really work when the new thing you want to use is more of a general trick than a memorised algorithm, though.)

Imo this is better since it gets you in the habit of always trying to be consistent in your solves, rather than arbitrarily allowing some of them to be a lot worse than normal.

 

[h2f]

About progress - I've thrown away all care about accuracy in 3bld but I've started to time memo. It gives me motivation to make it shorter and shorter and make as many solves as I can. Result: todays memo mean 21.43 with 14.xx best time. And around 20% accuracy. And I've made over 50 solves. So in this case using timer helps me to keep doing many solves and shorten memo time.

 

[mark49152]

Nice progress Grzegorz! Your memo is getting fast.

 

[Jason Green]

Sometimes, what I do when I encounter a situation like "I should know this alg, but recalling would take too long" is to just solve it the old/inefficient way, drill the new alg a couple of times, then get back to timed solves. (Doesn't really work when the new thing you want to use is more of a general trick than a memorised algorithm, though.)

Imo this is better since it gets you in the habit of always trying to be consistent in your solves, rather than arbitrarily allowing some of them to be a lot worse than normal.

The only problem with that is you don't get the experience of real solve integration. I think that's important. Anthony Brooks said in one of his videos you need to always do the new algs you have learned when they come up in practice also. But I think different things work better for different people, so maybe not as helpful for you.

 

[h2f]

Nice progress Grzegorz! Your memo is getting fast.

A few boys in age 15-22 add me to a chat for blinders in Poland. Their results give me motivation to practice. I have got also view pieces of advice. But the most important is: practice. And your results Mark in big bld and mbld are an ispiration for me. I'm also working on sq1 bld thanks to conversation with @Mike Hughey. Your results Mike are also inspiring. Thanks a lot guys.

 

[Jason Green]

About progress - I've thrown away all care about accuracy in 3bld but I've started to time memo. It gives me motivation to make it shorter and shorter and make as many solves as I can. Result: todays memo mean 21.43 with 14.xx best time. And around 20% accuracy. And I've made over 50 solves. So in this case using timer helps me to keep doing many solves and shorten memo time.

Memo in 21 seconds I can't even imagine that at this point! That's awesome!

 

[pglewis]

I can get very bogged down with details like this and I'm afraid it would make me burn out, or feel like I "have to" keep doing it to keep cubing.
...
Not trying to say don't track stats, just saying do whatever makes it the most fun for you.

Right on the nose. I love stats, and this hobby is ripe for them, but I fear hardcore session tracking would start to feel like a burden for my practice routine over the long haul. When I get busy and have to take brief breaks from practice it disrupts stat continuity. Some days I'm clearly off but I practice with the timer anyway without any anxiety over tainting my long stats. But that's just me, it could be the most enjoyable part for some people and we know you're not doing it wrong if you're having fun.

Since I "compete" in most events every week, I don't give myself a chance to go 2 to 5 weeks between timed sessions.

I still run the timer more often than not nowadays, I just don't normally bother tracking the sessions until it feels like a good time to go bust some records. At around :26 when warmed up right now I'm not all that interested if I drop my Ao12 by a couple tenths, I'm mostly interested in when I can get it down to 24-25 or better. With a few breaks in practice this year I've noticed my brain continues to process the things I've drilled for weeks even without practice. My LL might get a little rusty but after sharpening that back up for a couple days I've found I was noticeably faster even after 3-4 weeks without much practice. I'm sure a lot of that has to do with the fact that there is still a lot of easy time to be had for me, I wouldn't expect that to hold true if I'm sub 15.

I usually have a guess in mind for what my time is before actually looking. In most cases I'm reasonably close but I can tell when fatigue is setting in or I need a break when I notice a trend of "wow, I didn't think it was that slow." On the opposite end I tend to recognize a good groove not so much from a few faster solves but a trend towards faster than my guess. When I'm thinking "too many pauses, that's a 28-29" and then see a 23 I'm likely to open up csTimer and switch to "serious mode". In that mode I'll punt on newer algs and revert to full instinct to see where I'm peaking, basically considering it completely under comp rules and mindset, but it's rare I'm in that mode for more than about 20 solves. I'm either tearing it up, and thus landed the PBs I was after, or it was a false alarm... either way it's usually a short stint. 2-5 weeks just seems to be the rough window between those bigger jumps I'm after, usually quicker when I'm able to spam more at my level, go figure .

... and I rely on the timer for scrambles ...

I know this is somewhat controversial but I still hand-scramble most of the time. I've developed a superstition that reading notation while scrambling leads me to mental fatigue more quickly in a session but that's entirely empirical.

 

[earth2dan]

Is this normal? As I mentioned a little earlier, I've been practicing differently with 3x3x3 - trying to get more efficient with my F2L insertions. I now find that if I time a bunch of solves, and it's not that important to me, I can no longer average better than about 25 seconds. I guess that's because even when timing myself I'm always thinking, "but can I do that better?" Even if I don't actually do it better, I think that just thinking about it is slowing me down. And probably only about 1 in 20 of my single solves are sub-20.

And yet, for the last 4 averages of 5 that I've done in various of the online competitions (ours and some others), I've gotten sub-19 averages of 5. 4 in a row! (18.92, 18.84, 18.44, 18.81) I've gotten very few sub-19 averages of 5, like ever. It seems like when I feel like it really counts, I really have sped up a lot. But I can't see it when I'm practicing.

Is that normal?

I've been stuck at just over 20 seconds for so long now that I've forgotten what it's like to improve at 3x3x3. I've been stuck at just over 20 seconds for probably 5 years now.

I think that's totally normal. I pull down fairly consistent sub-20 averages when it counts, but I mostly get 21-22 averages when timing practice solves that "don't matter". When I'm practicing I'm thinking about improving, so I'm more critical of how I'm executing each step of the solve. For my weekly race averages, I'm on autopilot. Just get it done as fast as possible.

That said; It's been over a year since I've seen any real improvement on 3x3. One of my 2017 goals was to learn full OLL, but I've still got about 20 to go, and I've been lazy. I've even forgotten some of the ones I'd learned earlier this year :/

Though... Winter is coming (Sorry ) and I usually spend a lot more time cubing during the long miserable Winnipeg Winter. Full OLL in 2017 isn't totally off the table yet.

I go for it and throw the time away if it all goes wrong. Actually I kind of write the solve off at the moment I decide to go for the new alg, rather than afterwards when I see the time, but if it gives a good result I'll keep the time anyway .

Haha. I do this exact same thing. I assume it'll be bad, and if it is I delete it. But if it's good... Hey, why not keep it? I earned it

 

[mark49152]

I know this is somewhat controversial but I still hand-scramble most of the time. I've developed a superstition that reading notation while scrambling leads me to mental fatigue more quickly in a session but that's entirely empirical.

Using proper scrambles is a good habit to get into IMHO. It becomes fast and automatic after a while and now I am too impatient to hand scramble. For hand scrambles I have to use so many moves and a couple of tosses, and still don't trust the result, whereas I can very quickly knock out ~20 computer generated moves.

 

[Jason Green]

Using proper scrambles is a good habit to get into IMHO. It becomes fast and automatic after a while and now I am too impatient to hand scramble. For hand scrambles I have to use so many moves and a couple of tosses, and still don't trust the result, whereas I can very quickly knock out ~20 computer generated moves.

I agree hand scrambling seems too hard once you're used to scrambles. I think it was @Marcel P that recommended I use scrambles soon after I started, and I really think it helps you improve your ability to do algs also.

 

[openseas]

About progress - I've thrown away all care about accuracy in 3bld but I've started to time memo. It gives me motivation to make it shorter and shorter and make as many solves as I can. Result: todays memo mean 21.43 with 14.xx best time. And around 20% accuracy. And I've made over 50 solves. So in this case using timer helps me to keep doing many solves and shorten memo time.

Wow, 20s memo!

Last competition, my turn was right after Angelo (3BLD 3rd in US Nats 2017). The judge was one of the first timer cubing parents - I guess he thought what Angelo did was normal After 10s into my memo, he tried to put the paper inbetween my eyes and cube but realized I was still memo'ing, then put the paper back. Funny thing was he tried to do that like every 10s. It could have bothered me but I thought it was really funny during memo. So, all was well

Below picture is a month ago my status of 3BLD. I'm a data guy - so, I record (not filming though) every my session including memo and execution. After the failure, usually review and make a note what was the cause. I used to make a lot of trace error (like tracing flipped or twisted) but now better. As everybody says, pushing memo is the key to improve 3BLD while execution takes practice. So far, my best memo was 49s, long way to go for sub-30s memo.

 

[Jason Green]

Wow, 20s memo!

Last competition, my turn was right after Angelo (3BLD 3rd in US Nats 2017). The judge was one of first timer cubing parents - I guess he thought what Angelo did was normal After 10s into my memo, he tried to put the paper inbetween my eyes and cube and put the paper back. Funny thing was he tried to do that like every 10s. It could have bothered me but I thought it was really funny during memo. So, all was well

Below picture is a month ago my status of 3BLD. I'm a data guy - so, I record (not filming though) every my session including memo and execution. After the failure, usually review and make a note what was the cause. I used to make a lot of trace error (like tracing flipped or twisted) but now better. As everybody says, pushing memo is the key to improve 3BLD while execution takes practice. So far, my best memo was 49s, long way to go for sub-30s memo.

The charts are cool! I kind of asked about this on FB but don't think my question was clear. How do you chart C-Trace and C-Memo? I think you said you memo while tracing so I don't understand the two colors on the bar chart.

 

[openseas]

The charts are cool! I kind of asked about this on FB but don't think my question was clear. How do you chart C-Trace and C-Memo? I think you said you memo while tracing so I don't understand the two colors on the bar chart.

Memo fail is when I forgot some letter pairs or completely forgot what memo was. For example,(edge is easier to explain), edge memo was AB CD JL NO HG W but I forgot one of letter pair (like skip like HG). So, I know my audio memo toward the end was not good enough or audio sound was fuzzy.

The trace fail is when I traced wrong - instead of UR piece, if I memo'd as RU, then, I ended up with a flipped edge. In this case, I executed what I memorized but just target was wrong.

Not sure whether it makes sense but I tried to breakdown the source of error (which is what I do for a living )

 

[Jason Green]

Memo fail is when I forgot some letter pairs or completely forgot what memo was. For example,(edge is easier to explain), edge memo was AB CD JL NO HG W but I forgot one of letter pair (like skip like HG). So, I know my audio memo toward the end was not good enough or audio sound was fuzzy.

The trace fail is when I traced wrong - instead of UR piece, if I memo'd as RU, then, I ended up with a flipped edge. In this case, I executed what I memorized but just target was wrong.

Not sure whether it makes sense but I tried to breakdown the source of error (which is what I do for a living )

Oh so the chart on the right is only the reason for failures!? I thought it was a breakdown of what percentage of time each part of a solve was, including successes. I see now (I think).

 

[Lid]

Nine year to reach this milestone ... sub1 a12 on 4x4x4

{(52.418), (1:09.054), 56.215, 55.510, 1:00.173}, 1:02.393, 56.469, 1:04.942, 58.282, 55.980, 1:01.594, 1:08.021
= 59.958
also includes a new {a5} PB of 57.299

Spoiler: scambles

Average of 12: 59.958
1. (52.418) U' r' B2 U2 F' B2 D' r D' B' U2 L' D2 f' F' U f' B F2 U' B2 U' F2 r F L U2 r' F D2 r' U B2 L2 F2 f' D2 L' F r'
2. (1:09.054) U2 L2 F L' R r2 F' L2 F L2 f2 r u2 B' u r' D' u' f B' r2 D2 r' L2 f R U L D2 B2 u2 L f D F2 f2 u' D2 B u
3. 56.215 r2 F r2 R B' u' U2 B u' F' D' L2 f' R2 F' D' u U f' D F2 u f' D2 B2 F' D2 f u U B f2 D r2 R' L B2 F2 u2 r
4. 55.510 R D2 f' F' r' B' R u F' R B f D R F2 D' r2 R2 D r' R2 L B' u2 D2 r2 R2 u' R' F' f r2 F2 r' D F U' R r' B'
5. 1:00.173 R' F u' U r F2 R r2 U r2 u2 L2 B2 u2 F' D' r2 u2 D2 L R U2 f' U2 u2 f' R' u2 r2 f' L' r R' D2 u2 R2 F B L2 U'
6. 1:02.393 L U u2 R2 L' f2 D R2 F2 R' F f r2 D R' U r R u2 B f' r2 F R2 u L' f L' D F D B' f' F' L2 r2 f2 R U2 B'
7. 56.469 L R' B u2 R' D2 L D f' F r2 f2 L F L2 r R2 D' B2 F r2 B R2 B2 U2 f B2 F' L r2 u R' B L' F2 B' U2 u L B
8. 1:04.942 L' f R2 u2 L' R D R2 D2 r u2 L2 D' u2 F2 f' r2 B' r' D R f2 U F2 u f' B u' D2 B' U' u R' F U' B' u F2 B R'
9. 58.282 U R2 U' u' F f2 D2 B D' F' D' L R' f' B2 u F2 R B2 F' L2 B2 f U2 F2 L' F2 u' D' U' L B2 D' F2 U f D2 u R' f'
10. 55.980 F' R r2 f' R' u' r U' r D2 u' B2 u F' U2 B R2 B R L u' r u B' u f2 D2 F B r2 B D B L2 f U2 f F2 U' R'
11. 1:01.594 D U r2 D' U2 R2 B2 f2 u2 B2 D2 r' L f' D2 U2 F2 L' r' R' B f' L' f' r2 D' R u2 r U2 f' L u2 f' B2 r2 L2 R2 B' R
12. 1:08.021 L R2 D R2 U' R2 B2 u2 B' f' u2 D' f' D' u' L2 D L' r U' u2 L2 D f u2 F2 L2 u2 f' B' U2 u B u' f' u' U' r2 F L'