Older cubers discussions

Luckily one look cmll though I have a problem with bowtie.

I've learnt my L, U and T COLLs now... well, almost... I still have two of the T's to learn.

I think L cases are trickier to recognise than U and T but I'm sure they will improve with practice. :)

Edit: I think COLL recognition may become faster than PLL recognition, eventually.
 
That explains a lot to my untrained eye, thanks! Anyway, your times are quite good for 2L CMLL. Any plans on learning full CMLL?
More than half of my solves are with 1-look CMLL. I'm like 12 algs away from full CMLL, I was learning the U set but stopped a couple of weeks ago, also sune and antisune cases left.

@h2f How do you recognise L/Bowtie cases? From the position showing in algdb.net? I do an [y] to that, and I think waffle's are an [y'] away from algdb.net, but I don't know which one it's easier, I haven't tried those.
 
Welcome @One Wheel :)


I've filmed some solves (some lucky ones, I'm averaging 26.x), they look so different that Grzegorz's :confused:

Wow, nice dude! You are much smoother at Roux than I will ever be. I think slowing down and reducing the huge pausing (every now and then) will make your times so much better. Your L6E is flawless.
 
Think of the lead singer of Wham, and then switch first with last name. Now you will never forget. :)

:) Thank you.
@h2f How do you recognise L/Bowtie cases? From the position showing in algdb.net? I do an [y] to that, and I think waffle's are an [y'] away from algdb.net, but I don't know which one it's easier, I haven't tried those.

I do it the same like on waffo's page. I think they are with y' rotation to algdb.net. I think they are quite easy to recon from this angle. I was affraid of this set but it become much easier than I thought.

I've filmed some solves (some lucky ones, I'm averaging 26.x), they look so different that Grzegorz's :confused:


Wow. Very nice solves. With one-look and no puases sub-20 possible. I'm averaging around 28 right now.

Here is a solution I'm glad. Unluckily with Z perm in the end.
B2 U2 B' D2 B' U2 L2 B' R2 B2 U2 R B U2 R B' R' B U' L R2

z y2
R' U' R' F' u' F' //6/6
R2 r U r2 U' R' U R r U' r' // 11/17
R U2 R D R' U2 R D' R2 //cmll
U2 M U' M' U' M' U2 M' U M2 U' M' U2 M2 U2 M' // 42/42

alg.cubing.net
 
@muchacho: I agree with Marcel that your L6E looks awesome. Apart from full CMLL, which it looks like you're close to finishing up with the algs, how is your inspection working? Can you plan the entire FB? And how's your progress trend lately? I'm sure you still have room for improvement beyond sub-20. There are times where I wish I had learned Roux instead of CFOP in the first place. Have you seen some videos of Roux solves by this guy in the forum @PenguinsDontFly ? It's like I will do this and this...and ops! cube solved. :) Can it be that Roux has potential to be even faster than CFOP?
 
Yeah my story is something like that too :).

Here's my highlight solves from North London. Of course my best highlight was the 1:27 3BLD, but I have no footage of that due to untimely battery death.
Those 18's where really cool. I notice quite big pauses and you still manage to get full steps sub 19's. How do you do that? I think your LL is faster than most of us.
 
Wham? Wow this really is the older cubers' thread :). I wonder did anyone here used to have a hairstyle like Wham?
I was to young and later I had long hair.

@muchacho: I agree with Marcel that your L6E looks awesome. Apart from full CMLL, which it looks like you're close to finishing up with the algs, how is your inspection working? Can you plan the entire FB? And how's your progress trend lately? I'm sure you still have room for improvement beyond sub-20. There are times where I wish I had learned Roux instead of CFOP in the first place. Have you seen some videos of Roux solves by this guy in the forum @PenguinsDontFly ? It's like I will do this and this...and ops! cube solved. :) Can it be that Roux has potential to be even faster than CFOP?

I like Guroux's walkthroughs, Alex Lau's and PDF's too. There's no prove if Roux is faster or slower. Too few people use it and that's why it's hard to compare both methods looking at offcial ao5s. But the best ao5 with Roux was made by Alex Lau and was ER for some time and now he's 6th in the world.
 
Hi Jason. I uploaded the Python scripts (one to download the DB and build indices, one to generate the rankings) to GitHub and the link is in the original post.

Click the ipynb files to view - https://github.com/Logiqx/wca-ipy
Sorry I'm a little slow, I guess because I do more managing nowadays than developing :/
How do I download these to execute? Do I have to copy/paste the little code snippets together into a working script? If so could you upload the entire thing, or help me know how to download it?

Also, should IDLE on Windows work to execute it? I can see that I will need MySql but that should be easy to install, I have used it before.

BTW, welcome @One Wheel!
 
Sorry I'm a little slow, I guess because I do more managing nowadays than developing :/
How do I download these to execute? Do I have to copy/paste the little code snippets together into a working script? If so could you upload the entire thing, or help me know how to download it?

Also, should IDLE on Windows work to execute it? I can see that I will need MySql but that should be easy to install, I have used it before.

BTW, welcome @One Wheel!

They are iPython notebooks which means they are actual code with some surrounding markup to make them look like example code / tutorials. What you see on GitHub is the actual working code.

The easiest thing to do is install Anaconda then you can load and run my iPython notebooks. Have a look at the iPython website as they have a nice video showing how iPython works.

You also need MySQL installed and an empty database for the WCA data. Once you have Anaconda and MySQL it should be pretty easy to get everything to run.

One last thing. The scripts generate bbCode for the forum as opposed to HTML. They'll need some minor tweaking to generate HTML for a normal website.
 
They are iPython notebooks which means they are actual code with some surrounding markup to make them look like example code / tutorials. What you see on GitHub is the actual working code.

The easiest thing to do is install Anaconda then you can load and run my iPython notebooks. Have a look at the iPython website as they have a nice video showing how iPython works.

You also need MySQL installed and an empty database for the WCA data. Once you have Anaconda and MySQL it should be pretty easy to get everything to run.

One last thing. The scripts generate bbCode for the forum as opposed to HTML. They'll need some minor tweaking to generate HTML for a normal website.
Thanks I may play with it!
 
Those 18's where really cool. I notice quite big pauses and you still manage to get full steps sub 19's. How do you do that? I think your LL is faster than most of us.
As you know, I never videoed my solves until recently. It came as quite a shock to see how much I paused, even when it feels like a fast solve, because it doesn't feel like I pause much. I guess the solution is more slow turning practice, although I'm not a fan of that. I think it has helped my 3x3 to do more 4x4 and 5x5 because those puzzles force slower more careful turning. I might try doing a couple of weeks of 3x3 practice on a 4x4 cube and see what happens.

Regarding 4x4, Joey Gouly judged one of my solves and observed that I had a 27 second 3x3 stage in a 1:07 solve, with no parity. That's way too much. I checked my other solves and my 3x3 stage is consistently slow. So 3x3 on 4x4 practice might conceivably help with that too.
 
@Logiqx I was making pretty good progress with the scripts until our internet went out! I think I had a newer version of Python so urllib2 was in urllib.request but I got past that. Then I had some pathing issues I think I got straight. The only thing left I think is to figure out why the pymysql import is missing. I should be able to when my PC is back online. And then if I want to make changes for HTML.

Thanks though it's fun to play with!
 
Last edited:
I've learnt my L, U and T COLLs now... well, almost... I still have two of the T's to learn.

I think L cases are trickier to recognise than U and T but I'm sure they will improve with practice. :)

Edit: I think COLL recognition may become faster than PLL recognition, eventually.

I've forgot to mention that in my sheet 4 of 6 of them are comms. The first is of course only twister and only Yswitch is not comm and COLL.
 
@muchacho: I agree with Marcel that your L6E looks awesome. Apart from full CMLL, which it looks like you're close to finishing up with the algs, how is your inspection working? Can you plan the entire FB? And how's your progress trend lately? I'm sure you still have room for improvement beyond sub-20. There are times where I wish I had learned Roux instead of CFOP in the first place. Have you seen some videos of Roux solves by this guy in the forum @PenguinsDontFly ? It's like I will do this and this...and ops! cube solved. :) Can it be that Roux has potential to be even faster than CFOP?
My use of inspection time is pathetic, my worst part. Most of the time I can do a 2x2 block (not in an efficient way) and then while solving maybe follow the other 2 pieces. I've seen those videos, but just that, they looked useful but didn't pay much attention then, I'll get back to them. I was sub-24 before changing color scheme, I was improving lately mainly thanks to lookahead. The change affected mainly to my blocks, maybe some to CMLL recognition and almost nothing to LSE, but with the new cube I still get some catches during CMLL and LSE is not as smooth. I'll just do solves and learn and practice the other U cases until I'm again sub-24 (I'm 26-27 now), then I'll try to practice my FB.

A graph with my first 15000 solves and another without the firsts 1000.
 

Attachments

  • timeline-15000-a.png
    timeline-15000-a.png
    30.7 KB · Views: 7
  • timeline-15000-b.png
    timeline-15000-b.png
    68.7 KB · Views: 7
Last edited:
Back
Top