Old man (50+) starting to cube - (pønten's progression thread)

[Dan the Beginner]

Thanks for the tip. I will watch it later, as I'm still in bed this early morning, beside a sleeping wife

I do look around on the net for information (videos, articles, forums, streams, etc) whenever I get interested in something. Smartcubes I had mostly managed to steer clear of so far. Not because I don't like it, but on the contrary. I love statistics and numbers. I already know I will like the concept, but trying to restrict myself from too many temptations One of the main motivations for my physical training the last years (running, but lately fast walking due to knee injuries) has been my Garmin watch + extra sensor, collecting all kind of metrics (yeah, a bit of a nerd in that sense, I guess)

Understand, the temptations, the old nerd, and also the wife. When you look at the Cubeast I sent, click on the various buttons for Step, TPS, ... and you can also turn the cube using a mouse to the desired orientation for replay of the solve (animation). I think you will need to download Cubeast first, in order to try it or to see and also share your solve/data (like I did). Cubeast is free, and I particularly like it because it is the only software that supports all smartcubes, so the same common user interface is used whichever cube you should buy next. No re-learning effort, and no need to deal with unreliable or buggy software that some of the smartcube makers provide. And also the ability to compare data collected using different smartcubes, all in the same software (unlike all the different proprietary software, different for each smartcube maker).

 

[pønten]

When you look at the Cubeast I sent, ...

Just had a quick look now (at work, having first coffee). I mistook the link for a video early this morning, which I've made a rule never to click on in bed, as I've woke up my wife a few times too many forgetting to turn off the sound on the mobile first.

 

[pønten]

I only did untimed solves the last two days while passively watching TV etc. But the 25-30 timed solves this morning went well. No single PB, though I had 2 solves in less than half a second slower than the PB. But the averages keep creeping downwards (old PBs in parenthesis):

• Single: - (still 1:05.00)
• Ao5: 1:09.34 (1:16.18)
• Ao12: 1:19.17 (1:21.24)
• Ao50: 1:29.31 (1:42.69)
• Ao100: 1:39.54 (1:58.29)

My first goal when I started to cube, was to beat 1:40 in a single solve, which was (roughly as I remember) the PB I had as a kid. Nice to see that also Ao100 now has beaten that.

 

[pønten]

SUB 1 minute!

Sunday afternoon, not so consistent as during the morning session, but after a PB at about 1:04.xx, I finally managed 57.91

 

[pønten]

I got my second sub 1 minute solve today, and the list of my best solve/averages are now (old in parenthesis):

• Single: 52.25 (57.91)
• Ao5: 1:08.58 (1:09.34)
• Ao12: 1:13.47 (1:19.17)
• Ao50: 1:24.77 (1:29.31)
• Ao100: 1:28.54 (1:39.54)

EDIT: Solved a few more before going to bed, more consistent with no big errors, and made the averages written over a little bit better:

• Ao5: 1:06.00 (-2.5)
• Ao12: 1:09.02 (-4.5)
• Ao50: 1:18.68 (-6)
• Ao100: 1:25.46 (-3)

 

[Ayce]

Soon you'll consider 1 minute a bad time

 

[pønten]

Soon you'll consider 1 minute a bad time

For that to be true for me, maybe it can be translated to "when Ao100 < 1 minute".
I'll make that my current long-term goal, and then we'll see how soon that eventually will happen

I plan to slowly learn the reminder of the CMLL in the weeks to come. I currently only use 15/42 algs, which are the 7 when all corners are in right place, and the 8 when two corners have to be swapped diagonally. From now on I think I will finish each of the types (AS, H, L, O, Pi, S, T, U) fully one by one type. I've already started with O and H, since they only have 2 and 4 variants each, so only missing 3 algs in total. The others have 6 variants each, and after O and H, I plan to finish learning the Pi-algs first.
For the algs I've not learnt yet, I still need to do the Niklas-alg first to swap two adjacent corners, before I can use one of the 7 CMLL algs to turn the corners right way when they are already in right position.

 

[PikachuPlayz_MC]

Congrats! Welcome! Also great choice on cube, I love that one it’s really great if you are learning.

 

[pønten]

I've been doing a few solves every evening this week, both timed and untimed. No new PBs to talk about. I got one 54s-solve (second best I've done) and maybe a second or two on the Ao100, but it is hard to tell, because CubeDesk has been behaving strangely the last couple of days after they did an update on it.

I'm working on my CMLL algs. Until now, when I'm at the point in the solve where I should use the CMLL, I've been doing this:

• Check which corners that have to be swapped (if any)
  1. IF no corners need to be swapped, or two corners need to be swapped diagonally:
     1. Position upper layer according to the type (AS, H, L, etc)
     2. Do the actual CMLL alg
  2. IF two adjacent corners need to be swapped (these are the CMLL-algs I've not learnt yet):
     1. Position upper layer with these two corners on left side
     2. Do the Niklas-alg to swap them
     3. Turn all corners so they have the correct color upwards (yellow or white) by doing the correct CMLL alg according to the type (AS, H, L, etc) they have after performing Niklas.

What I want to do after I know all the CMLL algs, is like this:

1. Position upper layer according to the type (AS, H, L, etc)
2. Check which corners that have to be swapped (if any)
3. Do the correct CMLL alg

So the main difference, is that currently I need to check the corners first, because I might do a Niklas first, and then position the upper layer to do one of the CMLL algs that I already know, but what I want to do later when I know all the CMLL algs, is to position the upper layer directly before checking the corners, because I already know that I can use the right CMLL alg with no need for Niklas any longer.

When slowly implementing/learning the CMLL algs, I need to do a mix of these two methods. That is a bit troublesome. I've now learnt all the H-algs (and the last O-alg), but I've not been able to use the new H-algs in a solve yet, because my brain has not learnt to look for it yet (I go directly to checking which corners that have to swapped, and starting Niklas before I've checked if I had a H-type already (and therefore could skip Niklas).

I need an easy method to quickly see if I have a situation where I can skip Niklas. Therefore I need to quickly also learn the last 4 Pi-algs. Why? Because what Pi and H has in common, is that none of the corners have the right color upwards. So my brain can (hopefully) quickly recognize "no correct corner-color upwards", and go into CMLL mode, by checking whether it is Pi or H, position the upper layer, and start solving for correct corner swapping/turning without using Niklas.

 

[Imsoosm]

Hi ponten! I'm wondering whether you use beginner or advance roux.
Theres a great site called markfiend.com that has a lot of information on roux as well as algs
You can also watch JPerm (a Canadian youtuber), he has videos on beginner and advanced roux.
I'm personally only 13 years old, and I use cfop as main, but over the past few days, i've been trying to learn roux. I haven't found a lot of sites with algs and info on roux, but JPerm's videos explain everything
Roux compared to cfop is actually quite easier to learn because of the amount of algorithms.
I've been cubing for 4 years now and there are still some OLL cases i tend to forget now and then.
Can you post a time split of your steps? Like time your first block, second block, CMLL, and finish separately.
My average with cfop is 13 seconds, trying my best to get to sub 10, but it seems like an impossible goal. Hope we can all beat our own records!

 

[pønten]

Hi ponten! I'm wondering whether you use beginner or advance roux.
Theres a great site called markfiend.com that has a lot of information on roux as well as algs
You can also watch JPerm (a Canadian youtuber), he has videos on beginner and advanced roux.
I'm personally only 13 years old, and I use cfop as main, but over the past few days, i've been trying to learn roux. I haven't found a lot of sites with algs and info on roux, but JPerm's videos explain everything
Roux compared to cfop is actually quite easier to learn because of the amount of algorithms.
I've been cubing for 4 years now and there are still some OLL cases i tend to forget now and then.
Can you post a time split of your steps? Like time your first block, second block, CMLL, and finish separately.
My average with cfop is 13 seconds, trying my best to get to sub 10, but it seems like an impossible goal. Hope we can all beat our own records!

Hi Imsoosm, with 4 years of cubing, you are quite experienced compared to me with only a little bit more than one month of cubing

I suppose with "advance" Roux, you mean using the full 42 algorithms CMLL? I started out with beginner Roux, always yellow up, always red in front (solving first blue, then green block), and only using two algorithms: Niklas (0-2 times) for swapping corner pieces, and Sune (0-3 times) for turning corner pieces with yellow side up.

Now I'm trying to be a little bit more advanced every week. I can choose between yellow/white side up, and I can start with any of the other 4 colors for block-building. Also I'm trying to slowly learn all the 42 CMLL algorithms, but I've only learnt 18 of them until now.
I use the "Algorithm Database" (algdb.net) to find CMLL algorithms, and try to choose the algorithm alternative that seems easiest for me to remember.

Yes, I've searched around on the net quite a bit, and have read markfiend's page, as well as watched JPerm's videos and others. The videos that helped me most as a beginner, was a YouTube-playlist in 8 parts teaching beginner Roux by DeeDubb .

I have no split of my steps yet. I only take the time for the whole solve. My speed is much lower than yours, averaging 1 minute and 25 seconds currently. When I know all the 42 algorithms, I will maybe try to take the time more accurately for the different steps. Either by buying a smartcube, or maybe filming myself.

Good luck to us all to do the best we can!

 

[Imsoosm]

Hi Imsoosm, with 4 years of cubing, you are quite experienced compared to me with only a little bit more than one month of cubing

I suppose with "advance" Roux, you mean using the full 42 algorithms CMLL? I started out with beginner Roux, always yellow up, always red in front (solving first blue, then green block), and only using two algorithms: Niklas (0-2 times) for swapping corner pieces, and Sune (0-3 times) for turning corner pieces with yellow side up.

Now I'm trying to be a little bit more advanced every week. I can choose between yellow/white side up, and I can start with any of the other 4 colors for block-building. Also I'm trying to slowly learn all the 42 CMLL algorithms, but I've only learnt 18 of them until now.
I use the "Algorithm Database" (algdb.net) to find CMLL algorithms, and try to choose the algorithm alternative that seems easiest for me to remember.

Yes, I've searched around on the net quite a bit, and have read markfiend's page, as well as watched JPerm's videos and others. The videos that helped me most as a beginner, was a YouTube-playlist in 8 parts teaching beginner Roux by DeeDubb .

I have no split of my steps yet. I only take the time for the whole solve. My speed is much lower than yours, averaging 1 minute and 25 seconds currently. When I know all the 42 algorithms, I will maybe try to take the time more accurately for the different steps. Either by buying a smartcube, or maybe filming myself.

Good luck to us all to do the best we can!

I use a site called speedcubedb.com that I highly recommend for you to use, I use it for CFOP algs but it also has lots of Roux algorithms. But what I think is best about this site is the reconstructions in it (over 7000!). You can take a look at other Roux users and how they solve it. Just go to the reconstructions page, and check on the box on the left under method that says Roux. This gives you a list of all Roux method reconstructions. Hopefully this will help you improve your times

 

[Foreright]

I've been doing a few solves every evening this week, both timed and untimed. No new PBs to talk about. I got one 54s-solve (second best I've done) and maybe a second or two on the Ao100, but it is hard to tell, because CubeDesk has been behaving strangely the last couple of days after they did an update on it.

I'm working on my CMLL algs. Until now, when I'm at the point in the solve where I should use the CMLL, I've been doing this:

• Check which corners that have to be swapped (if any)
  1. IF no corners need to be swapped, or two corners need to be swapped diagonally:
     1. Position upper layer according to the type (AS, H, L, etc)
     2. Do the actual CMLL alg
  2. IF two adjacent corners need to be swapped (these are the CMLL-algs I've not learnt yet):
     1. Position upper layer with these two corners on left side
     2. Do the Niklas-alg to swap them
     3. Turn all corners so they have the correct color upwards (yellow or white) by doing the correct CMLL alg according to the type (AS, H, L, etc) they have after performing Niklas.

What I want to do after I know all the CMLL algs, is like this:

1. Position upper layer according to the type (AS, H, L, etc)
2. Check which corners that have to be swapped (if any)
3. Do the correct CMLL alg

So the main difference, is that currently I need to check the corners first, because I might do a Niklas first, and then position the upper layer to do one of the CMLL algs that I already know, but what I want to do later when I know all the CMLL algs, is to position the upper layer directly before checking the corners, because I already know that I can use the right CMLL alg with no need for Niklas any longer.

When slowly implementing/learning the CMLL algs, I need to do a mix of these two methods. That is a bit troublesome. I've now learnt all the H-algs (and the last O-alg), but I've not been able to use the new H-algs in a solve yet, because my brain has not learnt to look for it yet (I go directly to checking which corners that have to swapped, and starting Niklas before I've checked if I had a H-type already (and therefore could skip Niklas).

I need an easy method to quickly see if I have a situation where I can skip Niklas. Therefore I need to quickly also learn the last 4 Pi-algs. Why? Because what Pi and H has in common, is that none of the corners have the right color upwards. So my brain can (hopefully) quickly recognize "no correct corner-color upwards", and go into CMLL mode, by checking whether it is Pi or H, position the upper layer, and start solving for correct corner swapping/turning without using Niklas.

Not read through the entire thread (apologies) but just as an aside, 2 look CMLL is good all the way down to at least sub-20 - I’m just above that (last Ao200 is 21) and have just started going through the process of learning full CMLL. I do have a smart cube (Gan iCarry) and have been tracking stats for my last 500 solves. Right now I know the O, H and Pi sets fully to use in solves and I see that learning full CMLL would have saved me about 0.6 seconds on average albeit my “O” algs are pretty quick.

With that in mind, it’s the second block for me that the bulk of the time is spent / wasted. It’s approx 40% of the solve time so most of my time is spent improving that as when I a good second block I get sub 20 times.

it’s worth tracking exactly where you spend the most time to make most efficient use of practice time I’m in my late 40s myself with 3 very young children so don’t get as much time as I’d like for hobbies.

Having said that you’d probably be learning full CMLL at some point so it might as well be now - that’s why I’ve started now anyway - trading a month or two of slower times for better results long term and just because “why not?”

Good luck anyway!

 

[Imsoosm]

Hey Ponten, you can actually do split times on CStimer, by clicking options, go to timer, than make the multiphase 4. After you're done with the first block, press space, and don't stop, just immediately go on to the next step. Repeat that a few times, and you'll end up with a split time of your solve.