Muscle memory

[Llewelys]

Warning: This post won’t end the way you think it will (not clickbait)

My goal for the next half of 2019 is *drum rolls* to get sub 20.

I just did a 22.46 ao12 and here are the splits:
Cross/F2L/LL : 2.80 / 12.10 / 7.55

I know I can shave off up to 1.5-2 seconds of the LL stage by working on PLL recognition (I use 3LLL), but I want to focus on getting F2L down to sub 10 seconds first.

Now, before you hit me with urghh n00bs asking how to be sub20 when we've already talked about this at length, hear me out:

I've read all 30 pages of the "How to be sub 20" thread but the slow down and look ahead advice is getting old.
Here's the thing: to look ahead you need to look away from the pair you're currently solving, meaning you need to rely on muscle memory.
But HOW do you actually develop muscle memory ?

Here's what I've tried so far:
1. Just practice. Do nothing but solves, hoping repetition will eventually get those F2L cases into muscle memory.
Pros: practicing is always good
Cons: too many F2L cases so that's not very efficient

2. Blind solving. Look at a case, close your eyes, solve the case. I've done that a lot and can now solve pretty much any case with my eyes closed. Here's the thing though: when I have my eyes closed I still picture the pieces moving in my head, so I'm still focused on the pair I'm solving.
Pros: useful to know for sure if the algs you use are correctly in your muscle memory
Cons: only useful once the algs are already in your muscle memory

3. Drilling algorithms, just like with OLL and PLL.
Pros: Gets the job done (hurray \o/)
Cons: too many algs to drill (and the alg changes depending on which slot the pair needs to go in so that multiplies the number of algs by almost 4) so it's discouraging. And it's b o r i n g

4. During a solve (untimed), look at a case, solve it as fast as you can, pause to recognise the next case, repeat.
Pros: quite fun
Cons: I can't do the alg fast if I don't already have it in my muscle memory


I looked up threads on muscle memory but all they keep saying is to just practice, but I don't really find that useful (I've done thousands of solves already)

Point 3. was actually helpful for cases where both the corner and its edge are in the top layer (algs like R U2' R' U' R U R' and F (R U R' U') F' (U R U' R') are "fun" to drill) but I can't for the life of me seem to make it work for cases where either the corner or the edge is in a wrong slot.


Has anyone encountered the same problem / has consciously worked on muscle memory ?
Should I persevere with points 3. and 4. ? (The most useful points imo)
Am I missing something ? Am I just dumb ?

TLDR; How do I get F2L algs into muscle memory ? Please don't say Just practice.

------------

"Edit": I actually wrote what you just read a few days ago on my phone but didn't post it right away because I wanted to think about it a bit more.

I think I just found a solution to my problem (I'm sure a lot of people have discovered this before me but oh well, nevermind):

The previous points suck on their own but here comes ... point 5. !!! *applause from the crowd*

5. A mix of the previous points. You're gonna need two cubes. Solve on the first one using point 4.. If you encounter a case you don't have in your muscle memory yet, set it up on the second cube and drill it using point 3.. Drill it 5 or 50 times or any amount in between depending on how often the case comes up and how much it screws your solves when it does (you should have a pretty accurate idea of that thanks to point 1.). Then, you do that case on the 1st cube blindfolded (remember point 2.?) to check whether you still need to drill it. Continue your solve and repeat.



What do you guys think about this ?
Did this post go from "Please help me" to "Here’s a cool way to develop muscle memory" ?

 

[Tyler Fresh]

Getting things in muscle memory is just repitition. So when people say "just practice" they are correct. It isn't going to happen overnight. It may take up to 1000 solves but it will eventually happen.

 

[Llewelys]

I've done like 5000 solves over the last 10 months though (not a random number, I keep track of every ao100 I do to see my progress)

 

[Tyler Fresh]

I've done like 5000 solves over the last 10 months though (not a random number, I keep track of every ao100 I do to see my progress)

You have to be practicing very inefficiently, after 5k solves you aren't sub 20.

 

[Llewelys]

Well yeah that's the point of this thread, I'm trying to find efficient ways to practice.
That's why I said that the "Just practice" advice was useless

 

[Tyler Fresh]

Well "practicing" muscle memory is inefficient practice. That will come with time. Working on lookahead and more efficient f2l is what you should be doing.

 

[Llewelys]

Working on lookahead [...] is what you should be doing

........ Have you read the first message at all ??? This whole work needs to be done before working on look ahead

Here, I'm gonna quote myself for you:

to look ahead you need to look away from the pair you're currently solving, meaning you need to rely on muscle memory

To look ahead you need to have your F2L algs so ingrained in your muscle memory that you don't even THINK about the pieces moving and where they are to solve the case. I'm not at this stage yet, hence this thread.


Regarding f2l efficiency I 100% agree, but I think it's already pretty efficient for my level. I'll let you be the judge of that:
Here are reconstructions from a few solves I did in the ao12 I'm talking about in my first message:

Spoiler: 20.42 solve

B2 L2 F2 L2 D' U2 B2 U2 L2 D' L F2 U' F L R B R U2 R

z2 y’ // Inspection
L’ R’ (F B’) D // Cross
y U’ R’ U R // F2L 1
L U L’ U2’ L’ U L // F2L 2
U L U L’ U L U’ L’ // F2L 3
U’ F (R U R’ U’) F’ U R U’ R’ // F2L 4
U F (U R U’ R’) F // OLL
G perm U2 // PLL

Spoiler: 21.41 solve

U R' D' F' L B D R' F' D' L2 F2 L2 D2 L2 U B2 R2 B2 L2 U’

Z2 y // Inspection
L U2’ R’ (F B’) R’ D B’ // cross
U L’ U L U R’ U’ R // F2L 1
U’ (L F’ L’ F) L U L’ // F2L 2
(R U R’ U’) (L’ U’ L U) y’ R’ U’ R // F2L 3
U’ R U R’ U R U R’ // F2L 4
U’ F (U R U’ R’) F’ U x R’ U’ L U R U’ L’ U x’ // OLL
U2 Jperm U2 // PLL

Spoiler: 26.45 solve (worst solve of the ao12)

B2 F2 D2 R2 U' F2 R2 U' R2 B2 R2 B R F' L D2 F' U' L' U B'

z2 y’ // inspection
D’ U2 R’ D’ (F B’) R2 D’ // cross
U2 L U L’ U’ L’ U2 L U L’ U’ L // F2L 1 (made a mistake here -- wrong set up to pair up corner + edge, should've done y U2 R' U' R L U' L')
y R U’ R’ // F2L 2 (another mistake, should've done a y' instead of a y)
U L’ U’ LU’ L’ U’ L // F2L 3
U2 f R f’ U’ R’ U R // F2L 4
f (R U R' U') f' U2 R2 D R' U2 R D' R' U2 R'// OLL
Fperm // PLL

 

[Mike Hughey]

I'm curious about your age. I have found that being older seems to make muscle memory take much more slowly than when young. There are people who seem to have less problem with that, but overall I think it's what causes most older cubers to take much longer to progress than younger ones. Of course, there are probably exceptions on both ends of the spectrum, so even if you're young, perhaps you have trouble just because you're one of the unlucky ones.

I have done many tens of thousands of 3x3x3 solves, and yet only just recently got where I think I can say I'm actually averaging just barely sub-20. By that I mean that about half of my averages are sub-20, and the other half are 20 seconds or more. And I know I'm not the only older cuber who has taken that long to progress. I find that I still fail at algorithms that should be committed to muscle memory on occasion, and it takes me a very long time to add a new algorithm, so I rarely ever bother to learn new ones anymore. (I find it so much more fulfilling to work on blindfold solving instead.)

I personally really like your suggestion for how to handle getting muscle memory for these cases. It seems like a very good solution. I look forward to hearing how well it works for you; perhaps I'll even try it myself if I'm ever ambitious enough to try to learn to use new cases. I did something somewhat similar recently when I did add a few algorithms for inserting pairs in the back (actually I simply learned to set up the case so I could repeat it over and over again).

 

[Llewelys]

I'm 24 so I don't know if you consider that old or not haha (it certainly is older than the majority of cubers)
I hadn't thought of age as a factor here, but maybe you're right.

Thanks for your feedback! I'll let you know how it goes

 

[Mike Hughey]

Most people who start around 24 seem to still improve fairly fast (although generally not as fast as the kids). So yeah, you're somewhere in between.

 

[icarneiro]

I'm 24 too, I started cubing in 2019

and I'm jealous of kids who learn super fast hahahah