Road to Sub-30: What Ten Years of Getting Stuck Taught Me About Getting Faster
By: Dhakshit Gona
The first person I ever saw solve a Rubik’s Cube quickly did it in about 20 seconds. I was six or seven, and until then, the only cubes I really knew were the stiff Rubik’s-brand ones you would find in a toy aisle. Then a family friend came over with a GAN, picked it up, and solved it before I could really process what he was doing.
I thought he might be the smartest person I had ever met.
Naturally, I tried to copy him. My strategy was to solve one side, move on to another, and then wonder why fixing the second side destroyed the first one. Eventually, my parents got me a cheap generic speed cube from Amazon, and I started watching tutorials, constantly pausing and rewinding while trying to understand notation that felt like another language.
My first solves took somewhere around 10 minutes. Nearly 10 years later, I hover around the 30-second mark, with a PB in the low 20s. I can break 30, just not as consistently as I want yet, and I still use 2-look OLL and PLL while learning more cases.
Looking back, every big-time drop came from fixing a different problem. What got me from 10 minutes to two minutes did almost nothing to get me from one minute toward 30.
That is probably the biggest thing I wish I understood earlier: when you hit a plateau, doing the same thing faster usually isn't enough.
10 Minutes to Two Minutes: Get Comfortable First
My first improvements were pretty simple. Once the beginner method became familiar, I stopped spending forever trying to remember what came next, learned basic finger tricks, regripped less, and became more accurate with my turns. Those changes alone took minutes off my solves.
If you are still taking several minutes, I honestly wouldn't rush into memorizing dozens of OLLs. Get comfortable with what you already know first. Your method should feel automatic enough that your brain isn't constantly stopping to ask what step comes next.
There is also a big difference between turning quickly and solving quickly. At the beginning, they feel like the same thing because better finger tricks and smoother turning immediately lower your times.
Eventually, though, they stop being the same thing. For me, that happened around two minutes.
When More Solves Stopped Working
Two minutes were my first real plateau. I knew the beginner method; I could turn reasonably quickly, and yet my times were no longer dropping the way they had before.
Around the same time, my family moved, and I stopped cubing for somewhere between six months and a year. There wasn't some dramatic reason I quit. I moved, forgot about it for a while, and eventually picked the cube back up again.
I still loved it.
When I came back, I eventually started realizing that doing more solves was not automatically the same as practicing. You can do 100 solves with the same bad habits and become incredibly consistent with those bad habits.
If you are stuck, the useful question isn't just, “How do I go faster?” It is, “Where am I actually losing time?”
Maybe your cross takes forever. Maybe you stare at the cube between every F2L pair. Maybe you know your last-layer algorithms but hesitate before recognizing them.
The timer tells you that you're slow. It doesn't tell you why.
Then F2L Humbled Me
Improving my last layer eventually helped me reach roughly one minute. Then I discovered CFOP and figured I had finally found the thing that would make me fast.
Cross. F2L. OLL. PLL.
Then I tried F2L, and it was awful.
With the beginner method, I always knew what I was looking for. F2L meant finding a corner, finding its matching edge, understanding their relationship, pairing them, and inserting them correctly. My brain did not want to process all of that in real time.
For a while, learning a faster method actually made me slower.
If you just switched to F2L and your average suddenly looks terrible, I completely get it. You replaced something automatic with something your brain barely understands yet. That uncomfortable stage is part of the process.
The real breakthrough came when I stopped seeing everything as individual moves.
When Algorithms Started Looking Like Patterns
While learning 2-look OLL and PLL, I initially treated every algorithm like a password. I would memorize a long string of notation, repeat it until I thought I knew it, then somehow forget half of it the next day.
Eventually, I started noticing familiar pieces inside those algorithms: sexy moves, sledges, insertions, taking pairs out and putting them back differently. Instead of seeing 10 or 12 unrelated moves, I started seeing a few patterns connected together.
That changed the way I approached the whole cube.
Some of my favorite memories from cubing came around the same time. I taught my dad and brother how to solve, and we would sit in the living room late at night, pausing the same YouTube tutorial what felt like a hundred times while trying to learn an algorithm. My mom would eventually start yelling at us to go to sleep.
My dad and brother got to around two-minute solves before they stopped cubing regularly, but those nights helped something click for me: you eventually have to understand the cube, not just know how to finish it.
For me, understanding means recognizing familiar triggers, seeing why an F2L pair moves the way it does, and slowly getting better at predicting where pieces will end up.
I am still working on that.
Where I Actually Lose Time Now
Getting from roughly one minute to 30 seconds has been less about discovering huge tricks and more about removing small pauses.
Cross is still one of my weaker areas. I used to treat inspection like the annoying countdown before I was allowed to start solving; now I try to actually plan as much of the cross as I can before touching the cube. A simple drill is to scramble, inspect, solve only the cross, stop, and repeat.
F2L is probably an even bigger one. A solve can look fast because your hands are moving quickly, but if you solve a pair, stop for two seconds, find another pair, turn quickly again, and then stop, those pauses add up fast.
That is why slowing down can sometimes make you faster.
Try turning at a speed where your eyes have enough time to start finding the next pair before your hands finish the current one. Four fast F2L pairs separated by long searches can easily be slower than four calmer pairs that flow together.
And if you are worried because you don't know full OLL or PLL yet, don't panic. I am around 30 seconds and still use 2-look for both.
More algorithms will help, and I am learning them now, but another PLL isn't going to fix a five-second cross or massive F2L pauses.
Practice the Problem
For years, most of my practice looked like this: scramble, solve, check the timer, repeat until I got annoyed.
PB hunting is fun. It isn't always good practice.
Now I think a session is much more useful when it has one goal. If your cross is bad, drill crosses. If your F2L pauses are the problem, do slower solves and focus on keeping your eyes moving. If you're learning a PLL, practice that PLL until the recognition and execution feel comfortable.
Then go back to normal solves and see whether anything has changed.
Trying to improve cross, F2L, lookahead, TPS, OLL, and PLL all at once usually means you're not really focusing on any of them.
Find the thing costing you the most time and work on that first.
What 20 Seconds Means Now
After my generic Amazon cube, my first real speed cube was a MoYu RS3M 2020 that my mom gave me after I found it through J Perm. The difference felt insane. It was smoother and faster, and it made me want to solve it constantly.
I still love trying new cubes and lubes, but unfortunately, no cube has figured out how to look ahead during F2L for me yet.
I also still know the family friend whose 20-second solve started this whole thing, although we're not nearly as close anymore. He has never seen my fastest solve, so there hasn't been a perfect full-circle moment where I sit across from him and finally match the time that got me hooked.
Maybe one day.
When I was six, though, I assumed his hands were what made him so fast. Almost 10 years later, I understand that the bigger difference was what he could see in the cube that I couldn't yet.
That's still what I'm trying to improve.
I'm around the sub-30 line now, sometimes under it and sometimes not. My cross needs work, my F2L needs work, my lookahead can improve, and I still have plenty of algorithms left to learn.
If you're stuck at a minute, 45 seconds, or right around 30, don't immediately assume you need to turn harder or memorize everything.
Find where your solve stops flowing, figure out why, and practice that part.
The timer usually follows.
Try It Yourself
Working toward sub-30? Pick one weakness and spend some time focusing on it instead of only doing full-time solves. Once you've practiced it for a while, share what you worked on and how your average changed.
I'll use the anonymous responses to see which practice changes are actually helping cubers improve.
About the author: Dhakshit Gona has been solving the 3x3 for nearly 10 years and currently averages around 30 seconds, with a PB in the low 20s. He uses CFOP and is working on improving F2L, lookahead, and OLL/PLL recognition.
