Sub-30 cuber here. My data says I might need ~10,000 solves to reach sub-20

[vnsalaryman]

I'm currently around sub-30 (AO100 ~28s) and trying to reach sub-20.

I exported my csTimer history and tried estimating how many solves it might take to get there.

My current stats:
AO100: ~28s
Total solves: ~1,800

Based on my recent improvement rate, the estimate is:

~10,000 more solves to reach sub-20.


Full disclosure: I actually built a small dashboard to explore this idea.

It estimates “How many solves until Sub-X?” based on your recent solves.



You can upload csTimer sessions and see:

• solve time distribution
• session drill-down
• AI performance insights
• estimated solves remaining

If anyone wants to try it:
subx.guru

 

[cuberbutnotacuberV2]

Dang, this website is pretty interesting! I like the fact that it's very CubeDesk-esque; the data presented is done so in a very minimalist manner, and it's also quite comprehensive. Great website, seems like a good tool to have.

 

[Puzzle Maniac]

Reminder that "solve more" is not always the right answer to improvement. I average around 8.5 seconds and I have done around 9K lifetime 3x3 solves. Always focus more on learning and drilling algs, and building good habits especially during F2L.

 

[Isaiah Garrison]

Thats awesome! Unfortunately I can't do it because I use cubic timer on my phone. Its a .txt file

 

[CubeWizard23]

Reminder that "solve more" is not always the right answer to improvement. I average around 8.5 seconds and I have done around 9K lifetime 3x3 solves. Always focus more on learning and drilling algs, and building good habits especially during F2L.

I've done 18k solves in the past five months. How in the world?

 

[Michael Wimberly]

It did NOT take me 10,000 solves to get from sub-30 to sub-20. Just learn lookahead, better efficiency, and full PLL and it shouldn't take more than a thousand or two. Just doing more solves is never the answer.

 

[vnsalaryman]

Dang, this website is pretty interesting! I like the fact that it's very CubeDesk-esque; the data presented is done so in a very minimalist manner, and it's also quite comprehensive. Great website, seems like a good tool to have.

Thanks, I really appreciate that!

I wanted something that shows a lot of data but still feels simple to read.

Right now I'm still experimenting with the prediction model, so feedback from other cubers is super helpful. If anything in the stats feels off or confusing I would love to hear it!

 

[vnsalaryman]

Reminder that "solve more" is not always the right answer to improvement. I average around 8.5 seconds and I have done around 9K lifetime 3x3 solves. Always focus more on learning and drilling algs, and building good habits especially during F2L.

Yeah I completely agree.

The idea isn't that "just do more solves" is the solution. The prediction is more about estimating the amount of practice time people typically accumulate before hitting the next milestone.

Technique improvements (lookahead, efficiency, alg knowledge, etc.) obviously change that curve a lot.

I'm hoping that as more people upload their data the model can start reflecting those differences better.

 

[vnsalaryman]

It did NOT take me 10,000 solves to get from sub-30 to sub-20. Just learn lookahead, better efficiency, and full PLL and it shouldn't take more than a thousand or two. Just doing more solves is never the answer.

That's a fair point.

The 10k estimate is just based on my own current improvement rate (AO100 trend), so it's definitely not a universal number.

I believe some cubers progress much faster if they focus heavily on technique upgrades like PLL, lookahead, or F2L efficiency.

 

[Caleb Valenzuela]

i have 12k solves and i average 6.8
I think targeted practice does more than solves. I feel like solves are basically just: “combine all that you have learned and apply them here!!!”

 

[Michael Wimberly]

i have 12k solves and i average 6.8
I think targeted practice does more than solves. I feel like solves are basically just: “combine all that you have learned and apply them here!!!”

Yeah, I have 2k solves and I average 20

 

[vnsalaryman]

i have 12k solves and i average 6.8
I think targeted practice does more than solves. I feel like solves are basically just: “combine all that you have learned and apply them here!!!”

I agree!

that's actually why the dashboard has the “phase” section.

The solve estimate isn’t meant to say “just do X solves and you'll get there”. It's more trying to estimate how much total practice people tend to accumulate before reaching a milestone.

The phase section tries to describe what skills cubers at that level usually develop (PLL, better lookahead, F2L efficiency, etc.).

So the idea is more like:

solves ≈ amount of practice exposure
skills = what actually drives improvement

Still experimenting with how to model that better though.