My way of learning atcoder (~light blue)
It had been languishing for about a month just before light blue, but ARC106 finally made it to light blue. https://gyazo.com/f6135300b90474a9c0175d10ef3dfc84
https://gyazo.com/0c19e85e8e4d6363373c906c7b48e375
I don't usually look at it much, but you can see various stats on AtCoder Problems. So far, they have done 151 problems. I've been doing this for about 5 months, so I'm averaging 1 question a day.
https://gyazo.com/1e5e3a08cebe8e17297bbb8b359d9ed6
https://gyazo.com/3fb401a14bfca098dcd55d38b74d2dd1
The heat map shows that there is a lot of activity in June and July, and then mainly on weekends.
https://gyazo.com/a545e29a1fe805d14403c557efcb5833
I feel like crystallization is happening slowly, with multiple issues reacting at different times. The name "verbalization of awareness" is not appropriate.
To name the technique used in the deformation process and make it a tag on Scrapbox.
This definitely contributes to rate improvement.
Began to maintain the library.
Solved learning contests.
Contest about [dynamic programming
I did this around July when there was a lot of activity.
Contests on various algorithms added to the [AtCoder Library
He was stepping in front of the light blue.
The "contents of the toolbox" have increased with the above study and library maintenance.
However, I felt that it was weak to look at a problem and attribute it to a solution method or algorithm, saying, "If I use that tool like this, I can solve it.
This is simply put: "Instead of looking at a problem and writing a program, write a solution in natural language.
I've only been at it for about 4 days, but it feels good.
The detailed verbalization of "what we thought" clarifies what we misunderstood, missed, or didn't realize.
What we no longer do
Numba, Cython and Numpy.
If you think carefully about your current position, it is more important to arrive promptly at a solution of the appropriate order than to speed up such a process!
This type of acceleration is not appropriate as an investment of time.
So I'll just use PyPy3 until PyPy3 says "nothing is wrong, but I can't AC because the language is too slow".
Aiming for full completion after the contest.
The ones who try to solve all the problems they couldn't solve after looking at the explanations.
For a while, "solving it all" had become an objective, but was that an objective? I reconsidered.
---
---
This page is auto-translated from /nishio/僕のatcoderの学び方(〜水色) using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.