I just recently became an expert and i know becoming candidate master takes much more practice than what it takes to become an expert and it varies person to person. So i'm just curious how many problems did you guys solve roughly to reach your level?
Auto comment: topic has been updated by thedominator (previous revision, new revision, compare).
It is different for everyone. For me it was around 500 problems. Moreover, if you solve hard problems, it will take you less time to reach CM (my views).
What rating of problems would you say are hard enough?
I'd say around 1900-2000 non-inflated rated problems is good enough to reach CM
Still some people solve many 1900 rated problems but still stay cyan or green.
focus on concepts rather than the problem
How to pick a problem as there is a huge pool of problems on code forces?
Unfair life smh
Life is unfair to people who complain about their doubts about the efficacy of methods that worked for other people. :)
Anyway, what worked for me was solving problems rated 2200-2400 (both inflated and non-inflated). Hope it gives you enough of a challenge.
How would you explain People with 1000+ problems yet green. Don't give me crap that they solved easy problems. i visualized their handles they did solve enough no of hard problems. It's not like people like tourist have some magical method, it's just that some are a little gifted and some are not.
Problems solved after reading editorial do not count.
If someone wants to improve then it means they are not good enough and that's why they see the editorial and learn from it. So I think it also counts.
learn != think
Yes, but to think, you must learn first. By learning new things we learn to think in new ways. Also, I read in a book that we shouldn't have fixed mindset and feel excitement about solving puzzles and if we can't solve then we should be excited to know the solution. This way we grow more. Also, I am not totally against you (or I don't even know what I am arguing with you about LOL), I just think that sometimes we need to see the editorial (when we starting problems of harder difficulty) and eventually solve problems on our own.
If you can solve $$$> 70\%$$$ of the problems you try without reading the editorial, you are trying too easy problems imo.
Hello, I am able to solve around 40% of the problems I attempt without looking at the editorial. Is this a good enoough number?
Yes, but make sure you actually understand the editorial of the remaining problems (rule of thumb: you should be able to solve any problem that you've attempted in the past).
I've usually noticed a pattern when it comes to the magnitude of the returns on solving problems. People with more mathematical maturity (i.e., being able to deal with decent amounts of math and having mathematical intuition) tend to have higher returns on the same investment as compared to someone who isn't as mathematically mature.
Some other factors are that some people don't perform well under pressure, and some people are just better at thinking for longer periods of time.
Will someone please tell me what is meant by "inflated" and "non-inflated"?
For some people, old problems with rating X are harder than new problems with rating X, and for some, the opposite is the case. When I said both inflated and non-inflated, I meant both old and new.
Although I have dropped back to expert, back when I had hit candidate master for the first time, I guess I had solved more or less 500 problems. Also, I started solving quality problems (1800, 1900 — pushing myself in general) only recently (sometime around the beginning of May of this year, if I remember correctly), and I believe that's what helped me the most.
All the best!
dude hats off for your consistency, thanks for inspiring me :)
How many problems did you guys solve to reach expert?
around 300-320. I was green till 260 though
Can you please allow me to send you message on codeforces?
I did. You can send a msg now
Around 300-350, but at this point I realized I should solved harder problems and started to solve 1700-2000 and learning harder concepts. But I think it may change cuz of some factors, like your past at mathematics or something like contests/olympiads.
I had about 750 when I hit candidate master for the first time, though I think at least 250 of them were wasted, useless problems, as I practiced on problems that were too easy for me at the time.
How to realize if a certain rating problem has become a cakewalk for you?
If you solve it under 30 mins or you try 10 problems and AC 8-10,
30 minutes is a long time to be honest, at least for me. My threshold is much less, probably closer to 10 minutes or so.
Maybe that's why you're a 'master'
How do you realize if a problem is useless for growth? I personally can solve 1600-1800 sometimes(but not in contests ofc) but sometimes I'm unable to solve even B problems usually rated 1300-1400. (Ik what's the cause : it's just my mental state. I get tensed up a lot even tho I try not to) So how to determine not growth worthy problem?
Same thing I felt a few months ago. Now I realise it was bc of 2 reasons: 1.I was either underestimating/overestimating some problems. 2.The problem was similar to a standard problem or concept which I haven't solved.
Exactly! Whenever I solve A within 5 mins or sometimes even less(around 2 or 3 mins) I try to overthink for B as I mess B a lot. If I AC B (which happens quite less often) I totally overthink C as there are already 3k-5k solves (random figure) and I get tensed up like crazy. And guess what?! After the contest if I know the problem is rated somewhere around 1400-1600 range then I solve it within 30-40 mins.
I think a general rule I try myself, and advise to others is try solving problems that are at about your current rating +200/300 to maximise efficiency.
(At your own risk lmao)
Div2 A/B are often solved by making one observation, so they're not really algorithmic problems. And there are some especially annoying ones that are either solved in 2 minutes or bricked for an hour.
And about getting tensed up, don't worry if you can't solve some of them, even I couldn't solve Div2 B in the previous contest. And eventually, your rating losses won't do as much as you might fear. galen_colin tested this by recreating his entire codeforces contest history but without including any rating losses and the resulting rating was only 50 more than his actual rating.
True. A lot of people do brick A or B but they end up solving till D so they have a +ve delta.
694 on codeforces plus extra 100~150 on other ojs so about 800. Honestly just work on your intuition and math cuz all you have to do is to solve the first three or four problems quickly to get to CM (No need to solve some crazy advanced problems in a contest). So solve a handful amount of greedy/observation problems to improve your intuition and perhaps work on some intermediate techniques(dp and such) if you think your already very good at greedy/math. Nevertheless, solve problems rated about 1900-2000.
400~600 I guess. I think ratings nowadays are inflated to the point that you only need to solve at most one algorithmic problem (like D) with enough speed to reach CM. Knowing the very basic graph algos+binary search+fenwick is definitely enough.
I think i'm good enough at solving algorithmic problems. But i get stuck at problems which can be solved only via good observation and math. I gotta work on that
and how to find problem on observation to practice?
Everything < 1700 is pure observation.
For me around $$$1600+$$$ problems, I don't quite remember. It takes patience, but you'll get there.
To be honest , I think that the quality of the problems you solve and whether you figured out by yourself is much important than the quantity.
Personally , I solved about 1500 problems to reach Candidate Master , but I guess that only about 500 of them made an effect.
I had about 600-700 when I reached Candidate Master.
Why did your rating fell so much though?
Looks like codemastercpp just got infected with a "rainboy syndrome".
Rating adjusts to true skill with time
Your rating graph is interesting, for sure.
You just became expert by good luck due to easy problems you did fast,you just reached specialist and then this so first try to save this position which is going to be hard for you to,you will soon be specialist.
Thanks for the insight!
Maybe 300? I think I will reach within 300, maybe you too.
Sorry I didn't make it :( it's been around 325 and I haven't made it tho.
I think it took me around -250 problems
Bruh. That Means i am doing it all wrong from the first Place ..
According to me, to became CM both solving problems off the contest and in the contest matter. I have seen many coders solving 1000+ problems but still struggling to reach that mark. So with having adequate knowledge, how you utilise it in contests also matter! :)
I have solved around ~2500 including other platforms to reach CM. In my opinion, solving ~500 problems of 1800-2100 rating problems will be good.