same

Look bro it's not that hard

• Firstly, believe yourself that you can do it, If you don't believe yourself that you can do it, how will you do it.

• Practice has no alternative, Practice problems with rating 200 more than your current ratiing

• Always upsolve after contests, this is very important tip okay?

• Be consistent and participate in each and every contest to maximise learning

• Always review codes of top coders like [user;tourist] sir, It will help you a lot

First give me some advice for practice. How can I become a Master?

If you know the answers to my question, then surely you will find the answer to your own question.

Last of all , Congratulations for become a Master!!

Auto comment: topic has been updated by icyo (previous revision, new revision, compare).

step one: be born Chinese

step two: study for three months

I can't possibly comment on ways to become a GM but about about algorithms — I personally find Library Checker to be an excellent collection of standard(and non-standard) data structures/algorithms and their variations. It's maintained by an LGM, which definitely boosts its credibility.

Also, while dated, this blog by E869120 might help as well.

It'll be almost impossible to get to GM if you sometimes can solve problems only up to "difficulty" 2400. You simply have to get lucky in some round and find a problem that's just right for you such that you solve it and others don't, then you'll get IGM performance and go to GM. From there what awaits you is most likely just dropping right back to master though.

Forget about problem rating, just solve problems. You calling problem rating "difficulty" blinds you from the actual difficulty of the problem.

As for algorithms, just learn whatever you need to when upsolving. There aren't that many algorithms to learn that are common in CP so you should learn them quite quickly if you upsolve consistently.

I used to wonder how to get to GM, until I solved enough problems.

I guess some part of my brain naturally improved through time so that I can consistently have a GM performance without intentionally learning anything new and I still have awful implementing skills.(You will learn new algorithms when you tried to solve problems)

I am not sure whether this would be the case for higher levels, but I believe to reach GM you just need to solve enough problems(for me, I only brainstorm the problems without even implementing)

I think you should practice binary search, try to solve problem ABC fast you will be fine.

I think the binary search is one of the most important algorithms.

When considering difficulty of a problem, time is a very important resource. Being able to solve a 2400 problem in 3 hours and 30 minutes is drastically different. Sometimes you'll have to solve harder problems you know how to solve. Not a great performance, but a good example.

So consider using something like ladders (basically solve problems in increasing difficulty until you understand the editorial). Solving any Div1 round like this works well. Give yourself plenty of time (around 1-1.5 hours) before reading the editorial for a problem. After you've failed to solve the problem by yourself, try to implement it off of editorial.

This helped me greatly in achieving GM.

Stop learning useless algorithms, go and solve some problems, learn how to use binary search.

From Um_nik's famous blog

Always ask the question: "-is-this-fft-?"