Hi there,
As you can see in my rating graph, I used to be a candidate master. However, since I reached my peak, my performance has dropped significantly. I had to admit that it was probably my fault, as I stopped training as hard as before.
Recently, I did horribly in the December USACO Plat. contest. Then, I proceed to absolutely bomb the next three CF rounds I participated (including today's Div. 2). Not only was I unable to solve the harder problems (D and E), I had trouble in solving C.
I'm very sad right now and I hope to seek some help from you. How should I train in the next month to get back to candidate master? I will definitely do more problems, but what strategy should I use? Should I start from 1800-1900ish problems? (Previously I've done 2100-2200+ problems in my training, but since I wasn't as good as before, should I "restart"?)
Thanks. :(
You need to solve (almost) all problems which you weren't able to solve during contest.
Thanks. So you're saying that I should upsolve the problems that I missed?
Yes!
Regardless of whether it's true or not i wouldn't recommend to use advices from account with much negative contribution (at least wait for more advices from more people before using it)
ok i have positive contribution and completely agree with the first guy
No, it's definitely not you
If you look at my graph the same thing happened when I got to CM for the first time, also for GM and for IGM. Basically it means that you got lucky and luck isn't eternal so practice more and gitgud. The advice of upsolving is good but I wouldn't go as far as saying solve every problem, but 1-3 problems ahead of what you could solve. If you go back to those contests that you didn't do well, do you think you could've done something better or was it out of your reach? I'm pretty sure most people have lots of things to learn from that aren't that far away, for example having smarter implementation.
Thanks for the advice. I did let loose of my training after I hit CM (which luck definitely played a role). One quick question: do you think I should practice 1800-1900 rated problems (considering my current rating) or higher rated problems since my goal is to reach CM?
I don't think problem rating is worth considering. It's not good enough to be taken as a good estimative, especially in the case of someone aiming to be div1 because of the artificial barrier between div1 and div2.
I recommend doing contests and virtuals then letting those tell you what you need to practice. Unless you're AK'ing everything there will be something to learn.
Thank you so much for your inspiration and hints, especially for words "luck isn't eternal so practice more", and "if you go back to those contests that you didn't do well, do you think you could've done something better or was it out of your reach". I will keep this in mind and move on!
Exact same situation
You should wait for nearly a year...
Jokes aside, I suggest you to practice AtCoder(not new ABC)
you can simply change your rank to anything you want right now
Just give me your password (to boost your account ofc)
LOL
My password is the Ramsey number $$$R_{10, 9}$$$.
Good look figuring it out because it's an unsolved problem.
My password is the last 10 digits of pi
Then you don't even know your password
My password is the first consecutive 10 digit prime in the digits of e.
Easy to bruteforse, there are ~12k possible numbers.
Practice more? I'm not really sure anything else will get you there.
Time Machine, maybe?
At your level you will hit CM eventually if you give contests frequently.
Mental attitude to contests are very important. Having trouble getting positive delta isn't always skill issue. You just need to chill and try to solve everything you can without getting nervous
This right here. It's way more important than it may seem, and it's really easy to get anxious because you're taking a bit more to solve B, which makes you make mistakes, and you get even more anxious.
Do you train by solving virtual contests or using the "problemset" mode?
Investing too much into the "problemset" kind of training can lead to bad habits (i.e., the brain gets used to the fact there's no time pressure and there's "infinite" time to solve the problem).
Yes, that's a good point. Right now, I'm trying to do VP and upsolve 2-3 problems that I didn't solve in contest.
What is VP?
Virtual participation.