[EDIT]
I think at this point we all can safely agree that there is no such thing as "talent" and even if there is such a thing,hard work beats talent any day.So I request you all to please stop arguing about this.I genuinely made this post to get all of your suggestions and get some tips/tricks which might help me along the way but this post had turned into more of an argument about why all the LGMs were born with this so called talent and are invincible and can't be reached. I believe that with enough hard work, anyone can reach LGM and if that is not happening, it is because we are not putting enough effort.I agree I'm not the right person to say this when I myself am terrible at CP but I am willing to invest a lot of my time to get better at it.
[POST]
Dear all,
I am not good at competitive programming but I am willing to put in the effort to get better at it. I had a small question which I though might help me and a lot of people like me who want to get better at CP.
It is not a question asking how to solve problem X or which method to use.
1:
I just wanted to know if, let's say you start solving a question. Can you look at the question and realize.. Oh!! this question can be solved using DP or graphs or some paradigm Y.
Or rather, how long does it usually take for the idea to click in your head?
2:
Also, do you have a specific method to determine what to use(I know it is not possible in most cases) like looking at the constraints or the time limit or anything like that?
3:
Is there any particular pattern you follow while solving problems? Like.. do you select a problem to solve in you free time randomly or based on a particular tag?
4:
I found a couple of good blog entries by Morass and DarthPrince from which I practice topic specific problems and I am practicing algorithms from CLRS. Are there any resources that I am missing? If so can you please specify the names or post the links?
5:
Also, on an average basis, how much time do you spend on CP per day?
Thank you.
For the first question : First of all, i've just started learning CP for couples of months, so personally i think i dont have enough experiences, forgive me if i say something wrong. But i remembered i read an article about interviewing tourist, he said he tried many approaches for a specific problem better than ones.
Can you attach a link to the interview? It is always not a bad idea to read an interview of any LGM :)
Here But it was written in 2009 (at that time he was 14), too long until now :'(
Thanks a lot!!
"Tourist is no genius but he is simply good at it"
I really love tourist lol
You can't reach LGM or very high in CP if you are not born genius in maths or CP. Well, you can become red with lots of hard work. If a person is genius, he doesn't need to think, the idea/solution will automatically come to him. If you are a normal person, then try to deduce the problem to something simple, you might know if you have solved this kind of problem before, if not, then try thinking about what if i model the elements as graph nodes? what if i know the solution till this point, how can i extend it?
Rest will come with experience.
Thanks for your response mate.
I'll try to solve more and more questions to get more experience :)
That's just an excuse for one to feel satisfied with what they already have, geniuses aren't born, they're made.
I think it's okay for one to not have high goals, but using this as an excuse is not.
Yes, bro, I agree you. Any one can be genius with having high ambitions(goal) and working hard for it with a proper way and guidelines. Here one's should have mentality not to be lazy.
"You can't reach LGM or very high in CP if you are not born genius in maths or CP"
It's nothing but an excuse.
You are wrong. They were strong in maths and CP without working too hard from birth. Ask tourist Petr mnbvmar dotorya Um_nik fateice rng_58 SkyDec
Yeah, sure, ask me. My rating was 3000+ from birth. Also I didn't write this.
It seems everyone is missing my point (downvotes). I am just saying you can't reach "LGM" with just ordinary skills, you need to be talented at maths,CP or logic by birth.
I am saying that no "LGM" struggled for like even 3-4 contests in green/low cyan. They easily reached blue and then worked hard(less than other ordinary people) and with the help of their talent, they reached "LGM".
Also, your blog post above mentions about being red and not about reaching LGM which is a whole different matter.
I am not "LGM" myself so if there is some faults in my statements, please correct me.
I started participating on CF when I already has about 250 solved problems on Timus. And I was participating in math olympiads since 2nd grade. It is not from God or whatever. It is hard work for many years.
WoW. I know what it means to solve 250 problems on Timus. That's great. You changed my mindset. Thank you very much.
So it's just about exposure at early age and then lots and lots of hard work.
You had a huge rating boost in 2015. What were you solving at that time?
Anything I can reach
It seems like word "talent" is really helpful default excuse when somebody needs to find reasons for not being good at something — as it has been pointed out in another comment above.
They easily reached blue and then worked hard(less than other ordinary people) and with the help of their talent, they reached "LGM". — how many of them do you know personally? What stays behind your claim and how do you know about "less than other ordinary people" part?
I just want to add that I can think about more than a dozen of people for which I know for sure that they were actively training and putting a lot of work into developing their CP skills over multiple years (because of following them closely, or from private conversations, or from attending training camps together etc.), and somehow it turns out that the lowest rated among them is in high yellow / low red — and that's probably only because he was more focused on preparation for team events and developing corresponding set of skills, and he ended up with 2 ICPC medals as a result.
Low red is nowhere close to LGM, but still... I know a lot of people with low rating who claimed working hard, but for some reason I don't know any person about which I'd be able to say for sure that he actually worked hard — and didn't reach decent level. Maybe I'm lucky and it's a coincidence. Or maybe all non-talented people just dropped out early and decided not to put a lot of effort into their training.
Also I just spend a few minutes right now trying to recall any person for which I know that they reached level significantly higher than mine while putting significantly lower amount of time and effort into their preparation. I didn't manage to find any — but, however, it may simply mean that I'm incredibly freaking talented.
So I suspect that you are strongly overestimating role of innate traits in contest performance differences.
I really hate the word "talent". So what happens if somebody have talent over CP? Will you quit CF or what? It doesn't help people engage or practice in any mean, but just discourage (or flatter, which is also bad) someone.
"Talent" is basically subjective. How will you define a "talent"? an IQ? It's a thing that can't be measured or understood. Any claims about talents — they say something that they even don't know, and the result of that claim is nothing, if not disastrous. For your case, it was even proven false, quite easily XD
One person can reach red with much hard work in 2 year and no math/puzzle experience beforehand but another person who puts same amount of hard work reaches red in 4 years. The first person is called "talented" by me.
What do you think?
Great, you can now call that as notking Talent Index :D
I mean seriously, don't you believe this above statement holds true? If yes, then why does it happens? The first person has an edge over the second how?
This is exactly me, if you replace 2 year to 1.5 year. So I’ll try to answer.
If you think talent can be decided with notking talent index, then it’s obviously right, as it’s a circular argument. If we get to the cinematic notion of talent (Like, Sheldon Cooper or Ramanujan), I’ve never seen someone who is both very weird and very smart, so neither am I. If we get to a common notion of talent (this usually implies something like better brain ability) then we have no way to know about it, so I have no idea too. I’m not even interested in such — those things are unfalsifiable.
If you like to think about someone’s talent, then I can say nothing — maybe you can study neuroscience. For me, it’s actually an opposite : I’m very interested in someone working hard. I like the people who actively engages in competitive programming, regardless of rating scores. It doesn’t matter whether they are red or not. Some non red coders are also in my “favorite” list.
Why? They know the exact reason why people should practice — it’s not because it gives good rating scores. It’s because it is fun. It’s because it helps you to use your brain. By practicing, you learn algorithm, and more general problem solving strategy, and the way to “think”. In short, they know the virtue of learning. They know how to make their life better.
The biggest reason why I admire red coder is, because they are the one who best realize these by heart. There are some orange ~ blue coders who realized this, and I admire them — far more than some, “talented” red coders, who sadly didn’t realized this.
WoW You said many good things. I think now I should focus on learning and working rather than thinking about end goal. Do you mind sharing how you reached red in 1.5 years ?
XD
When I was in high school, I was training for IOI (not CF red), so I'm not sure whether it fits you or not. Whatever, my strategy is described here.
You can't make yourself 200 IQ
So what?..
I don't think there are any researches for competitive programming, but for bunch of other similar activities (like go and chess) plenty of studies have been done. While it is true that strong contestants there usually have higher than average IQ, it doesn't mean much. You can look for some old papers, like the research by Doll and Mayr (which is like 30 years old), which shows that there are chess GMs with IQ below 100. And I have to say that reaching chess GM is at least significantly harder and more time-consuming than reaching red color at CF :) Yes, usually top players have higher IQ — Kasparov has something like 135, and that's a lot. However, there are plenty of studies showing that it is very questionable if there is significant correlation between IQ and performance when all other factors are taken out of equation.
No. Geniuses are born, not made and cannot be made. Like a human can train himself to jump higher, but he can never fly by himself.
Practice can make an ant a better ant, an elephant a bigger elephant, but cannot convert an ant to an elephant. It can make fool, a smarter fool, a genius, a better genius, but cannot make a fool become a genius.
:v
If a fool can be a smarter fool then a smarter fool can be a smarter smarter fool.
Apply this a constant number of times and you get a genius.
unfortunately, the limit won't change, just like f(n) = 1 + 1/2 + 1/4 + 1/8 + ... + 1/2^n <= 2, no matter how large n is.
If a person becomes smarter, their training becomes smartrer, just like going from 0 to -n instead of 0 to n in your function.
Their training gets smarter, but surely the demand of their tasks will get just immeasurably harder. It is like a linear function versus an exponential function.
Not to mention that the more time you spend, the older you are and the closer to limit your brain cells are until they start degrading. Both logically and biologically, limit is born and cannot be made.
Tasks in codeforces are divided as A,B,C,D and E, not A and E only, so it won't get immeasurably harder by moving one step at a time.
A brain doesn't come with an expire date, it can be seen as a muscle, it's healthy as long as it's being used, if you mean it's possible that it's not healthy when one reaches age 80 then that's not a problem.
But regardless of what is true or not, I don't see the point as both will be the same for all people.
You are trying to prove that some people are born geniuses, can you give an example of a person who achieved something without spending any effort?
I said one's potential is born, but never have I said anyone can be at his or her 'full potential' at birth. The harder you work, the closer you get to your full potential, but no more than that. Like 1 + 1/2 + ..., the function does increase as n increases, but the sum will never exceed 2.
One step is immeasurably harder or not depends on how narrow your definition is. Codeforces organization of tasks does not represent the genius level. As you can see, sometimes they even mess up the order. And if someone can, for example, compute n = 1000000! in one second, if he now can compute n = 2000000! in one second, will you call it the second step? Will the effort to do that first step (learning a programming language and math) and the effort to do the second step (increase the array size or whatever) the same? Yes, but it just means that you consider a much smaller step, like from 1 to 1/2, and in the end, like I said, the limit will never change.
A brain doesn't come with an expiration date, or to be precise, no one can give a perfect expiration date for it, but it is not that hard to give an upper bound. Let's say 1000 years. Can anyone live 1000 years? So that is its expiration date. Brain is like muscle, but have u ever seen any sport player that can play forever? No, after they reach their limit, they will retire, sooner or later, but they will.
In the end, humans have limit, fools have limit, geniuses also have limits, but the genius's limit is just too high and cannot be made.
One step is immeasurably harder or not depends on how narrow your definition is.
You are right about it. I have started practising div2 D after solving 150 C problems and at least 50 problems of same difficulty from other OJs. But I am finding D immensely harder than C. Maybe it depends on our education process too. My country has less than 50 purples/above. No matter how much one tries, height doesn't increase after certain age. Maybe same applies for potential. But one can surely use more of his potential by hard work.
Well... OK, as you wish :)
Just curious, are you familiar with paper mentioned in comment below? It generally contradicts with your point of view and provides data and facts to prove it. I'd be happy to know what you think about it — are there some flaws in its logic which aren't obvious at first glance, or some data there is wrong, or they are using some tricks at some step, or what exactly makes it possible to have such a paper with conclusions different from what you claim is true?
You Can See many people Became Red Coders and LGM After Their rate was 1100
Can you give a few examples? Almost all LGM started from blue or high cyan.
zscoder had made it From grey and he became LGM But When Codeforces Changed rates he became International Grandmaster
He went to IMO 4 times and hence it proves my point about math talent.
I really don't know about his talent. I know zscoder personally, and he is really the most hardworking person I've ever know. Yes, he went IMO 4 times, but I don't know if it's due to talent or hardwork though. So I don't think you really prove anything.
I was someone who used to hate programming once. If I can reach pupil/specialist from there, I know I can reach to grandmaster from here. All I have to do is "not quit". By the way, I didn't reach directly to pupil/specialist. I solved tons of questions on hackerearth, codechef, hackerrank.
PS: I know pupil/specialist is not a great rating, but it is for me right now.
Reaching cyan from 0 is nothing compared to reaching LGM. You can even reach red, it's possible but LGM is something you can't .
One day, I will change this thread to "In your face" :D
ay2306 lets do this :)
.
I have a long way to go myself, but here are my answers to your questions:
-The time taken for the idea to click obviously depends on the question. If an idea does not cross my mind for a while, I try to think along the lines of "Can I apply X on this" with X being greedy, DP, Graph Theory, etc.
-I either solve problems from the A2OJ ladders in my free time or solve questions that my friends solved recently ("competitive coding" indeed)
-I do not know much about resources since I keep solving problems and learn the topic (from online resources) when I come across a question based on that topic and am unable to solve it.
Thank you very much for your response.
Also, do you solve questions from a2oj ladders or based on a particular topic(from a2oj)?
(Also thanks for not going into energy saving mode Houtarou senpai :P)
Mostly from the ladders (I'm trying to finish Div2C, Div2D and Div2E ladders right now)
But occasionally, I solve problems from Mathematics category on A2OJ because I love Maths.
(Lol @Houtarou senpai xD)
Thanks Ashishgup, your suggestions mean a lot.
Well generally there are two important things to know about it before attempting to solve it.
How hard it is? It helps you not get stuck on a crazy HLD idea if it is a Div 2. B.
The constraints and how good of a solution you re looking for. There are some specific constraints like when N < 40 it is probably some meet in the middle or when N < 20 it has something to do with bit masks.
Now regarding your first question. Usually when I see 109 + 7 I try to solve it with some DP or if I see some kind of links I try to apply graph theory. Actually, I have a bad quality that I often get stuck trying to apply some specific technique. But generally you should try to develop a sense of is something leading to the solution so you can faster go through different algorithms and in the end solve the problem faster.
Thank you.
I definitely think those points you made will help me :)
Don't know why I find this so funny.
more please.
That meme never gets old
This is exactly what I think when I solve a problem
....don't know, weather it will help you or not....recently I am reading a book, Think Like A Programmer: An Introduction to Creative Problem Solving... This may helpful...
[ Following Your Post :) ]
That looks good, thanks mate!
To improve, Codeforces rounds really work? even if you only can solve A,B and sometimes C?
I guess this blog somehow revolves around the nature vs nurture debate.
Anders Ericsson has worked on this topic quite a lot.The following is the abstract of his paper The Role of Deliberate Practice in the Acquisition of Expert Performance . WARNING It's a big read (~40 pages).
The theoretical framework presented in this article explains expert performance as the end result of individuals' prolonged efforts to improve performance while negotiating motivational and external constraints. In most domains of expertise, individuals begin in their childhood a regimen of effortful activities (deliberate practice) designed to optimize improvement. Individual differences, even among elite performers, are closely related to assessed amounts of deliberate practice. Many characteristics once believed to reflect innate talent are actually the result of intense practice extended for a minimum of 10 years. Analysis of expert performance provides unique evidence on the potential and limits of extreme environmental adaptation and learning
He has also released a book called Peak .Do check it out if you're interested.
It's great to see this post, nice work, I hope my comment helps.
I have a lot of work ahead still, but I think I can say a thing or two. I'm not red, far from it, but I went to the last ICPC World Finals (we did a pretty bad job at it, but just being there was great) and I help a lot of people in my university and in my country to get better at CP.
To the questions:
I usually analyze in which category the problem fits, then brainstorm the paradigmas, data structures, techniques and ideas I know for this kind of problem. The time it takes really depends on the problem and the problems you solved during your trainings. As Um_nik stated: solve many problems. The more you solve, the more it will be easier to solve new ones.
Looking at constraints and TL is necessary. You have to know the TL and constraints to discard some ideas during brainstorming. Solving a lot of problems is the way to know what you have to search while trying to solve new problems, since you will probably solve similar problems.
While I was training to ICPC (now I practice for fun and coach other teams), I had a routine of solving at least 5 problems a day, every week I studied and practice a new topic (if you didn't do it yet, get a Competitive Programming 3 and learn everything that's there. Solving at least 3 problems on each topic will give you a great algorithm base), and also one ICPC regionals every Saturday (with complete upsolving). During the semester that I did this I went from not knowing knapsack during South America regionals to be top 10 Brazilian team (see that my ratings didn't rise so much. Codeforces is another kind of competition, focus on you improvement and don't be totally focused on rating)
There are some blog posts on codeforces that are great (I'll comment the ones in my favorites later). Also I just found algo.wiki.is and I think it's awesome. CP3 is a great book to get base. Editorials of rounds and regionals are also a great resource of knowledge.
On my peak I did basically my whole free time (during some boring classes I usually studied, taking bus I read CP3, posts or editorials. At home I usually did some virtual contests). Now I read posts, select problems for people, help other contestants, do every round I can. It's something that I really like, and it's fun to me. Put some objective ahead, focus and work hard. If it's not being fun try to change. You don't have to be a red to learn from it and have a good time.
About the discussion about talent and stuff: I always worked hard on everything I love. I'm not a red and it's not in my plans to be one, but I'm from Brazil, from a not wealthy family, from a not great city, and I became a good Brazilian ICPC contestant, I had a fulltime job offer at Microsoft on USA, so I can say that talent is not the way to think. People that are red very early are usually good at math. This is not a gift, IMO contestants also had to study hard, they just studied early than other CP contestants. I say this to everyone: work hard is the key. If they know a lot of math it's because they studied. I was a math contest while I was young, and I didn't worked hard enough, I know people that knows math are not born with it, they just study and learn.
Anyway, good luck. Study hard and you will learn a lot.
Than you very much.
I was hoping for a response like this.
Your points were very valuable and I will try my best to learn a new topic everyday(or week at least) and solve problems based on it. Also, thank you for suggesting the book CP3 book and I will surely start reading it asap.
Good luck on your way to becoming an LGM :)
I don't aim to be a LGM. I already worked a lot and dedicated too much time for this. But I wish someone I teach have higher expectations and reach it xD I want to become yellow and maintain my rank (being on div1 seems hard enough for me right now)