Блог пользователя PORONGON

Автор PORONGON, история, 4 года назад, По-английски

Hello community,

I write a little entry here to open a small debate on Div. 1 Rounds. As we all know, these are "the main dish" in this platform. The most interesting competitions held within Codeforces have always been Div1 (or Div1+Div2) rounds, and it makes perfect sense.

I've been thinking a little bit, and I came up with an idea that I think may be helpful to improve the experience during the competition in Div. 1 rounds. I think that most of rounds would be far more interesting if we made them last one hour longer.

Taking a look on recent (and past) rounds on Div1, it is the case that in most cases the participants in the top positions do not solve the whole problemset. This is right, since the last problem is supposed to be (a priori) truly challenging. However, if at not even at least one among the best programmers in the world is able to solve all the problems within the time, this means that maybe the rounds are really short.

Also, since there are not so many Div1 rounds in general, I would suggest to make them last 3 hours. There's no loss here, and more people could be on the move of competing with challenging tasks.

Thank you, and if you don't agree, let's have a little discussion here. It is on the behalf of improving the experience on CF.

Cheers,

Roberto Esquivel Cabrera.

  • Проголосовать: нравится
  • +147
  • Проголосовать: не нравится

»
4 года назад, # |
  Проголосовать: нравится +45 Проголосовать: не нравится

I agree it would often be more interesting if it were an hour longer, plus I also think that longer contests reduce the randomness in performance.

There are two things to be careful of I think:

  • If this effectively results in div1's needing a 6th problem to be considered 'good', (or the 5th problem needs to be harder), making all div1s be 3 hours might make it harder to prepare a good div1 contest, which would result in fewer of them.

  • If you hold contests on weekdays, longer contests might make it more inconvenient to compete. Lots of people are either in college (and then a 3-hour wednesday contest is more likely to intersect one of their classes), or have a job (in which case they likely have to take another 30% more time off to compete.

Definitely I think on weekends 3-hour div1s would be fun though!

»
4 года назад, # |
Rev. 2   Проголосовать: нравится +33 Проголосовать: не нравится

Feel free to downvote.

It's an interesting idea. But as a Chinese, I can only participate the rounds at 22:35 in the evening, so maybe it's a little late for me to sleep at 1:35AM :(

(It seems i don't have enough qualifications to comment because i only participate in just one div.1 round then fall to blue:(

»
4 года назад, # |
  Проголосовать: нравится +35 Проголосовать: не нравится

On the contrary, I quite like 2-hour rounds, and would love to see 1.5-hour rounds as well. Longer, not so much.

With stuff going on around me, I can usually snatch 2 hours to take part in a contest, but with 2.5 and 3+ hours, it becomes unproportionally more difficult.

  • »
    »
    4 года назад, # ^ |
      Проголосовать: нравится +26 Проголосовать: не нравится

    Totally agree! 2 hours is fine, but when I see 2,5 or 3 hours that's already a commitment.

»
4 года назад, # |
  Проголосовать: нравится +25 Проголосовать: не нравится
Histograms for #664

If we look on distribution of number of submits by time, this suggestion looks fine for div1.
But what to do with corresponding div2 round? I think that if div2 round is extended to 3h too, than count of memes "me after solving ABC" will be small only by modulo 1000000007

»
4 года назад, # |
  Проголосовать: нравится +46 Проголосовать: не нравится

I think the problem is not the duration, but problem balance. It seems that problem setters try to optimize the results for the top competitors only, while you have nearly 1,5K participants.

For example today's contest. Problem A was way more complicated than A should be. That maybe fine if you give it 750 then and make B 1000, but then look at B, it is also not B on average. So you have 2 hard easy problems (with apparently weak pretests so we have massive B fail). I am not sure how was C, because I went for D after.

And that goes for most of div 1 rounds. It is sad to see scoreboard when you have 80% of participants with 1 or 2 problems and no attempts of other problems. Also many times you see submits for B on minute 20 and then silence for 1:40. So I highly doubt 3 hours are going to change things.

On the other hand look at Div 2 or Edu now. Most of Div1 guys compete there as well and it seems fun (for example Div 3 a week ago with 500 guys from Div 1 you had only 80 totals).

So my point is there should be improved problem ladder for Div 1. Because currently everyone below place 200 ranks by luck (like couple of minutes in penalty). And again, you have 1500 participants. So ideally you have A (super easy implementation but little trick in idea, or simple idea and some simple DS to implement), B same but potentially both. Then C and D should go like today's A and B. Maybe tricks, ad-hocs, maybe tricky implementations, but top 100 should still think those problems as boring. And then you have problem for reds and problem for Petr and tourist (like 2500 and 3000). So like 500-1000-1500-1750-2500-3500. And if you have 1500 participants then ideally you have A solved by almost everyone, B goes for 1000+, C 700-800, D 500-800, E probably 200-400 and then F is whatever.

Unfortunately now for me as someone who is doing that for fun, div 1 is not fun. I enjoy div 2 and edu way more. Like today was a stress, when I opened A and B and had 0 ideas how to solve it quickly. D saved my ass, and then I realized how to solve A, so ended up not bad, but still. From 5 problems only 2 in 2 hours.