CCC 2011 - Are you ready?

[mirhagk]

to get into stage 2, does everything have to be perfect, like what is the normal score to get into stage 2?

[Tony]

It greatly depends on the year, problem difficulties, and who else is participating. Waterloo says "about 20" of top contestants.

[bbi5291]

Based on the unofficial scores I have collected, there appear to be three students at Woburn, four at Don Mills, and at least one at Massey who have at least 60 on the Senior contest. Taking this into consideration along with the fact that several people claim that s1, s2, s3, and s4 were quite easy, I think the cutoff for stage 2 will be between 55 and 60. (Screwing up one case out of five on s3 seems to have been quite common.)

(In the past, I have usually overestimated.)

Also, in the years when I've been to stage 2, the selection rule has always been that all official contestants with scores greater than or equal to X are invited, where X is the maximum value such that at least 20 contestants are invited. These 20 include the Beijing and Hong Kong people though. (Don't let this intimidate you too much --- the best Chinese coders, who are insanely good, have better things to do than write Canadian contests.)

[MarkTwo]

bbi5291 @ Tue Mar 01, 2011 6:17 pm wrote:

If there's anyone here who finished the contest, used one of the languages supported by our autograder, and wants to know his/her score, you can email me your solutions (bbi5291@gmail.com) and I'll mark them for you. (Languages supported: C/C++, Pascal, Java, Python, Ruby, Perl, Haskell, OCaml --- no Turing.)

Would you pls send me the testdata, or the problems? I felt extremely uncertain about my progs... I'm so worried I must've made lots of mistakes pls...

[Shanethe13]

MarkTwo @ Wed Mar 02, 2011 6:53 am wrote:

Would you pls send me the testdata, or the problems? I felt extremely uncertain about my progs... I'm so worried I must've made lots of mistakes pls...

I'm pretty sure that within a week or so of the contest being written, the testdata and problems normally appear on the unofficial solutions website. Until then, I doubt anyone would want to release testdata out into the wild, since some people haven't written yet.

[ultimatebuster]

What do you get for Junior contests if you get a high score?

[antybash]

bbi5291 wrote:

... several people claim that s1, s2, s3, and s4 were quite easy, I think the cutoff for stage 2 will be between 55 and 60. (Screwing up one case out of five on s3 seems to have been quite common.) ...

Adding to this discussion, I do agree s1-4 were too easy for CCC. Unfortunately, as with many people I somehow missed a few cases in s3 (I really can't wait until I found out why.) and my score was just under 60. And based on results from dwite I know there are lots of people who are capable of >60. Thus, I believe the cutoff will be above 60.

Has the cutoff ever been greater than 60 in past years? And on another note, problem setters usually post solutions to their contests. Why doesn't the CCC do the same? This has always puzzled me.

[A.J]

Well, the cutoff was 66 three years ago. That year was by far the easiest CCC. However, I doubt that the cutoff this year will be that much above 60; it is most probably 60 or somewhere in the high 50's.

[bbi5291]

antybash @ Wed Mar 02, 2011 11:03 am wrote:

Adding to this discussion, I do agree s1-4 were too easy for CCC. Unfortunately, as with many people I somehow missed a few cases in s3 (I really can't wait until I found out why.) and my score was just under 60. And based on results from dwite I know there are lots of people who are capable of >60. Thus, I believe the cutoff will be above 60.

In my experience: other people screw up too. For example, nobody got perfect on stage 1 last year; yes, the problems were hard, but they were also pretty standard. Robin and I overlooked a case on Animal Farm, and Jacob overlooked an optimization on Nutrient Tree. Another example is that in 2008, the IOI team was more or less controlled by who screwed up least (on stage 2). Aaron Voelker (Sane) is an incredibly good coder now, but he was at best third in Canada at the time, yet finished first overall; Hanson screwed up and ended up second, and Guru screwed up and ended up sixth.

So there may be a lot of people who seem as though they're capable of > 60 on DWITE, but that doesn't mean they'll actually get > 60.

Quote:

And on another note, problem setters usually post solutions to their contests. Why doesn't the CCC do the same? This has always puzzled me.

They always say they're going to do it, but I guess they're just too busy. Also, they know that there's that unofficial solutions site; they even link to it on the official page.

[Dratino]

I agree that S1 through S4 were relatively easy compared to other years. If I can solve the first 3 problems in less than 30 minutes, it's probably too easy. I'm not exactly the best programmer out there.

antybash @ Wed Mar 02, 2011 11:03 am wrote:

bbi5291 wrote:

... several people claim that s1, s2, s3, and s4 were quite easy, I think the cutoff for stage 2 will be between 55 and 60. (Screwing up one case out of five on s3 seems to have been quite common.) ...

Adding to this discussion, I do agree s1-4 were too easy for CCC. Unfortunately, as with many people I somehow missed a few cases in s3 (I really can't wait until I found out why.) and my score was just under 60. And based on results from dwite I know there are lots of people who are capable of >60. Thus, I believe the cutoff will be above 60.

DWITE is a pretty bad indicator of who will do well in the CCC.

ultimatebuster wrote:

What do you get for Junior contests if you get a high score?

I believe it was $100.

bbi5291 wrote:

They always say they're going to do it, but I guess they're just too busy. Also, they know that there's that unofficial solutions site; they even link to it on the official page.

There's also no real need for official solutions - there are many ways to solve a problem, and having an "official" solution would be kind of strange. That being said, they could release some unofficial test data or something.

(Oh, and if anybody's wondering who I am, I'm Joe Zeng, from Don Mills Collegiate Institute, the same school as Cyril. I had an account here two years ago, but I haven't bothered to remember what the password was. Also, I decided to start afresh, without the whole "math, writing, and DDR" introduction.)

[helLOHELlo]

I just knew my "unoffical" score
it's not very good
difinitely not to stage two and not honourable mention
but since i didn't expected i could go to stage two or honourable mention anyways...
i'm happy about it

ALSO
I just find out that the unoffical solution website has post the problems and input/output files.. and some of their solutions

[ultimatebuster]

Have you guys seen this?

Quote:

Note 2: J4 does not require "q 0" (or any other command) to be entered if the program terminates due to "DANGER".

However on the contest it clearly had another line of command before the program quit due to DANGER.

Thiw is what i'm talking about as to "vague"

[trintith]

It was fun! Compared to some of the other years, I didn't think it was too bad. S1,S2 and S4 I finished within the first ~40 minutes. S3 took me 20 minutes just to get though. I'm curious as to what the correct solution for S5 is - can't wait until Friday! I thought I got 57 points, but I sent bbi an incorrect copy of S5, so I actually got an extra 4.5 points, I think. Hope the cutoff isn't too high...

(I'm new here, so hi!)

[Tony]

helLOHELlo @ Wed Mar 02, 2011 8:02 pm wrote:

ALSO
I just find out that the unoffical solution website has post the problems and input/output files.. and some of their solutions

Ok. I guess we can talk about the contest now.

[antybash]

Dratino wrote:

...there are many ways to solve a problem, and having an "official" solution would be kind of strange...

What I meant was a sort of analysis of the problems, much like AJ did for dwite, and Brian for the Mock CCC.

Anyways, I glanced over the test cases for S5 and it occurred to me the constraints in the problem set specifically said "4 <= N <= 1000". Yet the greatest test case that they tested one on was N = 25. This is well inside the constraints for a brute force style solution.