Page 1 of 1

ICPC '05 regional in Manila and Coimbatore

Posted: Wed Dec 14, 2005 5:35 am
by Ilham Kurnia
Hi all,

I participated in Manila and some of my friends participated in Coimbatore ICPC regional. Some (if not all) participants of both regionals suffered from the following things:

1. Poorly written problem set.
2. Lack of experience from the organizers.
3. Strange judging and mysterious compile errors.

Does anyone know if there is some sort quality control for the regionals from the world final organizers?

Posted: Wed Dec 14, 2005 8:08 am
by Cho
Absolutely agree with your comments on the Manila site.

Posted: Wed Dec 14, 2005 9:24 am
by misof
The correct way to file a complaint (an excerpt from the rules):

Complaints, Appeals, and Remedies

If irregularities or misconduct are observed during the contest, team members or coaches should bring them to the attention of the contest officials so that action may be taken as soon as possible.
After the conclusion of the contest and the results have been made public, coaches may file complaints or appeals as follows:

- Within 2 business days
The coach may file a complaint by sending an email containing a text message with no attachments to the director of his regional contest and copied to the Contest Manager.
- Within 3 more business days
The RCD shall respond to the complaint.
- Within 1 more business days
The coach may file an appeal with the Appeals Committee through the ICPC Registration System.
- Within 4 more business days
The Appeals Committee will investigate the circumstances of the appeal and notify the coach and RCD of their decision.

This process is governed as follows:
- The results of the regional contest are not final until the complaints and appeals process has run its course.
- Only coaches may file complaints and appeals.
- An appeal must be based on one or more of the following circumstances: violations of the Rules, misconduct by teams, or gross misconduct by contest officials with the intent to do harm.
- The decisions of the judges are final. Specifically, a decision on a problem submission MAY NOT be appealed.
- The Appeals Committee overturns decisions only under extraordinary circumstances.
- The decision of the Appeals Committee is final.
- No additional finals invitations will be given to remedy to a complaint.
- All complaints will be acknowledged.

The appeal will be automatically rejected if the above procedure is not followed.



My observations: It's far too late for the complaint to change the results. If you had some doubts or even better proofs that something went wrong, your coach was the one responsible to file a complaint as soon as possible.

You can't change the regional contest's results anymore. What you can do:
- write a letter that describes all the problems you encountered, gives all the proofs you have and requires the responsibles to comment on the issues you mention and to make sure that they won't happen next year. Get as many teams/coaches to sign this letter, then send it to your Regional Contest Director (he's the one to comment), to the Contest Director for Asia and to the Director for all regional contests (they are the ones you want to inform about these issues). You can find all the necessary information at http://icpc.baylor.edu/icpc/default.htm, in particular at http://icpc.baylor.edu/icpc/regionals/default.html and http://icpc.baylor.edu/icpc/People/

Posted: Wed Dec 14, 2005 11:42 pm
by Dreamer#1
Imagine an ICPC Regional contest going on with a ranklist showing only which team solved how many problems and no other information about which problems are being solved by which team or when or anything else. And at the end of the contest you figure out that one of the easiest problem remained unsolved but you'd spent hours behind that and all of your 6/7 submissions on that prob got WA just like many other teams except the very lucky ones who decided to skip that. This is exactly what happened at the other regional site in India, i.e., the Kolkata site. Even the better Indian teams couldn't do well and foreign teams like us were in complete shock. Even the Indian team that did so well to get a rank at WF last year (probably as the first indian team to get a rank in WF's history) were at the bottom of top 10. Its really unfortunate. :(

But it wasn't always like this, in 2004 the regional at Kanpur site, India was organized very well. Hope they take it back to Kanpur from Kolkata next year.

Posted: Thu Dec 15, 2005 12:01 pm
by Ilham Kurnia
Thx misof. But I was more in the line whether the chief contest director monitor the regional contests, rather than appealing the results. Does he believe whatever the regional director reports or does he crosscheck the result to find anything peculiar? Or does he need to be pointed to a problem before he will do anything about it?

Even appealing the results will be difficult as the organizers have a big shield with the only excuses to appeal the results are "violations of the Rules, misconduct by teams, or gross misconduct by contest officials with the intent to do harm." Can writing a poor problem set be categorized as "gross misconduct"? In some regionals, the organizers even give themselves more shield by stating in the local rules that a clarification request need not be acknowledged/answered. Another thing, I felt that local teams had an unfair advantage, but then again, what can I do to have a proof of that? How could I know if the judges responded more quickly to the locals?

Last year I also participated in the Manila regional, and while the organizers and the problem set were slightly better than this year's, those were not up to a proper standard yet (with unanswered clarifications, and, again, several poorly written problems). They didn't even want to release the testcases for the controversial problems... Like they were afraid that others will find a mistake that they probably already knew, but one that they didn't react to during the contest.

I was quite enthusiastic about the ACM ICPC, but after experiencing two rather unfair contests, I am more inclined not to do anything about it.

Edit: I just remember that the year before in Manila (2003, that is), a friend of mine told me that the judges gave out the testcases. And at that time, some teams found a huge mistake in the testcases, and protested. That caused the ranklist to be recalculated after it was finalized on the ICPC web. I wonder if they tried not to dig the very same hole again by not releasing the testcases...

Posted: Thu Dec 15, 2005 3:52 pm
by Per
Ilham Kurnia wrote:Thx misof. But I was more in the line whether the chief contest director monitor the regional contests, rather than appealing the results. Does he believe whatever the regional director reports or does he crosscheck the result to find anything peculiar? Or does he need to be pointed to a problem before he will do anything about it?
I think it is virtually impossible to detect all such quality problems without anyone complaining about them. I usually check out most of the regional contests, and while I sometimes get the impression that some of them are poorly organised, it is very hard to draw any real conclusion just from the web sites. The director of regional contests will probably face a similar problem. While (s)he has more information at hand than I do (just the web page), I think one has to actually be there and see for oneself how well (or not well) organised the contest really is. Therefore, it is important that you, the people who were there, raise any complaints you might have.

And of course, the more complaints they get, the more likely they are to take it seriously.

Posted: Thu Dec 22, 2005 2:33 pm
by shamim
I participated in the Coimbatore site and I too observed few irregularities.

As mentioned previously, the site had poor problem description.
But the most funny thing is the way they do the judging. In the practice contest, there was a problem of printing all possible path. But nothing was mentioned as to the ordering of the paths. It was also not clear from the sample I/O. As mentioned previously, they were either late or never replied to clarifications.

Later on, I talked to the site director and he told me that, the judging is done manually, i.e. human inspection, so ordering doesn't matter.
How shocking!!! I am quite sure that is not the way to conduct ICPC. Then, what is the need for PC2 software.

I think it is high time that, there is some sort of regularity board to manage the problemset and judging of ICPC contest. UVA has something like that, the Elite Panel.

I mean this is a worldwide contest and we can't have local arrangements ruin the contest of some site.

Imagine the World Cup Soccer qualifying round being played with different manner at different regions. One region using Coconuts for the ball while others playing on rocky,bumpy fields when everyone should have used the ball prescribed by FIFA.

Some regional contests are becoming like this.

ooff...

Posted: Sat Dec 24, 2005 1:58 pm
by sohel
I also participated in the Coimbatore site..
... other than the 3 chairs that they have provided, nothing was right.

Here are few points.

1. We got the question paper 30 minutues before the start of the contest...
.. the problems were given in the auditorium, and that gave opportunities for discussion not only within the team mates but with other teams as well. Its a 3 minutes walking journey from that place to the labs.

2. There was ambiguity/misprints/false information/unclear description in all the problems... and i mean all the problems.

3. They provided 3 sets of problemset for each team, they also provided soft copy of all the problems.

4. When it's written 'the integers are separated by commas', one should presume that the integers are separated by commas... but the story was different.

5. As shamim mentioned, the judging was done manually.. how ridiculous.

6. One team solved 2 problems in 1 minute.... !!!!!!!!

7. Most of the clarification were left unanswered... and the judge's reply was as slow as their brains....... sorry, but can't find any other words to describe them.

8. The volunteers were as 'smart' as the judge. We asked them, 'why aren't the judges replying'... their answer was 'if they judge now then you will know the verdict and that will ruin the fun'... what the hell !!!

There are many more.....

One more funny thing on Coimbaotre Contest

Posted: Sun Dec 25, 2005 6:14 pm
by DongJoo Kim
I also participated in Coimbatore contest.

One more funny thing is that our team CHANGED our computer DURING the contest because of manager's fault. (They mis-arrange the team seat.)

Of course, we are not allowed to move our source code.
Finally, we lost our time.
We re-typed the source code.

I'll say this contest is not about the algorithm but about the adaptaions about vague and unexpected environment!

Posted: Tue Dec 27, 2005 10:42 am
by Shamsul Alam
I participated in Coimbatore contest as a coach this year. To me everything was perfect there except the actual contest. The main problem I think was the inexperinence of the organizers rather than any bad intention. I am writing this not to change the outcome of the contest but to point out the problems that may ruin a contest that in turn causes a huge disappointment for the cotestants.

Let me once again mention the problems that occured in Coimbatore site.

1. It was very surprising that the problemset was given to the contestants in an auditorium where the contestants and their coaches were present. After that the contestants were asked whether they have any queries regarding the problems. It seems that the intention was good but the method was not right. It was very easy for the contestants to communicate with their coaches. I personally observed some contestants to talk with their coach and to other teams although it was not possible for me to know whether that was a general chat.

2. It was funny that after one hour of the distribution of the problemset the contest was officially started. For this one hour, the organizers direction to the contestants was NOT to use the computer and not to type any code. But, interesting thing is that some teams solved a problem in zero minutes and within one minute a team solved two. Obviously, some teams were actively doing their coding in this hour. Those who were honest were punished.

3. Response of judges were very slow. Verdicts came even two hours after submission. There is also some reasons to believe that judging was done by eye inspection. One of the submission was judged as accepted where there was a spelling mistake. In another problem the order of the output integer was not mentioned and according to a clarification any one will do. In an informal talk with a contestant with the associate site director the judging was done by eye inspection, so ordering was not a problem.

4. Response from the judges were confusing. A program which compiles in the team machine got 'compile error' from the judges. One particular submission used gets function which should issue a warning not an error. But this particular submission received 'compile error' verdict from the judges. After the team resubmitted the same code only changing the gets function, they recieved 'accepted'.

5. The standard of the problemset was extremely poor. Any smart first year student who have completed his first course on a programming language should solve at least five problems. Rest of the problems are not clear enough to make some guess. BTW, guessing correctly what was in the mind of the problemsetter was the most important determinant of getting a solution accepted.

6. Problems in the set were unimaginably flawed. The statements were ambiguous. There were major disagreement between input specification and judges' data. For example, in the input specification it was said that the integers will be separated by commas but in judges' data they are not. Some of the problems were also missing with possible input data range.

7. The score board mentioned only the number of problems solved and penalty points. No mention were there on which problems were solved by a team, when and in how many attempts.

8. Each team was placed in big separate rooms. To the contestants it feels like they are participating in an online contest from a restricted room. May be the intention was good but the excitement and charms of a contest arena was absent. Miserably, they have spent a lot but failed to spent a small amount for the balloons.


From the experience of Coimbatore contest I think the follwoing things may be considered so that future contests do not end up with this kind of disaster despite all out efforts from the organizing teams.

1. As all the major problems are occured due to poorly designed problemset I think a person with enough experience in problemsetting may be engaged with a site from ICPC head quarters. This is important specially for the first time organizers.

2. At least one person, possibly more, with prior experience of an ICPC contest should be in the organizing team who are organizing an ICPC contest for the first time. The person(s) should be present on site during the contest and involved in the judging procedure.