ICPC regional contest season is coming. Here in Japan, the first round of Japanese ICPC regional contest was held last week, and the result is up now. Link
You can see the passed team as shaded blue, and "A/B/C" means how the team was advanced (A is just because of the good result, B is the newest universities coming to ICPC, C is the director's arbitrary decision).
As soon as you take a look on the sheet, you can find how some of the top teams are eliminated from the selection. This year, at most 4 teams from the same university can advance to the next round, and 6 out of 7 top teams are from the same university, The University of Tokyo. The tragedy has happened. 6th and 7th team are eliminated from the contest, even though they were much (!) better than 38 teams advancing to the Asia Regional!
Some of you here knows how notorious the team selection of ICPC Asia regional is: complex concept of slots and multiple participation, as well as CJHwang's arbitral decisions. Here in Japan Domestic, it's not quite different: every year there are some troublesome arbitral selections, unfair choosing rules.
Of course, the official tries to maintain the number of universities they invite, since it's crucial for deciding how many teams from the site can advance to the World Finals. However, the current strategy is obviously not a good idea. I agree that they have to keep the number of distinct schools, yet it's nonsense that they are setting a limit of the number of teams from a university. If I were the admin, firstly I'd choose 30 teams which are the first rank of their university, then I'd choose unselected teams from the top. I can't believe that the admins don't feel anything about choosing the 40th team (even 2nd in the university) over the 6th team.
That is the (notorious) case of Japanese ACM-ICPC. How is it like in other countries? I'm the most curious about the system in Taiwan, since National Taiwan University dominates Taiwan regional every year, and I might be able to find some coincidence from that. Also, are there any other good ways of solving the issue? I'd like to see many people's comments.
To Japanese students: Ignore poor Japanese regional. You can qualify for WF from other regionals and they are MUCH nicer. My favorite is Korean regional.
To Taiwanese students: I would recommend Japan regional, which is good and stable. (and Japan is a nice country IMO :P)
In contrast, I have encountered several bad things like wrong testdata, trailing garbage in input, and duplicated problems in Taiwan regional. I don't know the quality of recent Taiwan regionals, hope it becomes better now.
Btw, I got rank 1, 1, 4, 1 in Taiwan but only 3, 2, 2 in Japan because I could not beat rng_58 lol...
The selection rules of Taiwan regional is even more ridiculous. I just checked that there are 8 different ways to qualify, and each of them has a strict limit for the number of teams from the same university. There were many teams of National Taiwan University failed to qualify just like the teams of University of Tokyo.
And the selection rules are changed almost every year. So the coach should study the rule carefully in the start of each year in order to miss some chance to get slots.
The selection rules will be changed again in the next ICPC Taiwan Council Board meeting, since Taiwan regional will increase the number of teams significantly in 2018. The meeting will take place in the end of March or the beginning of April, 2018. If anyone has ideas about the selection rules, please feel free to contact me or any other board member.
To Japanese students: Korea regional limits 1 team per foreign university in recent two years. Indonesia, Vietnam, Thailand regional contests did not have any limits and qualification on foreign teams. Taiwan regional accepts up to 20 foreign teams in 2017. We probably increase the number of foreign teams, since we are going to increase the total number of teams from 80 to 100 or 120. National Taiwan University will host ACM-ICPC Asia Taipei Regional 2018. If you need a ticket to WF 2019, you should come.
To Taiwanese students: If you are unable to win a ticket from any regional in PSP subregion, then you should consider Japan regional. IMO, Indonesia regional is also a good regional.
"it's nonsense that they are setting a limit of the number of teams from a university" — isn't it the same for World Finals? It is kinda often case that on CERC teams that advance to WF are from places like 1, 3, 7, 8, because Warsaw took places 2 and 4. Moreover, that second team would still be probably good enough to get a medal on WF outperforming vast majority of teams from all over the world that got selected to WF. I really don't like this rule, but your case is not far from selection to WF on the scale of ridiculousness.
However that is similarly ridiculous only as long as we consider just the A rule, C rule seems like a big scam, is it explicit room for briberies or what?
It is (especially in World Finals) an intercollegiate contest, so I can understand if the limit is 1. However, I have no idea why they set the limit to 2-4 even though they ignore the duplicated (2nd team of a college) in the result of regional contest.
If not to the 6th team, the last C tickets should have gone to the best unselected, and top in the college teams instead of second team with poor performance.
You could say that I am cruel, but I think it is better for university to reduce number of its teams participating in semifinals. Usually university has not more than two teams strong enough to struggle for medals on world finals. With very high probability they are between 4 best performed teams on quarterfinal. Taking two more teams to semifinal make them possible randomly win 1-2 best teams on semifinals. In MIPT we had several such cases =)
1 team limit would not help in any way just moving competition from regional to subregional.
Of course if limit will be 1 team it will be bad. But 4 teams limit imho better then 6 team limit.
Sorry, I misread the initial comment at first, thought you mean 1-2 best teams should go to the regional
This things also happen in Daejeon regional (Korea site), mostly because of weird WF slot rules of Asia regional. Asia region gives slot for WF to each regional considering how many different universities participated in each regional. So each regional have to invite as many universities as they can, but they also should invite good teams (regardless of their university). This makes unfair result for some universities, such as Tokyo university in Japan, Seoul National University in Korea, and so on.
I don't know much about Europe or America region's slot rule, but rule for Asia region is really weird..
What I learned from ICPC Asia Council meetings in WF 2016:
Site score is the major factor of slot allocation. However, the allocation schemes in different sugregions (EC, WC, PSP) are different.
Site score is roughly 0.7*#regional_univ + 0.25*#regional_team + 0.05*#provincial_team. In order to maximize the site score, each regional may determine their own selection rule.
When calculating site scores, the weights of on-site contests and online contests are 0.8 and 0.2, respectively. The weights are determined by ICPC HQ.
What I learned from Bill Poucher's speech in WF 2016:
ICPC HQ prefers equity (not equality). Therefore, there are some preference slots, and more regular slots for regions with more participants.
However, ICPC HQ plans to change the system in the future. There will be leagues finals between world finals and regionals.