A Look Back at the 2013 World Finals - Jonathan

Post date: Jul 8, 2013 3:51:27 AM

Here I will attempt to analyze the factors which ultimately resulted in our embarrassing performance at the 2013 World Finals in St. Petersburg, and hopefully provide a learning example so future teams avoid the same pitfalls. As much as I'd like to attribute it all to a case of bad luck, we were just not ready as a team. Oh, and contrary to what some believe, we didn't stay up until 5am the night before; our condition was great at the start of the contest.

We lacked serious team practices on full problem sets under actual time constraints, especially in the few months prior to the competition since two of the team members moved to the US to work, and instead opted to practice individually on single problems over a period of time. Though this raised our individual abilities, we were unable to develop many essential skills needed for an icpc style competition.

I do not believe that the order in which we attempted the problems was mistaken. Unlike last year where it was really clear what was supposed to be done for many of the problem, this year J was the only straightforward problem, and so it makes sense to start coding it first while reading the other problems. Though we were half serious about trying to get first solve for J, it was already solved at 30 minutes, so that was no longer a factor.

The main difficulty we had this year was not knowing when to move on as a team.

What we should have done was to give up on J immediately after the first wrong submission, and to give up on F after all three of us had read it and none of us had any ideas. Instead, we spent almost 4 hours of the contest on F and J combined and in the end solved neither of them.

I was aiming too high for the number of problems to solve, and probably affected the team with it. It was obvious from the list of people attending that the problems this year would be harder - the most solved problem this year was a greedy problem, whereas a greedy of comparable difficulty was one of the less solved problems last year. If we had aimed for 3 problems like Paul suggested, then that would've put us in a different mindset and perhaps we would have given up on those problems earlier to solve others and potentially have done much better.

There were also several cases of miscommunication, where different team members had solutions to different parts of the problem, but none said anything because they thought the part they solved was trivial. This can only be fixed by having more team practices...

But all that said, we really are still weak and inexperienced, not just as a team but individually as well. Even if everything went optimally, with our current level we could solve at most 5 or 6 problems during the contest, and probably with quite a high penalty time. We will continue to train hard for the future, and so should future UBC teams.

Go UBC, and good luck in your next world finals.