Also this year did we participate in the 39th ACM Intercollegiate Programming Contest World Finals, held in Marrakech, Morocco; the ‘we’ being: the “I Can’t Pronounce Catachtonic” team composed of Yaseen Hamdulay, Robert Spencer, and Sean Wentzel, and me as coach, from the University of Cape Town. We’re the only team from Sub-Saharan Africa, and one of 10 teams in the Africa & Middle East Region, of a total of 128 teams that participated, who were selected from 38160 contestants from 2534 universities of 101 countries on 6 continents that competed in the qualifying regionals.
One year more of studying, practice, and training wiser, our last training session indicated we might be a contender for A&ME regional winner. (For overall winner, we’d need to have and do what the medalist teams do, such as starting training in your early teens and winning IMO and IOI, weekly training sessions, monthly local contests, week-long training camps by previous medal winners, designated labs, competitive programming courses, scholarships and whatnot that other coaches talked about regarding preparations. At this point in time, we don’t have nearly enough such resources.)
The ‘first to solve a problem’ did so in a mere 5 minutes, from opening the envelope with the problems to having submitted the right code! This was Problem A: Amalgamated Artichokes, and the honour went to Peking University, setting a new record. The UCT team did so in 14 minutes, due to being sidetracked with another first. Then it was like: is this the only ‘easy’ problem, and the rest as grueling as last year’s problem set, and will it come down to ‘the team who solves the second problem will win A&ME’? Soon thereafter, the UCT team solved a second problem—but then so did two other A&ME teams, upping the ante that perhaps the 3rd solved problem would be the decider. UCT was still leading—at some point even on 27th position in the overall scoreboard. I got too nervous, and went for lunch, hoping they would have solved a 3rd one upon returning to the scoreboard. And lo an behold, they had, still leading for A&ME, though overall moving down to the 50s-70s in the dynamic scoreboard. To make matters more exciting for spectators, there were 5 PCs with shared screens and video, so one can see one’s team live on webcam, and see what they are coding, every single keystroke. Nifty, imho; nail-biting for some coaches of medal-contenders.
Right before the scoreboard was frozen regarding solved problems (for the last hour of the 5-hour nonstop contest), the American University of Cairo had solved 4 problems, surpassing UCT, but at the cost of a lot of penalty time due to a few wrong submissions, so if the UCT team would solve another problem, and Cairo not, then we’d win A&ME region on time difference. I could see UCT submitted a solution, hoping it was right. Then, sitting in the spectator area, and the Cairo team sitting near that, the scoreboard updated that they had submitted a solution for their 5th problem… and then came the involuntary reaction of its team members, being a mini-cheer. And UCT did not surpass that in the last 30 minutes. Overall, this placed Cairo on 75th place in the final standing, winning the prize for the A&ME region, and UCT just below that, as second in the A&ME region on a very respectable 83, therewith also receiving congrats from other participants, coaches, and interested spectators.
The “I Can’t Pronounce Catachtonic” team from UCT
So, relatively, they did well, having solved an impressive 4 problems, being A, D, I, and J, and all correct on first submission. This placed UCT ahead of other well-known, and arguably better resourced, universities, such as Uni Illinois at Urbana-Champaign, Virginia Tech, IIIT, Uni Western Australia, Cornell, Moscow Aviation, Calgary, and Rice. That said, at the other end of the spectrum, St. Petersburg ITMO broke the record of having solved all contest problems—a first of all the 39 editions of the ICPC world finals—and first to solve problem G. Moscow State Uni came second (11 problems solved out of 13, with first to solve B and H), Uni of Tokyo came third (also 11 problems solved, with first to solve J and K), and the fourth gold medal went to Tsinghua University (10 problems solved, and first to solve C).
If you don’t feel like solving the problems yourself, but still want to know the answer to, among others, cheese slicing, shooting asteroids, tile cutting, and the qanat irrigation system, then have a look at former UCT coach Bruce Merry’s analysis of the problems and directions of the solutions.
All in all, it was a good World Finals. An the food was good, the weather good, the other events too (including a fun camel ride), meeting up with coaches and some contestants met last year, the CLI symposium brought some useful information as well, and Steven and Felix Halim generously gave me a hardcopy of their Competitive Programming 3 book. Sean won the ICPC Quest, so a 1st prize was brought back to Cape Town.
The planning for participating with a strong UCT team next year has commenced; the 2016 finals will be in Thailand.
Keet blog 