Winter School on Programming, Day 6
On the final day of the competition, the problem-setter was once again from SPbU. This time it was 
Andrey Stankevich — a prize-winner of international championships, a coach of world champions, a jury chair and member of various Russian championships, and last year's best coach, and these are not all of the lecturer's accomplishments.
The topic of this lecture was "Data Structures for Statistics on Segments".
The lecture lasted right up to the end and was delivered at a very high level; unlike the previous ones, all the questions were covered in considerable detail, but grasping material at this level required some grounding in the fundamentals and some experience, which, I think, kept most of those present from understanding the content of the lecture, us included. Toward the end of the lecture there was a discussion of the multidimensional segment tree. In general, the emphasis was on the use of structures and optimizing the program for time.
In the main round of the competition, ten problems were presented in English; some of the statements were standard, others had no storyline at all. The difficulty level of the problems matched the author's level; in the end about fourteen teams managed to solve at least one problem. As for us, even when we did know how to make a working algorithm, we understood that it wouldn't pass within the time limit, or that running our version would require 4 GB of memory, while only 256 MB was allowed. As a result, we couldn't even attempt to solve most of the problems. One or two required working with a graph, but even that wouldn't have helped. Many problems required working with structures, and none of us could do that, plus we had no experience with optimization. I remember that only at the PTL did they teach us to write optimal and universal programs, rather than programs that fail to function on critical input data.
The lecture going over the solutions to the problems was the longest (it ended at 17:00) and the most detailed. The solution to each problem was examined down to the smallest details, with the nuances pointed out, but once again the solutions were beyond what most participants knew, and the detailed walkthrough was of little help, except perhaps that it gave some people a headache. Toward the end it was noted that the entire set consisted of slightly modified problems from the Petrozavodsk training camps.
During the upsolving session, few people were working on solving today's problems for the reasons stated above, and besides, some of the teams were leaving today. As for us, we set about finishing off the sponsor's problem, since by midday there had been only three solutions to it. In the end, most of the upsolving time went into getting it to read comma-separated numbers. After that we ended up with an almost working program; after a quick modification of the main algorithm the program worked — the solution wasn't the most optimal, but it passed the main interesting tests. After that I sent the source from two of my mailboxes to the address they gave. My coach and I (the rest of the team had left the upsolving session long before the sponsor's problem was solved) arrived at dinner later than everyone, when most people had already left the dining hall. After a quick dinner we headed off to the wrap-up of the results.
Strangely enough, the concluding part of the day turned out to be fairly short compared with the previous ones. The VIP team from DonNU received a one-gram platinum bar for the best solution to a problem, since their code for problem C came out more concise than the author's. The best team of the day once again turned out to be the Pointless team from KNU, with nine problems solved. After that the final day of the competition came to an end, and we headed off to the hotel, where this time we made no attempts to start a holy war, but instead discussed achievements and impressions of the interactive computer-entertainment industry and more — for example, we ran a couple of camera tests.
Now all that's left is to wait for the closing of the School, where they'll announce the participants' placings and sum up the final results, and they'll hand out the elephants. But I doubt that the final day of the competition had much effect on the statistics of the teams' ratings.
Edited on 22.02.2008