Algorithm Semifinal Room #1 - 06.27.07 2:30 PM EDT
I'll try to give some on-the-fly coverage of the semi here.
9:02 - The round has just started, and ACRush has already submitted the 250 problem for 247,67. Let's see if he has any bugs...
9:10 - ACRush's solution seems correct, as the problem is quite easy, and the latter is confirmed by almost all cometitors submitting the easy.
9:18 - the problems for the round seem really strange, as the medium is straightforward 'code the recognition of the poker hands', and the hard seems pretty unsolvable during the round. So my bet is that someone who codes the poker hands faster will win.
9:32 - the poker takes more time than expected to code. The cost of single challenge is thus rising, so probably faster coding could not be enough to advance.
9:38 - meanwhile, everyone has submitted the easy.
9:40 - still no submits on the medium, despite three compiled solutions. Is everything OK with the sample tests?
9:47 - rumors say ACRush is still struggling with example 0.
9:49 - cyfra!!! here goes the first poker expert.
9:50 - Per submits the hard. Let's see...
9:52 - Per's code is quite understandable. For every value of n we check all the possible pivots, and try to interleave the two sequences for the smaller and for the larger elements to get a lexicographically sallest sequence where such pivot will be chosen. The interleaving, however, is quite tricky (the 'zip' routine). I still expect it to fail some corner case.
9:57 - no resubmits yet, and with the easy being straightforward the challenge phase might become a little boring.
9:59 - Ying submits the poker, too. The semifinal is starting to become as Chinese as it was expected.
10:02 - lovro is there, too. I wonder what is he going to do in the remaining time - solving the hard in 25 minutes should look like a crazy idea even to him.
10:06 - yet there's less than 50 points between all the poker experts.
10:08 - Jan_Kuipers decided to abandon the hard and is now back in play with an amazing 400+ score on the medium. He's now the most probable finalist in my view.
10:15 - it's pretty calm in here now. Even the chatting is not so intense.
10:16 - lars tries to get back in play with a medium submission.
10:23 - a last-second submission on the hard from xhl_kogitsune.
10:24 - the hard is actually very easy to challenge, as the input consists of only four numbers. I expect all the submissions on the 1000 to be 'tested' in the first minutes of the challenge phase.
10:25 - a little on-the-fly interview with misof revealed he doesn't agree with me on the difficulty of the hard problem, and expects it to be solved. Although he does agree on Jan_Kuipers being the most probable finalist.
10:28 - the challenge phase is almost starting. Let's see if anyone has enough courage to blind challenge. I expect that.
10:29 - the first couple of blind challenges go harmless.
10:33 - not much action now. -25 for Eryx.
10:35 - Tomy's easy is down. Quite surprising.
10:37 - Jan_Kuipers' 500 seems to attract a lot of attention. In some sense it increases his hopes on it being right, and in some sense decreases :)
10:40 - the challenge phase is over. Let's see if systests are lightning fast this time.
10:50 - and the results start to come in.
10:50 - no failures for the 7 bottom people.
10:50 - one medium has failed.
10:51 - xhl_kogitsune's hard is down.
10:51 - as is Per's.
10:51 - ACRush has only the easy.
10:58 - a 5-point gap for the finals, a 1-point gap for the wildcard, and the -25 kicking the favorite ACRush out of the competition. Tough semifinal.
10:59 - although I still believe such toughness was quite artificially brought by super-easy easy and super-hard hard.
11:02 - the decorations are already being changed for the Component Design and Development round. And that flags the end of our live coverage.
11:09 - join us later today (at 4:00 PM PDT) for the live coverage of the second semifinal at http://petr.mitrichev.ru/r2.html