Hello from Berlin!
Get ready for some action as participants are ready to battle for Grand Prize of 300000 russian roubles and coveted title of Yandex.Algorithm champion
Stay tuned for live commentaries
If you have any questions about the competition feel free to ask. Competition starts at 11:30 CET, about an hour from now. Currently participants set up their PCs
-0:33 — We have quite tough field this year including ACM ICPC World Champions, TCO, GCJ and IOI winners. snarknews had provided some info on finalists here
-0:10 — Contest is delayed for 10 minutes
-0:09 — You can follow live results here
-0:05 — Currently contest area is filled with several journalists and cameramen
0:01 — Contest had been started, participants reading problems atm
0:16 — there are several submits on D, some blind and some open. and 1 submit on both B and E, blind
0:19 — peter50216 is gettings TLEs on problem D, seems like map just do not cut it
0:23 — First submit on A. Also second one on B. All submits not on D are blind now
0:24 — marek.cygan is first one to submit 2 problems and currently in the lead
0:26 — Problem A is a tricky one — sometimes you need to add something for stack before removing
0:32 — During test run I solved A, B and D during the contest and C a couple of minutes later, though I have seen A and B before
0:33 — Currently marek.cygan, pperm, s-quark and LayCurse are in lead with 2, while usual favorites have only one problem under their belt
0:36 — Some more subbmits, most of participants with 2 solved D and either B or E, though there are also one blind submission on A
0:42 — Merkurev may spend a lot of time on A as currently he have the mistake I mentioned. To his defence authors originaly made the very same mistake
0:44 — A lot of photos made by some professionals
0:46 — Contest almost reached its equator, and we have out first participant with 3 problems — s-quark. He solved all the usual suspects — B, D and E
0:50 — vepifanov joined him, though remeber — as Yandex empoyee he participates out of competition
0:51 — And now we have someone we sure solved 3 problems — peter50216
1:00 — Both Petr and tourist are now in group of people who solved 3
1:01 — eatmore is actually quite close to solve A — he only needs to understand that sometimes it is good to make 2048
1:04 — SirShokoladina has most chances to be first to solve 4 as he still have B to submit. Also Eryx is the first one to solve C, he uses all open tactics and guaranteed to has at least 3 problems
1:07 — Seems like polish participants have good 3D imagination — pparys solves C as well
1:08 — And first submit from GCJ and TCO champion rng_58 — correct solution on F. We now have at least blind submits on all problems
1:13 — Petr and origami
1:14 — And s-quark claims the lead with 4 problems. He solved A
1:16 — As evident from the posted image Petr works on C
1:18 — And peter50216 joins him with all open submits
1:21 — 3 people with 4 tasks — are this our winners?
1:22 — Though one of them participates out of competition, so no
1:27 — Both Petr and tourist debugging problem C
1:29 — And Petr takes the lead with submit on C
1:31 — The contest currently fulfills all 4 criterias for ideal one — everyone solved something, nobody solved everything, every problem is solved by someone, no problem is solved by everyone
1:33 — 7 people with 4 problems atm, SirShokoladina in the lead
1:35 — Both Petr and tourist trying to solve A in time
1:37 — Winter is coming. Oh, I mean, end of the contest
We have out final results. Congratulations for Gennady Korotkevich for winning it all!
Is vepifanov participating in competition?
As he is Yandex employee now, he will participate in the final round out of competition.
Winter is coming since 1999.
Something wrong with results:(
Each person is duplicated (More clever people to our world!) and the second version have minuses instead of '?'
Now everything is fixed.
Please try to find better keyboard next year...
The given one was German keyboard layout. This is very inconvenient for those who don't remember other keyboard layouts.
Or allow participants use their own laptops :)
Hm, isn't it possible just to change keyboard layout in system settings? Then it doesn't matter whether 'z' and 'y' are swapped, or whatever is written on keys at all. Though I agree that letters on keys might be irritating, the proper approach in this case would be providing smth like Das Keyboard.
You can change keyboard layout in system settings, but it won't help you much if you usually look at your keyboard when typing and don't remember exact locations of particular keys.
Lol, are there coders who need to constantly look on the keyboard? Sometimes I look on the keyboard, but I can write text with my eyes closed with not that many mistakes and I thought that applies for most of pro coders — especially two swapped labels won't change anything.
Surely there should not be a huge problem if one can change the layout. But..
Is the punctuation in the same places in the German layout? E.g. using vim while configuring a remote server with an unknown keyboard layout is a lot of fun, but very dangerous : )
Participants were able to choose their own layout
Upto the last line of the blog, it looked like tourist could not do well. The last line came as a typical behavior :P
By the way, is it normal that marek.cygan has -5 on his penalty time?
Read the rules. It's the feature of TCM/Time glorious scoring system.
I don't really understand, in what sense it's glorious, especially when it's mixed with lots of easy problems (compared to, for example, the problems from ICPC World Finals) and very small amount of time.
Can you, please, explain?
I said "glorious" with some irony, you can replace it with "fancy" if it sounds better to you :) Nevertheless, I can explain why I think it is better than "normal" ACM ICPC scoring.
There are 2 big disadvantages of traditional scoring systems (not only ACM, but also Codeforces/Topcoder, if we don't account challenges).
There is not much space for strategies. On ACM-like contests, it is always optimal to solve problems from the easiest to the hardest. On TC/CF, you can choose to solve hard problem first and it's sometimes optimal, but then again, how to determine the difficulty of the problem before the contest? On TC, 1000 is sometimes easier than 500 (and 900 is almost always easier than 600); on CF, it's even worse.
One of the fundamental principles of game-design is the interaction between players. Sadly, with the traditional scoring systems, the interaction is minimal: you compete not against other participants, but against tasks. This significantly reduces the fun of the whole process.
Systems like TCM/Time at least try to solve these problems (to some extent). E.g., you can look through the past contests, and find out that a certain participant usually submits his solutions blindly, and also compute his success rate depending on difficulty; and then you can modify your strategy accordingly to maximize the probability of being ahead of the participant.
Of course TCM/Time is not perfect and doesn't completely resolve the problems I mentioned above, but at least it's a step in the right direction IMO.
hos.lyric is a boy or girl?
Boy.