Comments
|
0
I encourage you to implement it (it shouldn’t be that big, should it?) and test it against the above testcase :) |
|
+1
Let’s wait a bit. |
|
+24
Yeah, now I understand — it looks like the selection system is quite prone to this kind of thing happening. When I was a participant at the Russian IOI camps, there was about 12 days on average that decided the team, not 2, so it was much more stable and most gaps were more than one problem. |
|
+56
Yeah, it’s a pity that IOI selection was decided on that :( But how come it’s just one problem that decides who gets on the team and who doesn’t? How many problems in total affect the selection? |
|
-4
Apparently there was a technical problem with the recording, and most probably it’s lost :-( We’re trying to see if we can recover it from backups, but most probably it’s gone. Sorry. |
|
0
Yes. Meanwhile, you can ask and vote for questions at http://goo.gl/eoWf0 :) |
|
+1
Oops, I’m terribly sorry — the scoreboard contained a name in unknown language, but now that I look closer, it’s indeed in Korean. Apparently Po-Ru doesn’t have a Codeforces account. |
|
0
Never mind, I've read the editorial and the part that Gaussian elimination is not supposed to pass. It's strange that the assertions that verify that the Gaussian elimination error is not big do not fail in my solution, though.
|
|
0
Could you share the testcase my solution fails on Random Walk?
|
|
+10
Suppose the entrance and exit are fixed. For a given edge x, what is the expected number of times that the DFS will traverse it before reaching the exit? For edges that lie on the path from entrance to exit, it's 1, because we will always traverse them exactly once. What about, say, another edge from entrance - what is the possible cases for it? Can we traverse it 0 times? once? twice? more? What are the probabilities of those cases? Please tell if that's not enough to figure it out. |
|
0
Yeah, Egor is right.
|
|
+75
Yeah, let's lose some karma :) I agree. Problem C: a simple idea + standard problem. Implementing solution for standard problem takes much more time than inventing the idea. Problem D: a simple idea + standard problem. And an awful statement. "We'll find a list of such pairs of numbers for which the corresponding substrings of string x are equal to string y. Let's sort this list of pairs according to the pair's first number's increasing. The value of function F(x, y) equals the number of non-empty continuous sequences in the list." Seriously? What's the point in encrypting n(n+1)/2 that way? I've spent several minutes trying to convince myself what "continuous" means (I've used the Russian statement, but the English translation is good, so the statement is equally awful), I couldn't believe this long paragraph is just about n(n+1)/2. Problem E: a nice, no so easy idea - but still with a tedious standard problem after it! Why 100000? Why not 5000? I'm sorry for being so negative, I hope you understand it's just my impressions. See http://www.mit.edu/~jcb/tact.html |
|
0
I have no idea.
|
|
0
And the contest is over with no major changes at the top. meret with 100, then rng_58 with 90, then ivan.metelsky with 77. Let's wait for the large results to be revealed.
|
|
0
But meret retakes the lead with B-large and 100 points.
|
|
+20
|
|
+12
47 minutes to go, 3 people with 77.
|
|
+17
Meanwhile, we have a new leader - ivan.metelsky leads with 77 points by solving D-small. If we add up points for all datasets solved by at least one contestant, though, we get 140.
|
|
+12
|
|
+19
That long silence from rng_58 was for a reason - E-small+E-large and tentative second place!
|
|
+12
And that means we have a new leader - donehl
|
|
0
B-large is now submitted by 2 contestants.
|
|
0
|
|
0
Somebody else has attempted B-large. We'll learn who in 8 minutes :)
|
|
0
According to https://plus.google.com/110928860973677468355/posts/8wLb6QCw7eG, nika didn't make it. All others from Egor's post are there.
|
|
0
meret finally solves all "opened" problems. Since he doesn't have wrong submissions, one needs to solve something else to overtake him.
|
|
0
Wow, an unexpected submit from voover - I didn't even know he was participating - apparently somebody from http://apps.topcoder.com/forums/?module=Thread&threadID=706558&mc=114#1387375 couldn't make it.
|
|
0
At mid-contest, he's still in the lead. The problems solved by at least one contestant are A-small, A-large, B-small, E-small, but nobody got all of them (yet?).
|
|
0
Now meret takes the lead by solving A-small and A-large in addition to E-small.
|
|
0
|
|
0
|
|
0
1h18m - pashka still leading with 30 points. He was in the lead (shared for the first 25ish minutes) during the entire contest so far!
|
|
0
And a time-expired for him on B-large. Ouch.
|
|
0
More than an hour into the contest, still just 11 contestants with successful submissions.
|
|
0
There's probably too much submissions now to enumerate them. ACRush is the first TC target to submit a problem. B-small, after 2 wrong attempts. A 8-minute penalty to carry through the entire contest.
|
|
0
So does donehl
|
|
0
ivan.metelsky gets A-small.
|
|
0
There's also a long-standing attempt at D-small which I guess was unsuccessful since it's been more than 4 minutes.
|
|
0
And hanshuai. None of the submitters is a TopCoder target, while 5 of the competitors are. What are the heavyweights doing?
|
|
0
|
|
+1
pashka submits A-large as well. It looks like there's not much to get WA on, so hopefully it's good.
|
|
0
Meanwhile, there are also attempts at A-large and E-easy.
|
|
0
From vepifanov, also correct. Go Russia!
|
|
0
And there's also an attempt at B-easy.
|
|
0
Finally pashka gets A-easy!
|
|
0
Or maybe too tough? 25 minutes, still not even an easy input submission.
|
|
+10
Still no submissions at 15 minutes. Apparently the problems are tough, good!
|
|
+5
|
|
+15
No, I'm in Juan-les-Pins now :)
|
|
0
I'd guess people are still setting up because of some previous event (breakfast? :)) running late.
|
|
+10
No. I'm on vacation without contact with the event :)
|
|
+5
Apparently the contest gets postponed over and over again :)
|
|
+5
What is the main place where the spectators discuss what's going on?
|
|
+13
Yes, I was referring to the fact that hacking is usually much easier in a div 2 room (if the same contestant were to hack there).
|
|
+6
Oops
(a < b && t[i] > cutOff) || (a > n && t[i] < cutOff) and I've tested in on many cases :) |
|
+10
(this is similar to the debate about "IronMan" seeding in TopCoder)
|
|
+46
Thanks for the contest!
There's one thing that confuses me: were contestants still separated by division? If yes, it looks unfair because tournament advancement may be easier if you're in div 2. |
|
+3
Yes, I do.
|
|
0
I couldn't solve the first understanding in about 10 minutes, and solved the second one immediately, so I think that this issue has affected me significantly.
If you knew that the second interpretation is solvable, why did you opt for changing the problem instead of changing the testcases? |
|
+5
Same question about test 33 for problem F (although there it may well be true that all solutions are wrong).
|
|
+29
Lots of people seem to have WA on test 62 on D. Any clarifications about that?
|
|
+8
Yes, that's fine with me, too.
|
|
-4
This is of course a biased opinion since my round went poorly, but I think that the change of statement of problem E midway through the contest should make this round unrated.
|
|
0
As always, you can use your favorite search engine:
|
|
0
Yes.
|
|
+2
Oh yeah, that's a good point indeed. I take my words about D back :)
|
|
0
With the idea being?
I'm trying to find previous occurrences of B. |
|
0
About other problems: I believe problems B and D were too well-known :(
|
|
+12
Problem E is very nice :) I've got it accepted 3 minutes too late. The final solution is really short and cute, but it took a long time to get to.
|
|
+12
Anyone wants to share ideas about solving D?
|
|
+14
I believe this is a an example of how using Java makes you write more readable code :) I think if-elseif-else is more appropriate here.
I don't think this has anything to do about "smarter". I'd call this "with more attention to tiny details". And I'm perfectly happy to admit my compiler has more attention to tiny details :) |
|
0
C#
|
|
0
I don't understand what are you asking about.
Could you explain your question? |
|
0
And I believe the availability of various libraries is also not that important. So if I were to choose between C++ and Pascal, I'd choose Pascal because of the same argument (much more strict checking of everything).
|
|
+6
I think Java/C# (I don't see much difference between them except speed) are best suited for programming contests, since it's so much harder to make a mistake and so much easier to find and fix a mistake in a Java program than in a C/C++ program.
Much more strict type checking (implicit casts from long long to int and from int to bool??), out-of-range checking, code flow checking (allowing to read from uninitialized variables? why would a language allow that?), fantastic IDE which finds a lot of other mistakes for you, fantastically convenient debugging, more explicit syntax (a language with less power actually leads you to writing more readable programs), more explicit error messages (and the errors are always reproducible!) - to name a few advantages, but I've probably missed some more. I think that writing correct programs and fixing them quickly when they're not correct far outweigh the disadvantages mentioned above (slower execution, longer programs). Even a 2x slowdown is almost never important in programming competitions, while a WA always is :) And I believe that most of the time at a programming contest is spent in thinking (including the thinking you do _while_ writing code), not in writing code, so the length of the program (or the typing speed, for that matter) is irrelevant. |
|
+20
Thanks for the nice problems!
Although when I submitted D for the second time, I was almost completely sure it will fail again, since there're too many if's :) |
|
+5
It was great that all feedback was almost immediate. Good job on optimizing the judging!
|
|
+5
I think it makes sense to have this round at a significantly different time from previous ones.
Maybe 9:00 MSK = 7:00 most of Europe = 6:00 U.K. = 2:00 US East Coast = 23:00 US West Coast = 13:00 in China = 15:00 in Japan? It seems to be terribly inconvenient for US East Coast, but for others it looks OK. It also seems to not overlap (and not even come close) with other contests. |
|
0
It would be great if you published a "post-mortem" explaining what was the culprit.
|