Hey All!
UPD: Editorials https://www.topcoder.com/blog/single-round-match-764-editorials/
Topcoder SRM 764 is scheduled to start at 12:00 UTC -4, August 10, 2019.
Please Note: This SRM will have a common problem set for both Div I and Div II. The difficulty level will be overall a bit easier than usual.
Registration is now open for the SRM in the Web Arena or Applet and will close 5 minutes before the match begins.
Good luck to everyone!
Topcoder Java Applet | Next SRM 765 — August 23
Some Important Links
| Match Results | (To access match results, rating changes, challenges, failure test cases) |
| Problem Archive | (Top access previous problems with their categories and success rates) |
| Problem Writing | (To access everything related to problem writing at Topcoder) |
| Algorithm Rankings | (To access Algorithm Rankings) |
| Editorials | (To access recent Editorials) |
Commenting to bring this to top. Reminder: 2 hour 10 minutes to start!
Wow, this table of links is amazing, I didn't know it is possible for any human to find all these places. This blog goes to bookmarks xD
:D
Still unorganised but they are right here on the top of the page. https://www.topcoder.com/community/competitive-programming
We will be redoing this page in the next few months to look something like this. Still working on it, This is the v1.0! There will be many iterations to make it better.
That's fourteen pictures of tourist in one place xD
Sixteen pictures here
Maybe this is a stupid question, but how do I view the division standings in web arena? It's possible in the java applet, but the web arena tells me that "the leaderboard you are trying to view is currently unavailable. Please try again after the match is complete."
What am I doing wrong? I think I somehow managed to make it work a couple months ago but I can't remember how I did it anymore. :(
What the fuck was that?
Giving five tasks with big amount of coding is not a good idea for a 75-minute contest. Especially the one with paths and the final DP with given table.
I think there were only two tasks with big amount of coding... Yep, the ones you mentioned.
Just when you think SRMs can't get worse ¯\_(ツ)_/¯
Are constants in last problem random or do they have some logic behind?
It's some board game where two players roll a die multiple times, I don't remember name though.
Seems to me that it is a slight variant of [Risk](https://en.wikipedia.org/wiki/Risk_(game))
I don't think that this round was totally bad. I really liked 300-pts and 500-pts. And also, 650-pts was fair.
I recommended 300-pts because it is not easy to see that the simplest condition of being rectangular is that "angle of four vertices are 90-degrees". Some submitted algorithm (judge if the opposite side vectors are same) failed in rhombus case.
And for 500:
$$$a^2 + b^2 + c^2 + d^2 = (\frac{a+b}{2})^2 + (\frac{a-b}{2})^2 + (\frac{c+d}{2})^2 + (\frac{c-d}{2})^2$$$.
Thanks for contributing to the problem set, but for the future it would be good to take into account the feedback provided here.
First, "not being totally bad" is not the quality one should aim for. :)
Second, the other tasks that you did not mention bring certain frustration, which makes it hard to enjoy 300 and 500. In (such short) contests the psychological component of participants is quite important. Also, 300 is Googleable (I Googled mine). I do agree that 500 is a cute one (unfortunately, I got the intended idea only during the intermission). But, 1 (or potentially 2) nice tasks out of 5 does not seem to be a great tradeoff.
EDIT: To make more precise my comment about the psychological component: One of the main reasons why tasks are nice is because there is "aha moment" in the solution/idea. But if one sees three tasks where it certainly does not feel like there is any aha moment but rather the opposite, why one would expect the fourth task to be much different?
Right. I am blue, I failed B but I have no motivation to go see the editorial or upsolve it. Earlier when I failed Div1 A, atleast I would try it afterwards.
Possibly another good way to judge a problem: "Would people want to upsolve it if they fail in the contest?"
BTW, is the last problem inspired by a board game? Looking at the denominator in the probabilities (36, 216, 1296, 7776) it looks that the result of the attack if obtained by throwing from two to five dice.
With misof busy this week with IOI, we decided not to postpone and go with a simple experimental round with a mixed problem set. We were planning a round like this since long to see how it goes with combined divisions.
I cannot comment much on the side of the problems. However, we compiled a quick set to run some easier/traditional problems and kind of also announced that problems will be easier. (Probably should have mentioned the difficulty better). My Bad :( OR should have called it an experimental round.
As you know, good really-hard problems aren't easy to make and are only ever enjoyed by a few people, so it's often a limiting factor.
I will pass the feedback to the problem coordinator and will try to avoid implementation heavy problems. Also have more testers test and give feedback if we do something like this going forward.
Just a suggestion: consider creating div2 only srms. Like on cf, div2 only rounds. Because this was pointless.
Even being blue it felt like a punishment to implement those silly paths. I can only imagine what reds must have felt.
I love 500pt problem. Why admin use this problem for such a contest?
It was a fun experiment. Perhaps a couple more iterations, and collecting feedback from the community, can turn it into an actually viable format.
Personally, I'd like to see more ideas and less implementation in the problemset. The implementation-only problems, such as the third one, are irritating for most participants, so they don't go very well with the format where the timer is ticking only for opened problems.
Yea that went as I expected.
Look up "four squares sum", find a theorem that it's possible to represent $$$N$$$ as a sum of 3 squares if $$$N = 4^a (8b+7)$$$. Ok, let's pick $$$w$$$ as large as possible such that $$$N-w^2$$$ satisfies this and bruteforce the rest. Most likely, $$$N-w^2$$$ is $$$O(\sqrt{N})$$$, so it works.
UPD: Rofl, Markdown edited "500. My solution." to "1. My solution."... yeah, that's for the 500-point problem, not the 1-point problem.
Welp, I probably should have given it under $$$a,b,c,d \leq 10^{18}$$$.
Just thought the problem is too easy for this to matter...
Yeah, let this be a testament that the approach with lower limits to trick people and hide the intended complexity doesn't always work — in fact, it can trick people into finding a different solution.
My approach is still more time-consuming than just immediately making the required observation, but at least it means there's no need to wonder if there's a nice solution or what.