TCO18 Algorithm Round 3A and Parallel Rated SRM
TCO18 Algorithm Round 3A and Parallel Rated SRM are scheduled to start at 12:00 UTC -4 on July 7, 2018. You will be able to register for the SRM in the Arena or Applet 4 hours prior to the start of the match. Registration closes 5 minutes before the match begins, so make sure that you are all ready to go.
Click here to see what time it starts in your area.
Algorithm Round 2 Qualified Competitors: https://tco18.topcoder.com/algorithm/online-round-results/
If you are participating from web arena — You can switch to the parallel rated fun round in the mentioned way:
You can compete using:
- Topcoder Web Arena(Beta) - Please watch this video for step by step guide.
- Topcoder Java Applet - You can refer to this guide here to set up the applet. (Note that those who have Java 8 installed on their machine will see a security issue — You will have to add Topcoder in security exceptions in Java Control Panel. Please refer to the details in the guide here)
I have qualified for round 3 in round 2C, but my name is not there in the list. Can someone check this?
Hey Yes, you have qualified — not from 2C but from SRM 735 Round as you had by mistake registered and opened problems in SRM 735.
You are eligible to compete tomorrow :)
What is the maximum number of participants? (400 or 600)
I missed something, but what is Round 2C?
Recipe to a crappy contest: use TC interface (probably worst CP interface), set obvious easy and medium (yeey speed contest, cool right?) and a hard that is easier to solve than to prove. And you thought that's all? Extra: don't put any disconnected example for a graph problem so that people who do dfs from only one node would fail. I'm not the best, nor am I close to that, but regardless of my performance, TC keeps letting me (and probably anyone else who has a taste of decent problems) down — the problems are not as interesting and challenging as they used to be
I'm not particularly excited about 250 and 500 either (not particularly complaining either, they're just ok), but problems that are way easier to code than to prove are the whole point of TC. And it's not very hard to prove either: if the DFS-tree is all 0, then only leaves/root can be on a cycle, each leaf has a 1-edge to a non-cycle vertex and a 0-edge, so it can't be on a cycle either, the only thing left is the root so there's no cycle.
It's a pretty nice problem with a non-obvious, but not crazy solution; it would've made a good 500.
Ha, failing because the graph is disconnected is such a common mistake. I made it so many times I specifically watched out for it this time. TC samples are notoriously bad, learn to not rely on them for anything but a final sanity check before submitting. I once wrote a solution to a 250 that was utter garbage with 1+ brutal bugs and some smaller bugs — it passed samples easily and a decent part of systests too.
I have a similar feeling. Solved the easy and the medium for a decent number of points, was working on the 'Solution 1'-approach for the hard (in quadratic time), but couldn't quite finish it, got a challenge and still only placed 57th.
In retrospect I maybe should have looked at the standings about 20 minutes before the end and after seeing the large number of submissions for the hard problem reconsidered my aproach; maybe that I would've thought of the 'Solution 3'-approach then. But it is hard to change plans when you feel like you're on the right path.
So for me it also feels wrong to (almost) only let the 'do you see (guess) the Solution 3-approach' decide the advancers, but I might be biased :). Anyway, I'll retry in 3B.
Test your solution before or after submitting. You wouldn't fail hard today.
It's not that. The problem was still way too easy to be a hard. Basically all problems were: read and then code the first thing that pops in your mind. I agree that hard was a bit cute, but that's not enough to make a good problem. Actually hard was a good math problem, but not really CP one. No cycles of a color -> forest for a color -> you kind of want to work with spanning trees, so randomly take one, small degrees -> yeah probably not cycle based on transversal edges. The fact that taking only a ST of the graph solves the problem makes it be a pretty troll problem, but definitely not hard. And well here it wasn't really about testing, for some reason I thought the graph is connected (I honestly don't see any reason to not give it like that, but whatever, that wasn't the biggest issue of the contest).
'taking only a ST of the graph solves the problem`
That is simply not true, not every ST works.
Yeah you're right. Sorry. Taking a ST in the most usual way (using dfs). When referring to spanning trees, I'm also referring to those generated by dfs, but that's not a common agreement so I misexpressed
Submission times suggest that the hard wasn't as easy as you describe.
A problem doesn't have to be complicated to be hard (i.e. hard to solve). The number of submissions suggests that the difficulty wasn't terrible. It could be harder to allow some people with easy+med to pass, but still, it was fine.
The solution doesn't require anything about the connectivity, so IMO it would be even strange to add that condition. I would add to the examples, though.
EDIT: btw. my complaint is that the batch testing still doesn't show the total result and shows something like "successful test" just to mess up with a contestant. This is why I submitted a solution for easy that passed only 3-rd example test...
I had the solution but didn't trust it to be the solution for the hard (as it was quite simple, wasn't it?) so I just wasn't confident enough and looked for potential mistakes in thinking.
And about simplicity meaning difficulty. I completely agree, only that here the simplicity is actually just technicality — it's the thing you're taught to think of, not something unusual but simple, rather just direct.
As for connectivity, it's debatable. Indeed nobody would care if there was any disconnected test in the samples, but even so, in all graph problems, I find it annoying to deal with several components even in such easy cases — it's usually a matter of attention and not of skill / thinking.
Anyway, I don't think it's normal to have ~50+ people solving hard (and other tasks). In the nicest rounds the hard is solved by ~5 people (and for that reason I couldn't believe what I was thinking is true)
It's debatable but not something that you should use as an argument for the contest being crappy. You didn't like it and you would do it differently, that's ok.
Well, qualification rounds for TCO were always easier than SRM's. Early rounds can be sometimes solvable in 20 minutes.
I find it annoying to care about even and odd palindromes in problems about palindromes. But should the statements say "count only even-length palindromes"? I find the connectivity to be a similar issue.
Maybe not, but that was just an "extra". The major problem was that the tasks were too easy to select 50 people (obviously, in my opinion). I can't deny I'm a bit frustrated, but that's definitely not more than half of why I'm disappointed.
I thought round 3 is already a bit advanced.
"But should the statements say "count only even-length palindromes"?" — good one:))), I think they should.
I did and I failed. I somehow considered disconnected graph an invalid output, so it was natural I didn't test on such cases. Quite possible that the whole "simple graph" story assured me (and apparently many other people too) that there will be no unusual graphs.
I'd say the opposite. Before opening hard, I checked the scoreboard and was surprised by how fast people submit it. This also made me not doubt my solution when I came up with it quite quickly.
I find it classy when writer takes care that the set of people who come up with the correct solution and the set of people who pass the problem are as similar as possible. Focusing on even-length palindromes only is absurd, but it's very rare (and you would have to intentionally do it) to have a set of samples that let your solution seem correct if it is completely wrong on odd-length ones.
I agree this doesn't make the contest crappy though. I only may argue this makes the scoreboard a bit more random, although the variety of possible solutions for Hard already created this issue. Whether it's good or bad for a qualification round is a separate discussion.
I believe writer/tester's intended solution for the hard was much harder than those simple solutions, and so many solutions for the hard was unexpected for them. The editorial says "I’ll describe two different approaches" and describe two hard solutions, then the easy solution is attached at the end as "Solution 3". It seems they missed it and Solution 3 was added later.
The difficulty of problems were not suitable for round 3 but this is an accident rather than their intention. This kind of things happen from time to time. And there's round 3B to save people in such cases.
Why will printing the whole array into stdout cause TLE for the medium one???
Because topcoder:))
PS: It does seem like a huge issue indeed, just making a joke based on their worthless interface
I did this mistake on my very first TC match, over 10 years ago. I wouldn't expect it to change any time soon.