I got a login page, and then I tried to log in. Now, I'm getting never-ending redirections on https://www.topcoder.com/my-dashboard and https://accounts.topcoder.com/#/member?retUrl=https%3A%2F%2Fwww.topcoder.com%2Fmy-dashboard%2F. Therefore, I can't browse TopCoder's website anymore.

Are you getting the same problem ?

Just learn it, It will become the easiest service you will ever use for CP.

Take a look: fusharblog.com/apps/testerdream/

Just mail them.

I think it's DP. dp[pos][gcd(current product, K)]. DP for n — 1 elements. The last element will be chosen!

Well here you have to actually provide tree, so it is harder?

Yes there was an announcement of this.

I had an idea using inclusion-exclusion as the number of divisors of numbers uptil 100 would be very small. Didn't have enough time to code it up.

Is there an alternate solution?

Here, your biggest hint is that n is very big and k is small. Keep an answer array of k values, the number of ways you can form 0, 1 2, etc so far (mod k). Now, to compute it for a certain n, you find the matrix you need to multiply it by to transform n to n + 1 (actually, you need to construct this matrix in your program since it depends on k), and exponentiate this matrix and multiply it by the initial answer array(should be just 1's). It's O(k³ log n)

That's because people who failed systests have their score wrong(see negative scores). I went from 82nd to 73rd :D

That's because people who failed systest get negative score instead of zero ;)

This does indeed appear true, here is link to the solution of somebody to that problem, however does anybody have link to full set of problems from that contest, and / or a place to submit? I can only find standings, not problems.

Login into topcoder.com first. Then hit arena.topcoder.com

http://mirror.codeforces.com/blog/entry/48850?#comment-333916. It passed.

Let K = p₁^e₁... p_s^e_s. The answer to the problem is just the product of answers to the prime factors (CRT) so WLOG K = p^e.

Observation: if and . then answer for v = x and v = y is exactly the same.

So we just want to compute how many n-tuples has product divisible by exactly p^f.

Consider the polynomial W(x) = (p^e - p^e - 1)x⁰ + (p^e - 1 - p^e - 2)x¹ + ... + (p - 1)x^e - 1 + x^e.

i.e. coefficient at k-th power of x is number of remainders from [0, p^e) which are divisible exactly by p^k.

Now what we want is just more or less W(x)ⁿ, (powers greater than x^e should be replaced by x^e). Now if and p^f + 1 doesn't, the answer is f-th coefficient of W(x)ⁿ divided by (p^e - f - p^{e - f - 1}).

You have the center, it is the node/nodes of minimum eccentricity (there can be only 2 centers at most). Now about the ones that have eccentricity higher you can note a couple of things:

1. They always come in frequencies of 2 or more. Why is this? Think about the point where the excentricity is max. There's a path that's the maximum path and it obviously ends in a leaf so you can note that this leaf also has maximum eccentricity. For the ones in the middle of the path you can think the same way, just note that the maximum length path goes through the center.

2. A point closer to the center has lower eccentricity.

Now to construct the graph just order the input, if there are 2 centers you connect them and connect the vertices to some vertex that has a immediatelly inferior eccentricity making sure that there are at least 2 branches while checking if there's a number with freq<=1 (that means it is impossible). Check if this is a real answer (the numbers could be higher than possible for example) and it's done.

You are right, I was wondering how my rating went down, and this somewhat explains it, failed submissions are apparently as good as correct ones these days! If I had known I would have submitted the 500 and 900 for max points and had my rating go through the roof!

Yes, we are looking into it.

Sorry for the issue, it was caused by my mistake (there is a bug in checkData() for Div2-Medium, so someone submitted an invalid test case in challenge phase.)

This should now be corrected, just waiting on cache to expire so it shows up properly online :)