I registered for TCC South America and when I switched to the tab for SRM 761, it said that I was registered for that as well. Turns out that I wasn't ... Is this only a problem with the web arena?

For real, find a sum of two repeated digit numbers?? And it's div1Easy?? Are you insane??

Problem like this maybe were probably okay in 2007 but it's 2019 now! It's freaking 2019!!!

And the more important question is who was that insane tester who approved the problem?

Argh, I'm so angry this is the most tedious and irritating problem I've seen in live contest for a looong time.

Worse than that are probably only problems about handling dates and calendar formats manually.

UPD: Here's another one instance of exactly this problem alongside with the one I mentioned a bit earlier.

Bonus for Div1 250: Minimize the number of total digits in return string. Of course, if there are multiple optimal solutions, prioritize the period "0" than "9".

Thanks for educational problem 250. Now, I have a template for that kind of problem :)

I wrote the hard for this round (and I think misof wrote the other two). Here's a brief solution sketch:

Consider the following problem: You are given an undirected graph with $$$n$$$ nodes and $$$m$$$ edges. Each edge has probability $$$p$$$ of being destroyed, independently of each other. What is the probability the graph will be connected? It's easier to solve this problem for some fixed $$$p$$$, and you can do it in $$$O(3^n)$$$ time.

The probability can be written as a polynomial of $$$p$$$ with degree $$$m$$$. Call this polynomial $$$h(p)$$$. You can compute $$$h(p)$$$ by doing the addition/subtraction of polynomials in the above dp for an additional factor of $$$m$$$ in the runtime.

Consider another way to compute $$$h(p)$$$. We can compute $$$h(p) = \sum_{i=0}^{m} f(i) \cdot (1-p)^i \cdot p^{m-i}$$$. Thus, given the coefficients of $$$h(p)$$$, we can extract out $$$f(i)$$$.

There are some small optimizations you also probably need to do in the dp. For example, you can loop up to the degree of the polynomial in the submask rather than all the way up to $$$m$$$. My final number of operations is something like $$$3^{n-1} * n + 3^{n-2} * m$$$.

Medium is also a classical problem. I remember KOTEHOK said that he read about it in Cormen and then it appeared in one of ICPC world finals that he won.

Note that in Python solution for 250 is as easy as:

from decimal import *
prec = 300
prep = 20
getcontext().prec = prec

def get_period(s):
    ans = len(s)
    for x in range(len(s)):
        if s[:x] == s[-x:]:
            ans = len(s) - x
    return ans

class AddPeriodic:
    def add(self, A, B):
        A = A[:-1].split('(')
        B = B[:-1].split('(')
        A = Decimal(A[0] + A[1] * prec)
        B = Decimal(B[0] + B[1] * prec)
        
        C = '{:f}'.format(A + B)[:-prep]
        
        period = get_period(C[-(len(C) - prep):])
        
        while C[-period-1] == C[-1]:
            C = C[:-1]
        
        return C[:-period] + '(' + C[-period:] + ')'

So much different from the hell you go through in C++...

But even here there are some caveats like the necessity to use '{:f}'.format() instead if str() to avoid zero being represented as 0E-500...

Hi, i participated in the round ( Div 2 ), how can i see the problems of both division?

There is no editorial for SRM 760 and SRM 761 here https://www.topcoder.com/blog/tag/srm-editorials/ . I am unable to find that, Can anyone provide me the link of editorials. Thanks in Advance !!