Google Code Jam 2020 has been announced. Here is the schedule:
- Registration: March 3, 19:00 UTC ‒ April 5, 2:00 UTC.
- Qualification round: April 3, 23:00 UTC ‒ April 5, 2:00 UTC.
- Round 1A: April 11, 1:00 UTC.
- Round 1B: April 19, 16:00 UTC.
- Round 1C: May 2, 9:00 UTC.
- Round 2: May 16, 14:00 UTC.
- Round 3: June 6, 14:00 UTC.
- Finals:
August 7, 10:30 UTCAugust 8, 13:00 UTC.
The finals will be held in Munich, Germany ONLINE, the first Code Jam finals in mainland Europe.
The dates of other Google competitions, Hash Code and Kick Start, has been announced as well.
Unfortunately, there is no announcement regarding Distributed Code Jam, which probably means that the competition is dead. UPD: Google confirmed that there will be no DCJ in 2020.
UPD: Code Jam finals will be held online this year due to the COVID-19 pandemic.
Another trophy for tourist, congratulations!
Thank you! Probably, it is a good idea to add the rounds into https://mirror.codeforces.com/calendar
No DCJ second time in a row.
what's the minimum score needed for advancing from qualification round ?
AFAIR eps > 0As Swistakk noticed previous information was misleading. You actually NEED 30 POINTS TO CLASSIFY.
.
That's dangerous misinformation. I would be at risk of not advancing because of you :P
"Unlike every other Code Jam Round, there is no limit on the number of potential advancers from the Qualification Round. You will advance to Round 1 if you earn at least 30 points, regardless of your penalty time. You may wish to review Visible Verdict Test Sets versus Hidden Verdict Test Sets and optimistic scoring; it doesn't hurt to score some extra points just in case!"
You can read more at the Code Jam FAQ
Can anyone tell me how to prepare for Code Jam? I had qualified Round 1B in Code Jam 2019, yet could not do a single problem in Round 2 (not even the brute force partial test set) :(
That tells you that you need to practice precise implementation.
That's certainly true, but that wasn't what went wrong. I couldn't come up with any idea to determine if the permutation was correct in https://codingcompetitions.withgoogle.com/codejam/round/0000000000051679/0000000000146183 (which was the simplest problem in the round as per no. of submissions).
This was a problem where I was completely unsure how to proceed. I am usually stuck in a scenario where I know the brute force but can't optimize it. Being able to try all 6! permutations and still having no idea, is another level of embarassing.
I want to do more such questions, but maybe gradually increasing in difficulty.
[Deleted]
I hope you are joking, since it's pretty obvious (even if you don't open the link, just by looking at when I posted my comment) that the problem I mentioned in my comment is from last year's Code Jam (2019), not this year's.
lol sorry for that my bad, deleted.
the round will start in 2 mins
don't forget to take part if you can .
In case some of you also haven't noticed because of the amazing design (I write this because not only I didn't notice)
The round has 5 problems, not 4
Ok so I modified the Codejam CSS I used last year, here you can find the new CSS code for use in Stylish this year: https://pastebin.com/jd3jGPmL.
Improvements:
Feel free to modify it.
P.S. as far as I understand, they've deliberately hidden the last problem (.hideDesktop in CSS) on all devices wider than 1175px, it is not like it is not shown because there is no space.
I used last year style of Mukundan Senthil (I don't remember where I got it) and had no problem with it. Now I modified it a bit and now it isn't so awful as it was
It should have some bugs, feel free to comment
Install with stylus/tampermonkey, you know it better than me. CSS
Is it just me or is the interactive runner still creating mental pain?
I'd say it's also creating physical pain.
for linux
no output = AC
if you want to see your output:
What the hell, guys?
I don't like nor google neither codejam, but they provide testing tool for interactive problems (nice ones, IMHO) and it works perfectly and lets you to debug locally without ANY actions from your side. Nobody else provide interactors, you have to code it yourself. Usually it is easy, but takes time and has tedious implementation
For running it simply download testing tool and then:
And in another console
To see any output just print it in stderr. As it was mentioned earlier no output of interactor means AC, otherwise it prints WA and message what went wrong. Just run several times it and submit it to gcj
I think that windows also supports pipes, but has no ideas how to do it. Or just use wsl
10 mins to think of solution to Interactive problem. An hour of struggle to try submit it. I've given up. Hopefully can cross few of the future rounds too without touching the interactive problems.
"print Case #x: y"
why make us print the case number ? why...
They are influenced by codeforces recent rounds
Likely to make it easier for newbies to analyze their output, and not to make them confuse between input and output when they paste a test with multiple test cases. (I'm against this though)
AFAIK, they started consistently using the "Case #" part back when you still solved the test cases on your own machine and then submitted the outputs, and (one of) the main reason(s) was that newbies were submitting all kinds of junk instead (programs, binaries, arbitrary numbers). The absence of the case labels in the submitted file was a good heuristic they could use to tell the contestant that they are probably submitting the completely wrong file.
As one of my favorite sayings goes, "tradition is when you keep doing something for so long that you forget why" :)
Is there any plans to support C++17/C++2a?
I think this is an OK place to ask: I am currently at score $$$42$$$. There is no possibility of the score displayed decreasing right? I am too tired and sleepy to continue and will look into the problems later if I am sure that my current score will clear the cutoff of $$$30$$$ points.
Yes, you will qualify the round.
sorry
What the hell are you doing? This is an ongoing contest. If it were up to me I'd instantly ban you for this.
Edit: Apparently collaboration is allowed, but sharing codes is discouraged, so still don't paste codes.
It's OK to do this for the qualification round.
It's my bad
Fair enough, apparently collaboration is allowed — my bad on that one. Rules still ask people not to publicly post codes, so still a good thing he removed it.
Any tip on 3rd one
Think of a greedy, keeping in mind that both people are indistinguishable, i.e. they can both do all jobs.
Approximately what CF rating should one have to advance to Round 2?
Round 2 would probably require a good knowledge of algorithms and fast implementation as only 4.5k participants get selected. So, maybe 1600-1800+
4.5k? Do top 1.5k people from round 1a not get counted for subsequent round 1s?
Top 1.5k selected from each round 1A, 1B and 1C. That's 4.5k for Round 2. Top 1k from Round 2 selected for Round 3. Top 25 from Round 3 selected for WF.
No, I mean suppose the top 100 people from round 1A , again participate and get into top 100 again, will they be not counted in 1.5k for round 1B?
AFAIK you cannot participate in the further round 1s after you advance in one of them.
Oh, I didn't know this. Thanks for informing.
I couldn't solve the last question. Can anyone tell how to solve it after the round is over?
I solved the first 4 questions correctly
, good for you now stfu.Also, thank you sister_of_contestfucker for this nice, polite and inspiring comment.
Wow, they haven't changed anything since last year. Competing on that system is so painful. :(
What are your constructions for E?
I got accepted with a merge of 4 cases:
1) The obvious cyclic shift construction for $$$k = n * p$$$ for some $$$p$$$.
2) It's easy to prove that for $$$k = n + 1$$$ and $$$k = n^2 - 1$$$ we have no solution.
3) A construction for $$$k \in [n+3; n^2-3]$$$ based on picking three numbers that will appear on the diagonal $$$x$$$, $$$1$$$ and $$$n-x-1$$$ times.
4) Weird Monte Carlo (e.g. rejection sampling like approach) with random matchings for $$$k = n+2, n^2-2$$$.
As I'm pretty sure that's not intended I would love if someone can share a clean solution.
By Hall's, if there is a diagonal where no number appears exactly N-1 times, then you can construct a grid with it.
Therefore, I generate the lexicographically smallest valid diagonal with recursive backtracking. After that, I forcibly place every number if some instance of that number appears on the diagonal with bipartite matching. After that, I place the remaining numbers in increasing order.
Can you explain these parts? I must be not seeing something as at least from my perspective, different values on diagonal should be treated as various commodities types, so I don't get how you can apply Hall here.
I think seeing this problem should help motivate what I'm trying to do here, except instead of filling the grid one row at a time, you fill the grid one number at a time. The application of Hall's on that problem should hopefully be more obvious.
Also see this problem.
I did some casework to fill up the diagonal with at most 3 distinct numbers and then used flow to generate the positions of the other numbers.
Heh, that is problem that I gave my students yesterday as a homework on a course that I teach xD (or to be more precise — its mathematical version). So you can assume I knew that one xD. But since you say that is similar, I will just think about it by myself tomorrow, cause it's 5AM here and my mind is not that clear.
If it helps, the editorial that is released for that problem contains a proof for why Hall's applies here.
The claim I'm asserting consequently is that inserting the numbers row by row as opposed to number by number are isomorphic operations.
Nice! Thanks :)
Although I feel like they left out how they treat the already initial values. This seem to prove only the case for empty rows, but in this case we have rows already filled...
Hey, can you maybe explain how you can prove this by Hall? I don't get the why for every subset of numbers there is at-least this much matching cells in a specific row, after filling each row behind it.
I think that is close to intended except for the case with $$$n+2$$$ and $$$n^2-2$$$ that you clearly messed up. I have something similar, but a bit different construction for three numbers (frequencies in my code are n-2, 1, 1) and I have also a case when there are two numbers appearing $$$c$$$ and $$$n-c$$$ times for any $$$2 \le c \le n-2$$$ (but $$$c=2$$$ will serve your purpose) which is significantly easier than the case with three numbers.
EDIT: OK, seeing the explanation of xiaowuc1 I don't think that caseology was intended solution :P.
I had the same construction for cases 1-3. I couldn't figure out 4. Also, another thing to note is you can't represent all numbers in $$$[n+3, n^2-3]$$$ by picking three numbers that appear $$$x$$$, $$$1$$$ and $$$n-x-1$$$ times.
For eg. there is no such representation for $$$n = 4, k = 10$$$.
Good news is all these numbers can be represented by picking $$$2$$$ numbers, one appearing $$$2$$$ times and other appearing $$$n-2$$$ times. So, you can use the square obtained from case 4 here.
I think your code is automatically doing this, so it should be fine.
6+2+1+1=10? UPD:sorry, I may be blind
Oh you are right. I might have written my brute force wrong then. My bad.
Edit: Oh no, my code is correct. 6+2+1+1 is wrong because if n = 4 you can't use 6.
For me randomized greedy swaps worked well for almost all cases. I started with a random cyclic shift matrix with 1s on the diagonal if $$$k < n^2/2$$$ or with $$$n$$$s if $$$k > n^2/2$$$ and swapped greedily, minimizing the difference of the trace with $$$k$$$.
Surprisingly, it did not work for $$$k \in {n+2,n^2-2}$$$ when $$$n$$$ is odd. Had to come up with a construction for that. Also the impossible case is $$$n=3 k=5$$$.
If we use $$$T$$$ distinct values on the diagonal and each of these values has multiplicity at least $$$T$$$, then each square determined by a segment of identical values on the diagonal can accommodate the remaining matching for all the other $$$T - 1$$$ values, by a simple construction (basically each time we start out with a square of side at least $$$T$$$ with only one diagonal filled and we add $$$T - 1$$$ more diagonals in cyclical fashion).
It turns out that the values of $$$K$$$ that cannot be represented in this manner exist only for $$$N <= 8$$$ $$$(*)$$$ and those can be brute-forced $$$(**)$$$.
This solution has no pen-on-paper caseology, but $$$(*)$$$ and $$$(**)$$$, although intuitive to me, were leaps of faith and had to be tested, so the code is probably significantly longer.
Even though I got more than 30 points(68 points).It shows a cross in front of Round 1A in the schedule in the qualified section.Is this a bug?
Oh thank God ! I freaked out thinking it's just me
No It will be updated within a day. Last time it happened for me in Round 2 and got updated within a day.
I see on my page that I did not qualify for next round, even tho after finalized rank list. I have a score of 75. Why is this happening?
Can anybody please tell me where I went wrong for 3rd problem? I got a WA everytime. The sample test cases are working fine. Thanks in advance. //package main;
import java.util.ArrayList; import java.util.List; import java.util.Scanner;
public class Solution {
public long getStart() { return start; }
public void setStart(long start) { this.start = start; }
public long getEnd() { return end; }
public void setEnd(long end) { this.end = end; }
private long start; private long end;
public static void main(String[] args) { // TODO Auto-generated method stub Scanner sc = new Scanner(System.in); long test = sc.nextLong(); for (long t = 1; t <= test; t++) { long N = sc.nextLong(); List setJ = new ArrayList<>(); List setC = new ArrayList<>(); String ss = ""; boolean isImpossible = false;
for (long i = 0; i < N; i++) { Solution sol = new Solution(); sol.setStart(sc.nextLong()); sol.setEnd(sc.nextLong()); boolean J = setJ.stream().anyMatch(item -> { if (sol.getStart() >= item.getEnd()) { return false; } if (sol.getEnd() <= item.getStart()) { return false; } return true; }); if (!J) { setJ.add(sol); if (!isImpossible) { ss += "J"; } } else { boolean C = setC.stream().anyMatch(item -> {
if (sol.getStart() >= item.getEnd()) { return false; } if (sol.getEnd() <= item.getStart()) { return false; } return true;
}); if (!C) { setC.add(sol); if (!isImpossible) { ss += "C"; } } else { ss = "IMPOSSIBLE"; isImpossible = true; }
}
} System.out.println("Case #" + t + ": " + ss);
} sc.close(); }
}
Rather than down voting my comment is there any one who can really help out. Please?
I don't think anyone would want to read that code. At least put it in pastebin or something
What was the idea to solve the 4th problem (the interactive one)?
I scan through symmetric pairs $$$(i, N + 1 - i)$$$. Note that if the two values are different, they stay different after any allowed transformation, and if they are equal — they stay equal. So I simply query the pairs and save the rsut, and then at 11th, 21th, ... queries I recheck one of the "equal" pairs if it has changed (1 query), and if it has, I update all the known "equal" pairs. Similarly is done for "different" pairs.
I am sure many of you must have noticed this but for those who didn't:
The name ESAb ATAd is Data Base after reversing the letters and taking "complement". (If we take complement to mean switching cases).
This is similar to how a similar problem last year was named "Dat Bae" which is "Data Base" missing two lettets a and s. And that problem was about losing bits in data!
Need help in the last problem,
It turned out for me that there are only 3 types of matrices needed. Where in the diagonal : 1) x appearing n times. 2) y and z appearing 1 time and x appearing n-2 times. 3) y appearing 2 times and x appearing n-2 times.
1) is simply shifted identity permutations.
2) is just swapping first row and second row,then we can swap 2 digits to get intended diagonal.
3) I did not found any direct solution. (btw we can easily do it when n is even.)
Does anyone know a constructive approach to build the matrix of type 3 ?? Thanks.
Edit : By constructive approach I mean a direct pattern based solution which we can perform by hand. I am not good with flows.
I found such a construction for type 3 (assuming n is odd). It works as follows when n is odd and k=n+2:
In all steps the subgrid is circular, i.e., after the last row/column comes the first row/column. For example, the construction for n=7 and k=9 is here:
This was brilliant !! How did you come up with this approach ??
Thank you.
And thank you for your wonderful book Guide to CP. I started my journey from that.
Great, thanks! I assumed that there has to be a construction where the main diagonal has n-2 1's and two 2's, and then tried many things using pen and paper until I found this.
Please Help! i have done a lot but unable to figure out :'(
I am trying to solve for D just for TestCase1 — only 10 bits are there (B=10) so we should just query each and output that. But why i am getting TLE even if i am exiting the code when i receive a 'N' from stream.
i did this too but still TLE.
Leave it! I resolved by my own... Ahhh!
What was the problem?
If you look at the 3rd last line of solve function of my code... I have written -
Here was the problem. I was writing the correct output but was not going to the next line and hence it was unable to read "Y" or "N" even if I was reading it so waiting for it and results in TLE. So I must go to next line in order to read "Y" or "N".
Here is the correct one
Ohh, won't work for me(
Ok, thanks
I have the same problem, only attempting to solve testcase 1. If i play the judge manually, it works. Why do I get TLE?
range(b) gives 0 to b-1 if I am not wrong. But you have to query 1 to b so you are not even demanding for 10th char and so judge response will never be N and hence leading to TLE.
Thanks, that solved it!
Problem : Parenting Partnering
Verdict : WA
can anyone tell me what's wrong? If i calculate whether jaime or cameron are not free, results in an "IMPOSSIBLE" or Otherwise assign activity to whosoever is free.
...
why don't you put it inside the sopiler tag?
Analysis for the last problem says that there is a solution for all $$$k$$$ except $$$n + 1$$$ and $$$n^2 - 1$$$, but if we take $$$n = 3$$$ and $$$k = 5$$$ then only possible diagonal values are 1 2 2 and 1 1 3 which clearly does not work. Am I missing something here?
From the analysis:
In N=3 I think there are some exceptions you need to handle(like K=5, which is like you said only possible with 1 1 3 / 2 2 1), otherwise if (n-2) > 1 (which is True for any N>3, you can prove it.
Hi there,
I have a question about "Indicium". Is the following part of the editorial correct?
Now that we know what the diagonal looks like, how do we actually find a Latin square that has this diagonal? To do that, we will fill in the unfilled cells row by row. We will use bipartite matching to find a valid row. In one bipartition, we have N vertices for the N cells in that row. In the other bipartition, we have N vertices for the N numbers that can be placed into the cell. Make an edge between the cell vertex on the diagonal and the number vertex that was decided on. For every other cell, make an edge between a cell vertex and a number vertex if that number can be put into that cell without breaking the Latin square properties.
For example, N = 6, K = 8. This is my initial matrix.
Then, as the editorial mentioned, I used bipartite matching between N numbers and N cells row by row. However, I ran into the following case.
I can only fill 6 into * but these * are in the same row.
In the end, I used bipartite matching between x and y with edges which were made by empty positions (x,y). Then I put a number into matched positions. It worked but I guess it would be a different idea...
If I understand correctly, they're fixing only(!) the values on the main diagonal. But you are fixing some additional values which are clearly not on the diagonal.
Hi Vladyslav, thanks so much for your advice.
I fixed numbers on the diagonal. However, a similar issue occurred.
N=5 K=23
It is already impossible to put 4 into *'s row.
From what I understand, Hall's Theorem guarantees that you will always succeed if you fill the square in one of these ways: a) complete row by complete row, or b) entering all instances of a single number at a time. (These are actually isomorphic methods, different views of the same mathematical objects.) However, you can't necessarily mix and match these approaches, and you can't do it row by row with some numbers fixed ahead of time.
So if you start with:
You can then place all the 1s, then all the 2s, then all the 3s, and it will work.
Oh, well, I misunderstood the editorial. But now it makes sense. Thanks a lot!
Can you elaborate on why Hall guarantees that?
See slides 5, 6 and 7 here http://www.math.caltech.edu/~2015-16/1term/ma006a/28.%20Latin%20Squares.pdf
But this only shows it for completing it row by row (which is not what they're doing in the editorial since the diagonal elements are already fixed).
It's isomorphic. ie, instead of choosing N values to place into row 1, you're choosing N heights at which to place each of the 1s.
Thanks! although this lack the last couple of slides in which he explain where it is true for a already filled matrix :P
Also he left the title אלגוריתמים, which is Algorithms in Hebrew. Apparently some Israely guy is giving this course and he forgot to change the title for Cal-tech student Haha
I mean, it's not like I had any clue about this during the contest. I only learned about it from reading the editorial. :)
But I think the idea is: suppose you've already placed numbers 1..k, so each row/column has N-k slots free. Take any set of i rows, and suppose there are j total columns that have free slots in those rows (the union of column choices, which is what Hall's theorem cares about). Then, counting by rows, we must have exactly i*(N-k) free slots in that i x j submatrix. But counting by columns, there can be at most j*(N-k) free slots in that submatrix. So j >= i, and Hall's theorem is satisfied. Thus there's always a perfect matching that allows you to place all the "k+1"s.
Just to help visualize this, suppose N=8, k=6, and you try to construct a case where i=4, j=3. You can try to fill in the rows thusly (# is a filled slot, — is an empty slot):
This looks fine if you're only considering the rows, but when you look at the columns, you can see that the j=3 rightmost columns have i*(N-k)=8 empty slots, more than the j*(N-k)=6 they're allowed.
Thank you so much! I get now why Halls applies here at last <3
Single question that still bothers me, is that while what you explain is true in the general case, in our case when filling the numbers we have another constraints on where in the diagonal they are supposed to be.
So after filling the first K=2 element of the diagonal in the matrix, we get something along this:
Now the first row have 4 empty locations, but the forth has 3. So when conducting Hall for 3, the rows available are {0,1,3,4,5} and the columns available are the same, but some columns appear on edges N-K times, and some N-K-1. The same for rows.
This means that when selecting a set of rows you get between (N-K-1)*I to (N-K)*I, and then there is at most (N-K)*J. So there might be a situation in which I > J.
What am I missing?
But how do you get the initial matrix? The editorial shows how to get the numbers on the diagonal but how do you fill in the 4s and 5s (in this case) that are off-diagonal before you start the "entering all intances of a single number at a time" method? You cannot prove that bipartite matching is going to work for those since the 4s and 5s are already partially filled, right? Can you use backtracking, is it guaranteed to be fast enough assuming we have at most 3 distinct diagonal elements as they do in the editorial?
Right, Hall's Theorem doesn't guarantee that you can place the rest of the 4s and 5s after fixing them on the main diagonal (and in some cases you actually need to place three distinct numbers). But it's not nearly as hard to place these initial numbers as it is to fill in the rest of the square. Like you said, you can run a backtracking search and observe that (except on obviously impossible cases like 5,5,5,5,4) it always runs quickly — even though you haven't proven it. Or you can figure out a general construction by hand.
Thanks, that makes sense. And also thanks for the screencast, I didn't think I was "just wasting my time" ;)
I looked at xiaowuc1's AC solution, and his idea is to only consider the lexico-smallest valid diagonal (except those with one number appearing exactly $$$n-1$$$ times) and try to fill the rest of the grid number by number, first taking care of the numbers that appear on the diagonal. If I understood the above discussion correctly, this should not always work right (or at least it needs some more proving)?
In stress-testing, my solution RTEs on the case N=7, K=32.
Can anyone help me to find the bug of Problem C?
You do not keep track of the initial ordering (after sorting, the order of events may differ from the order in the input). In your code, you store the index
i
of every pair, so instead of concat, you can writess[ind] = 'J'
orss[ind] = 'C'
respectively, whereind = v[i].second.second
.Thank you real.emerald.
I don't know if this is the right place to post those but I couldn't find any better place where I will find so many relevent people.
As we all know Google Code Jam 2020 Qualification Round Ended. This is my first ever Google Code Jam and I don't have much information about that also. Currently Its showing 42points but I am confussed that its the finalised scoreboard or not. I have uploaded screenshoots of Code Jam Website and it seems like they haven't finalise the scoreboard but on the other hand I have seen so many posts on all social media platform people are saying that they are through to the next round.
Can Anyone please clearify this?
Thank You
Probably, there's some plagiarism checking or similar thing going on and thus the results aren't technically "finalized". But according to the rules you have qualified.
I found this one https://vstrimaitis.github.io/google_codejam_stats
Slides with constructive solution for E: https://github.com/Errichto/youtube/blob/master/GCJ/2020/qual/E-slides-construction.pdf
Video explanation for all 5 problems: https://www.youtube.com/watch?v=KbXk_-M0kw8&list=PLl0KD3g-oDOG7cBqmx0NMqmemM5uDohwH
For problem D, can someone explain for me why just querying once could find the answer? It makes no sense at all since after the first query, the array already changes.
the array changes just before you get answer to the first query so you can assume that the first change is after 10-th query.
Gentle Reminder: Round 1A starts in 12 hours.
Can I participate in Round 1B if I did not participate in qualification round and 1A?
If you haven't participated in qualification round then you are not qualified to Round1. However, for those who qualified in Round1 and they did not take part on Round1A, they are allowed to take part in Round1B.
I just received the following mail: (TL;DR: Codejam World Finals will be online)
"Hello, We wanted to provide an update regarding the Code Jam World Finals. We‘ve been closely monitoring the coronavirus (COVID-19) situation and after very careful consideration, we’ve decided the Code Jam World Finals will be entirely virtual this year taking place on Aug 8 at 13:00 UTC.
While we’re disappointed that we won’t be able to connect in-person in Munich, the health and safety of all finalists and volunteers is our top priority. We're still working out the details, but we'll be in touch again soon.
Please note that Round 2 and Round 3 will proceed as planned during their scheduled dates/times.
We will keep our Code Jam website updated with additional information. If you have any questions please contact codejam@google.com.
See you in Round 2, and best of luck! - The Code Jam Team"
I hope that this means that they can consider slightly increasing the number of finalists, to maybe say 50 like before!
"I hope that this means that they can consider slightly increasing the number of finalists, to maybe say 50 like before!" — that would certainly be an interesting idea
Or let's go back to 2008, where 100 finalists took part in the onsite contest!
So that now I will be outside top-50 instead of top-25? No thanks, being outside top-25 is embarassing enough.
There's also a possibility to invite top (x-1), where x will be your position in round 3.
Judging by current CF ratings. E(X)=2.
But to be honest, I believe that the number of finalists won't be changed, because the rules for this year were already established.
If you try real hard, I believe that you can think of another established rule of the competition that was broken because of covid-19.
True, but there is a difference in changing the rules that in current circumstances are impossible to hold.
Maybe that is true. But I see value in changing the number of finalists, considering that Google invites top 100 contestants to onsite even for Codejam to I/O for women, and that field is definitely weaker than the open Codejam. Now they don't even have cost constraints, which should make it much easier for them to have more people participate. (If there are issues with cheating, rules such as in the Magnus Carlsen Invitational in chess can be established — webcam all around the coding arena keeping everything clearly visible, etc)
I think it will also make for a much more interesting Round 3 — last year, I see that 500 people had non-zero scores, with only 598 people participating out of 1000. That means that many people like me probably scheduled something else on their weekend which was more worthwhile than a long shot at top 25.