Yes, that is true. Congrats.

It was a hint to AprilFoolsClairvoyant receiving solutions from the future.

No it’s not. You forgot to add a comma (,) and c++ merged your last two answers.

Doesn't the same argument with Google results apply to standard rated contests as well? You are still allowed to use search engines after all. Would you say AI should not be banned for those contests either?

I'm curious, could you explain why you think using AI on these problems would be fun? I personally cannot imagine having fun. What's the point of solving puzzles if you're not even solving them?

Thank you. We have added it to the announcement to solve the ambiguity.

https://mirror.codeforces.com/blog/entry/133941

Furthermore, this contest is prepared for participants to have fun. What's the fun in making AI solve the problems?

If anyone does, you can send reports to sg-epk@jtk93.x29.jp

Thank you for letting me know, I added it.

I didn't really miss it though, I just have no way to know when he edits his comment (it has 9 revisions). So please send another comment for new teams.

As a responsible tester, I forgot to test this contest for 6 months.

Thank you for pointing it out. I add everything manually, so I must've forgotten to change it. I fixed it now.

I'm using the ;teamrate command on the Discord bot called TLE, which calculates the rating you would need to have a 50% chance to beat everyone in the team, as this comment explains in more detail.

I could add a column for that but I'm not sure how useful that would be. Many teams are already added without it and not all teams would do virtual or Universal Cup contests under the same name.

I cannot speak for everyone, but in my circle people seem to prefer spending their free time for team training rather than individual training. In my opinion, good individual skills cannot compensate for bad team harmony (unless you are good enough to carry the whole team, of course)

I can see it on the dashboard of icpc.global. You might have to change the year to 2024-2025 since it seems to default to 2025-2026.

A.K.A. second best girl

Here is the blog! Sorry for the wait :)

https://mirror.codeforces.com/blog/entry/141024

Think outside the box.

Since I'm involved in the national olympiad process, I'll try to give some general suggestions as far as my observations go. (These should apply to anyone, so I would suggest everyone preparing for the contests to give it a read.)

In case of time management, my biggest suggestion would be doing as many timed olympiad contests as possible. Some people mostly prepare with Codeforces, AtCoder, and similar websites. From my observations, this can cause stamina issues as you are getting used to 2-3 hour contests instead of 5 hour ones.

In terms of olympiad-style problems, it's best to make great use of subtasks. If the problem is prepared well, they should guide you towards the actual solution. You don't have to implement every subtask since it would likely be a waste time if you full-solve the problem in the end. However, only focusing on full-solve is usually a bad idea as well unless the problem is too easy. Just thinking about the subtasks and writing short summaries of your ideas for them can be quite beneficial.

But here's the thing—I’m pretty sure our olympiad doesn’t have any communication, interactive, or constructive problems.

I think it's quite risky to make such strong assumptions. I'm not saying there will or won't be such kinds of problems, but since we use CMS, there are no technical obstacles preventing such problems.

I need advice on how to deal with long contests without burning out, how to stay calm when I’m stuck, and how to keep my confidence from totally disappearing.

What I personally do for this (which might not work for you) is to drink water when I am stressed. If I still cannot calm down, I eat some food or go to the restroom to wash my face, as these get my mind off the contest. You might think that these would waste valuable time, but you waste more time by being stressed and not working as effectively as possible.

And lastly, I would suggest giving a break on the last 2-3 days before the contest and doing something fun. Motivation is always very important.

Wow. After more than 2 years of him dreaming about it (and me scamming him), the time finally came for jampm to become a Codeforces problem setter. I'm so proud of him and happy I could help with testing.

They should be visible to the public now

The implementation is extremely easy though.

You can try to write the shortest code if you want a challenge :)

Do you love April Fools?

Yes:

Definitely:

Absolutely:

Without a doubt:

Of course:

As VIP testers, this was our testing room:

Considering those problems are also in the div1, making them the same difficulty would make the div1 contest speedforces or at least not fun. Also, if they were to have the same difficulty div2 contests would still be speedforces, just with a different definition. If you make some problems the same difficulty that means participants have to solve all of them, which makes speed a higher factor than it is now.

What people normally call speedforces usually happens when there is a high difficulty gap between two problems that most div2 participants usually solve (which happens mostly with the problem D). The issue is, authors don't set the problems to make the difficulty gap like that. Even if you invite a lot of testers, it's still not easy to decide the exact difficulties of the problems. Usually, the problem order is decided based on the difficulties of the problems relative to other problems.

That's not how Codeforces decides the ratings of the problems. It's not a perfect representative of difficulty and gets affected by many things. For example, if the problem statement is too scary or it's an interactive problem (like your example), most div2 participants give up, hence making the rating of the problem higher.

Of course there are some contests where difficulty order is actually broken. However, in my opinion, it's not the case for most of the contests that get called speedforces, and making some problems the same difficulty definitely isn't a solution.

Actually, that number only takes $$$1805$$$ steps. It only works because you forgot about overflow and it's getting overflowed into negatives.

not BucketPotato

Previously known as ‘STAR Contest’, the competition attracted more than 1,600 contestants across 82 countries over the past three years. Renamed to Constructor Open Cup in 2023, this contest welcomes all regardless of age, qualifications, or skill level.

Is this the same contest as SIT & JUB STAR Contest 2022? It's been almost 1 year since I've been told that I would get a reward for it but there weren't any news.

You can solve in $$$O((m + q)\log(m) + n\log(n))$$$ using persistent segment trees. Normally it's $$$log(C_{max})$$$ but you can make it $$$log(m)$$$ by compressing values. The idea is to use persistent segment trees to find the sum and count of elements $$$\leq x$$$ from a node to root. You can do this by keeping a persistent segment tree on root and for every node creating another segment tree by updating its parent's segment tree persistently. Then, you can find those values for a path with $$$f(a) + f(b) - 2 * f(lca(a, b))$$$. When you can find that, you can binary search for the max price you'll take. This way, one query is $$$O(log^2(m)$$$. However, you can use binary search on segment tree trick. You walk on the segment tree as usual but instead of 1, you keep track of 3 pointers.

1806F2 - GCD Master (hard version)

As a tester, I must say that this contest has a problem that got into my favorite problems list. You should definitely participate in it.

Thanks a lot for sharing this with us. At first, it might look like it doesn't have much of a use case. However, it is actually helpful. For example, there was a tree DP problem in my second NOI (National OI). It was quite simple, but it required $$$O(nk)$$$ memory where $$$n \leq 10^5$$$, $$$k \leq 200$$$, and the memory limit was 32 MB, which is why the main problem was reducing memory usage. Even though other people managed to squeeze in wrong solutions because of bad test cases, I tried many different methods, but I couldn't get AC. If I knew this back then, I would've solved it easily and won a gold medal.

I have a different $$$O(n + k)$$$ solution for the problem 1799D2 - Hot Start Up (hard version).

Let's say we currently need to run program $$$x$$$ and we want to run it in $$$hot_x$$$ seconds. To do this, we can use the same CPU that we used for the previous $$$x$$$, and use the other CPU for every program between these two. If the index of previous $$$x$$$ is $$$i$$$ and current $$$x$$$ is $$$j$$$, this means $$$i$$$ and $$$j$$$ will have the CPU 1, and every program in $$$[i + 1, j - 1]$$$ will have the CPU 2, or vice versa.

So, if we know the answer to finish the first $$$i$$$ programs we can calculate the answer for $$$j$$$. To calculate the amount of time needed to use the other CPU for the tasks $$$[i + 1, j - 1]$$$, we can use prefix sums where the $$$kth$$$ program takes $$$cold_k$$$ seconds if it's different from the previous program and $$$hot_k$$$ seconds otherwise.

However, there is something we're missing here. Please try to find it on your own.

             i          j
..., y, ..., x, y, ..., x, ...
     2   1   1  2   2   1

Here is my implementation: 195175862

I tried to write it as clear as possible. However, I was in contest so if there is a mistake I'm sorry.

Please let me know if you have any questions or if there is any mistake.

Who's the cool Pikachu in problem B?

Accepted — Green

Wrong Answer — Red

Time Limit Exceeded — Yellow

Runtime Error — Orange

Memory Limit Exceeded — Purple

Compilation Error — Gray

Pending Judgement — Same as background

I tried to think of them as the color spectrum. TLE and RE are yellow and orange which are between red and green. TLE and RE can be swapped but TLE is the one we see the most and using the traffic light colors for the main ones makes more sense. Purple is next to red if we think of the spectrum as cyclic. CE is gray because it's usually not penalized. Pending judgment is the same as background to make it easier to notice when the judgment is done.

Edit:

I was thinking of them as background colors. If you want to change the text color, the background color won't work. In that case, you could use a lighter shade of gray.

Hello world, can you teach me how to become gm? Also, does being red taste better than the sea?

Jokes aside, congrats! I'm very happy you reached red, and I'm sure you'll achieve your next goal too. Just don't forget about us yellows :)

I gave him the idea that another anniversary could be in the 13th year. Anniversaries might not be following a pattern, but it's fun to think they do and try to find it.

I had 2 ideas about a possible pattern:

1) Every 5 years there is an anniversary. 5 years is a long time so between every anniversary they wanted to add another one. There are two possible middle years between two years and they chose the bigger one. With this pattern, the chosen year between 10 and 15 would be 13.

2) Anniversaries are Fibonacci numbers (5, 8, 13, 21, ...) with every 10 years being outliers (Except Fibonacci numbers that are multiples on 10). With this pattern, 13 is still the next number.

There are many different possible pattern ideas that you can find if you use OEIS but these are the ones I came up with myself which are simple enough. If anyone has any other pattern ideas that aren't too obscure, I'd love to hear them.

prvocislo orz

Who's atodo?

omg saarang round

As a tester I must say that one of the authors, namely ScarletS, has a huge skill issue when it comes to Codenames.

As a tester I suggest everyone to participate in this contest. Even I want to participate in it but I obviously can't :(

I wish everyone good luck and a good experience

I think renaming them to "Standings of Followed Users" and "Rating Changes of Followed Users" could look great. Might be a little long but I can't think of a better idea.

As a tester, I'm really happy to be a part of this contest :)

Hi, thanks for the questions.

In the fourth paragraph, we're fixing the maximum as $$$v$$$ and creating segments for every $$$u$$$. A segment consists of numbers $$$[(u - 1)(v + 1), u(v + 1) - 1]$$$. When it becomes impossible for a segment to contain any elements of the array we can stop early. We only create segments for $$$u$$$ where $$$(u - 1)(v + 1) \leq a_n$$$ which means for every $$$v$$$ we only create segments for $$$u \leq \frac{a_n}{v + 1} + 1$$$. We're essentially only doing $$$\frac{a_n}{v}$$$ operations for every $$$v$$$ (This is also written in editorial). In total it makes $$$O(a_n log(a_n))$$$ because this is the same as harmonic series * $$$a_n$$$.

but at that point, aren't you just iterating through all values in the array regardless, since you need to check for the minimum of every value at every "u"?

Like in the previous answer, the amount of $$$u$$$ we have to check is $$$O(a_n log(a_n))$$$. We find the minimum element in the segment with the help of the $$$next$$$ array in $$$O(1)$$$ so this is also $$$O(a_n log(a_n))$$$.

how we generate the "next" array in less than n^2 time

Like the editorial says $$$next_j$$$ means minimum $$$a_i$$$ where $$$a_i \geq j$$$. There are many ways to do this. You can either use lower_bound or if you want to do it in $$$O(n)$$$ you can notice that for every $$$a_{i - 1} \lt j \leq a_i$$$, $$$next_j$$$ should be $$$a_i$$$.

and how that will help us for a fixed maximum rather than a fixed minimum.

Please correct me if I understood this question wrong. We're fixing the maximum as $$$v$$$ so in order to minimise $$$v - min(\left\lfloor \frac{a_i}{p_i} \right\rfloor)$$$ we have to maximise $$$min(\left\lfloor \frac{a_i}{p_i} \right\rfloor)$$$.

For a better understanding, I suggest you to check my code in the editorial. I hope it's an easy to read code.

Yes you're correct. Thanks for noticing it. It will be fixed soon.

As a tester, I can guarantee that vanilla is beneficial for health.

As a tester, I must say that it was quite hard to solve those 1048576 interactive problems.

My coolness is what keeps them in place. Don't you see my suit?

Honestly, you should definitely participate in this contest. Problems are very good.

Also, as the only tester with sunglasses on, please give me contribution.

As a tester, please give my salary as contribution so I can be as rich as TheScrasse someday.

When I first started Codeforces, I was confused because of its many different features. So I think it's a really well-written blog and it will be very useful for newcomers to Codeforces. Thanks a lot for the guide.

As a tester, I think problems are great. I highly encourage you to participate in this contest and check out all the problems.

Tip: Having a cool profile picture like mine might help you do better :)