Hello Codeforces.
Unfortunately a few hours after the end of Round #819 we were pointed out that one of the problems was stolen from another online judge.
Codeforces team strongly condemns such an action from the author. By no means is this acceptable behavior. This is the first and the most important rule of problem making: your problem should be new up to your best knowledge.
A lot of people put lots of effort into creation of every single round. You can find these people mentioned in the announcements and they indeed deserve these acknowledgments. And yet a single cowardly decision can ruin this huge work and experience of thousands of participants.
In this particular case the copied problem was not widely known, and it seems that this made little or no impact on the course of the round. Nevertheless, we stick to the decision that in such case the round cannot be considered successful, and thus no rating changes can be applied. We will also remove the problem from the round and from the problemset as well as blacklist the author.
We hope that this case will be a lesson for future authors and this situation does not repeat in the future.
Can you share Which problem?
refer to this blog.
Making the round unrated is unnecessary, IMO.
Agreed
well, though I agree on this personally, there are still reasons to make this round unrated. I have seen many people suffer with F, both in the same room (room for hacks, not the physical room) and in the standings. Assume, in this situation, someone knew it was a stolen problem. In this case, they would have an unfair advantage! This unfair advantage is the exact reason to make this round unrated. While I am not a representative of Codeforces, I consider all such "unfair advantages" undesirable and therefore enough reason for making a round unrated.
UPD: I understand both thoughts on either keeping the round rated or making the round unrated. However, as problemsetters, we (and by we I mean myself too, I prepare problems for other platforms) have many different values to pursue. These include fairness and justice. I hope you understand how this "unfair advantage" could be unjust and understand my idea as well.
Do you think a round should be unrated if one of the problems was similar to a existing problem because of actual coincidence?
This is quite hard to decide. But really, no, in this case we can attribute the advantage to the luck and wisdom of the ones who found a similarity. Some problems can be similar, it happens. When this happens, we can say that these people who found a similarity are very wise, being able to find a connection of the problem they're solving to one they did some time ago. That is a skill, after all. However for stolen problems, there is no "skill factor" over this. (If Googling is a skill, then maybe there is, but imho it's really not) When two problems are exactly the same, people can just copy-paste one side's solution to the other. There is then no more thinking about it, they just copy-pasted, got AC, that's it. While I personally think removing duplicates "entirely" is impossible (problemsetters are humans after all), they at least need to try their best to find out duplicates as much as possible in the problemsetting process.
Ok, let me change the question. Do you think a round should be unrated if one of the problems was THE SAME AS a existing problem because of actual coincidence?
Okay, if that's the case, then why not just make it unrated. We have some things to consider in this case though, such as 1. was it easy enough to find out the original problem? 2. did the coincidence give some participants an unfair advantage? 3. was the coincidence easily preventable? if 1 and 2 are true, then making it unrated should be a good decision. 3 though, is a hard question. there are tens of thousands of problems in the CP scene, if not millions. they should have tried their best in preventing coincidences, but as problemsetters are people, they can't review the massive amount of problems spread across different judges. if 3 is true, then the problemsetter may need a (weak) punishment, maybe a warning. but it's not an easy thing to simply say either 3 is true or not, after all detecting duplicates without the source problem in mind is hard.
They even copied sample cases that can not be a actual coincidence.Codeforces has set of rules to be follow . That's what happened.I fully agree with the decision of KAN.
Last time (last point).
What about those who touched new color first time and rolled back?? Genuinely feeling bad for them.
I am one of them:-( After a week i got the new color
Me too
negative 1
plus 1
plus 1
me after getting positive delta in the contest and getting rolled back:
+94 and the round is unrated, of course. It sucks, but ultimately I agree that there has to be a strong message sent out that plagiarism won't be tolerated for future problemsetters.
Thanks for making this announcement. May I ask why you plan to remove the problem from the round/problemset when the same was not done for 1710E - Two Arrays?
Agreed: barring copyright concerns, it would be helpful to have the problem preserved so that e.g. the standings can be viewed as they were in the contest, the contest can be VCed, etc.
I also think that removing the problem from contest/problemset is unnecessary.
The difference between this problem and other actual/notorious coincidences is that this time one of the dmoj staff wrote to coordinator Artyom123 expecting an appropriate action (against setters). Ig this is one of the reasons influencing the unavailability of this problem on codeforces.
https://dmoj.ca/ there is a discord link beside "comment stream"
In #general they were discussing contacting Artyom123 about this.
It's just a grey area (I'm not a lawyer, and CF and DMOJ have their licenses wrt to problems); probably that's what Artyom (and KAN?) want to avoid, even tho they may have not asked for it.
I doubt at present Artyom (and KAN?) are going to revert back unless DMOJ staff explicitly says they don't have any problem wrt this problem being present on CF, after that people ask Artyom (and KAN?) again to add this problem back on CF.
You can try going by this route if you care too much about this problem being available on CF.
I do think this is a good decision, even if there was no impact, the behavior of the author should never be encouraged, and being unrated is also a warning to future authors.
Making the round unrated isn't important for the author who's copied the problem, Why should the rights of other authors and people who participated in the round be ruined?! Do you think it's fair?
Nice. Next time when I drop ratings, I will connect my problem setter friend to make the round unrated.
my friend got about +90 and he became a specialist for the first time and he was planning to pay for me in a restaurant, you spoiled it for me, that's not fair guys
so do i,now i back to Pupil,what bad luck.
me too :(
It's not the first problem that was copied for the contest. Neither it affected heavily the standings. What's so special about it to revert rating changes for everybody?
last time it was a Div.1 and a Div.2 separately, this time it's a Combined round. probably that is why the only options were to make it rated for all or make it unrated for all, in which the former would have been undesirable (last time's Div.1 was unrated too)
For those with low scores, they need to spend more energy and time to solve the first few easy problems in each game, but for them, the later problems have nothing to do with them , but they have something to do with them when they are leaked.Unrated is really fair?
unfair? okay, maybe. but did they have a choice? maybe not. It's a combined round, meaning that their only choice was to rate all or unrate all. the former is unfair to people who solved F and the latter is unfair to the ones who did not get to that point. do you think they really had a choice?
In your opinion, how should we deal with this theft next time? Won't stealing stop just because of one or two warnings, or do we only choose the round that can do most of the questions? The number of participants in the 1+2 competition is even lower than that in some div.2 competitions.
If Codeforces had a way to recalculate the ratings with only the stolen problems excluded, that would be much better for the future.
that's not fair we spent 135 minutes doing our best
we made a great effort in this contest we want our money back
pardon me, you want whose money back?
it's an expression sentence meaning that we want our rates back. pay me some focus bro, my life is short:)
oh. I understand you fam, I lost rating due to this rollback too.
yes bro and I don't like it at all that they have reverted the ratings just because of a single person's fault
If you do it so you need to do the same for CF1710E bro.
What a pity.
I'm sorry, but I think this is unfair. We spent a lot of time solving these problems and the first 3 were good problems. I get that this should be a message to future authors, but you said that it made little to no impact on the round. I also got Pupil for the first time ever, and now this happens.. I don't know what to say, but I'm not happy about that
I wanted to participate in the contest, but skipped. When morning came, I heard the contest has become unrated? :/
Please make it unrated next time when I drop rating, thx.
Was not the first time.I think that codeforces should have some clear rules to curb such behavior.
Making the round unrated is overkill in my opinion since so few people actually solved F. I hope this decision can be changed.
In ROUND810, div.2 also has the original question, and the result is still rated. This time there was unrated.
So why the Codeforces Round #759 (Div. 2, based on Technocup 2022 Elimination Round 3) is rated?
Was there ever a conclusion reached that the problem was deliberately copied in that case? This seems to be the determining factor — coincidences can happen, but deliberate copying is forbidden.
How do you think about my ideas about why this round is unrated and the difference between coincidence and theft? There's a lot of downvotes, but I have no idea on the reason except for "toxicity". I feel like basically this community upvotes on every comment who sympathize with their own loss and downvote the rest. For all people who don't consider diversity of thought and think only for themselves, I have only one thing to tell you, shame on you!
Yea this is common, I remember when I wrote comments about FBHC they were initially upvoted, but then later became downvoted, possibly by people who want FBHC to move to the "code runs on server" format. Hard to know though.
I upvoted.
You can see 1591F is just like ARC115E. But the round is still rated, though more people found.
Was there ever a conclusion reached that the problem was deliberately copied in that case? This seems to be the determining factor — coincidences can happen, but deliberate copying is forbidden.
Okay, you guys make the rules. You have the right to explain.
I spent a lot of time solving this round, but when I was happy that I was about to get promoted, you told me not to rate it. Although I can only solving A、B、C, it is obvious that not many people can solve F questions.
I thought you were deleting the cheaters and i will get my rating back but this time the author is cheating
oh no
me after getting bad delta
I got +13 this time but I think this decision is perfectly valid. In order to make a completely fair decision from the perspective of the administrator, it is necessary to consider even participants who wasted time due to problem F and could not focus more on other problems. This is literally impossible.
I participated in the contest with $$$102^{\circ}F$$$ fever, and now it has become unrated...
upvoted in wishes for your recovery. by recovery I mean both rating and health!
please don't make the round unrated just remove the problem and its points plz :( :( :(
Only a little participants solved the problem F,making the round unrated is unfair for other participants
Current state of codeforces after this blog.
For those who are upset about the round being unrated, please appreciate the experience you had with all the other problems. The rating is just a number that simply estimates your ability, and having one or two unrated rounds wouldn't have much impact on its accuracy. But the real value for the contest is on the problem-solving experience, honing your skill and reasoning ability as you worked through a problem (even if it was the copied problem) and tried to implement a solution. None of this is being affected by having this round unrated.
While it's unfortunate that your efforts aren't reflected in your rating, this does not, in any way, erase the actual experience of the contest and its corresponding benefits.
I just got Specialist last night,now back to Pupil
Woke up to my +150 rating gone, actually gonna commit
As usual, I did well in the contest and now it's made unrated :((
If you make a round unrated only for one copied problem (Which didn't affect someone else's rank as mentioned in the blog), Then you should do it for another rounds too, You could make rounds 805 and 808 unrated too but you didn't, That's unfair, Also IMO making this round unrated wasn't necessary, Hope asap codeforces becomes real codeforces not googleforces.
This is unfair. We spent 2 hours and 15 minutes to solve these problems. Now this round of contest has been unrated. This is not a punishment for the author but for the participants.
Woah got+101 in this contest and now it's unrated. Hard luck. I didn't quite get what's the point behind making it unrated.
So I get to reach my max rating with a 105+ in this contest and then wake up in the morning to see this? (sighs...)
I support making it unrated, but why remove the problem? This would ruin the standings and make contestents lose their submissions.
I suggest removing problem E in last round but not this time because the impact on this problem is smaller, as the majority of contestants who passed this problem haven't seen this problem before.
Is there any punishment to the fake author themselves? in my opinion stealing a problem intentionally is worse than cheating
Imo this was way too unnecessary and disheartening.
Please make this round rated.
Making this round unrated would be a punishment for the participants.
Out of 10000+ participants just 50+ have solved that problem and also most of them were red or orange.
So orange or red participants are not participants? Only you guy the green participant is participant?
No
I mean they are most probably authentic people which implies the problems were most probably not cheated.
Approx 50 people solved F. Making this round unrated is just unfair for almost everyone who participated in this contest. This actually seems to be a punishment for participants rather than author. Can't we just remove the problem F, this way it will have much lesser impact.
Wtf, that's not fair
So disappointed. This round had great problems in my opinion. It's unfortunate that the hard work of the problem authors has been diminished.
Rated div.1 contests are too rare nowadays, and on top of it, this happens.
can you explain why they are rare ?
are they running out of ideas for new problems or because the majority are in div2,3 so they don't focus on div1
Finding ideas for hard problems is difficult.
Could you explain why this has to be done as it seems like multiple people are against it? KAN
Pretty sure it is a violation of DMOJ terms of service by the author, assuming we include in material the intellectual property rights of those authors.
But that's not even the main issue I would say, do we really want to see the platform become a host of stolen problems, one where thieves get to rejoice at all the oblivious upsolvers years from now solving the problem while overlooking the fact it was copied? Deleting is the way to make it clear that any effort put into preparing stolen problems it will be wasted and thrown way once exposed, especially since legal recourse is pretty much infeasible.
I suppose one could argue for a message in the statement exposing the event with a reference to the original source.
In fact, I don't quite understand why the last copied problem has not been deleted as well along with this one. I feel like the same rationale should apply...
I think this will be a good idea.
The "Two Arrays" problem, as far as I recall, was copied from a source (Bytedance Grand Prix) which wasn't publicly available (or perhaps was but not in English) at the time of the contest (and it was also noted that a disproportionate number of in-contest solves were from people in China). I have been informed that there now currently exists a public online judge mirror here, but the user interface is all in Korean. (You can still kinda blunder your way through registration and submission by Google Translate and/or using the URLs as hints, but as a non-Korean-speaker I consider this a significantly less usable experience than DMOJ)
KAN Sir as people are heartily requesting to make this contest rated, I would like to suggest you one thing. Remove this problem from the contest.. and just give ratings based on the other problems all of us have solved.
Calculating the score just based on the other problems does not work. During the contest, participants try their best to score as much as possible in the given set of problems. This often includes strategies like looking at the standings, skipping problems that seem difficult, etc. The removal of one problem can change participants' strategies significantly.
I am good at data structure problems, so I often skip an adhoc-looking problem and try to solve a data structure problem. If after the contest it was decided that the data structure problem will be removed and ratings will be recalculated based on the score of other problems, I will feel that it is very unfair.
I'd like to know what you think will help solve this problem. After all, this competition has consumed too much energy.
I think just having the contest be unrated is fine.
What do you think if you count the people below 1900 who can't write F questions as div.2?
What do you mean by "can't write F questions as div.2"?
Sorry, there is something wrong with my expression. I mean, people with scores less than 1900 can't solve the F problem. So, can we say that most people with lower scores don't feel unfair? Then, can this group of people's achievements be regarded as a div.2 competition?
.
It's ridiculous that the author didn't dare to go online, admit this behavior and apologise to the public. This is something that deserves our despise most.
It has happened again.
No, it's generally a strange situation, to be honest. If then (Codeforces Round #810 (Div. 1)) the Chinese contestants directly knew the solution and it was clearly visible that the stolen task had a strong influence on the results, then here is the task of 2017 from some super unpopular resource, where it was solved by one and a half people. In general, what are the probabilities that at least one of those who solved the problem then participated in this contest, and even remembers when and where exactly he solved it 5 years ago? I don't see anything wrong with making the task public. Apparently, the author thought the same and I agree with him. But the community, as usual, came up with a problem and unrated the contest for themselves, and now they are dissatisfied
Dont get me wrong. Its not about "i ve got positive delta — i am maaad now". Its all about undeserved hatred towards the author
On both of my last 2 div1 rounds I had plus delta, and both of them were made unrated. In round 810 plagiarism had a strong impact, but this time the impact is NONE. I think you should not make this round unrated. And if you do, I think you should permanently unrate all of div1 rounds from this day.
No I think it should. It will serve as a warning to future problemsetters.
The warning will not involve the time and energy cost of so many people. :)
If this round is rated, it will be my first time to get Master. I can say I've been waiting for this day for five years. But the reality gave me a big blow.
Just for one problem, the contest is unrated. I think it's unfair for the 13,101 people who didn't solved this problem. And the most of us, even didn't reach that problem statement. This means that most people's efforts throughout the night are in vain. Additionally, I think the testers for this round also need to take some responsibility. This could have been avoided during testing, rather than announced it's unrated after the round. And throughout, the author little_angel didn't come out and give us an apology.
I cannot accept such a decision. I sincerely hope that this round will be rated again, KAN. Thank you.
As I see it's unrated not because some people solved it, but because this has a bad effect on the image of Codeforces.
Well, negatively negative delta this time :D
Sorry for ones who now has negatively positive delta D:
Lucky day for everyone with negative delta
I lost 70 points this round. I think im the most lucky contestant this round.
I feel that I'm so lucky because I will drop 100+ rating if this contest is rated.
Meanwhile me who lost 150+ rating T_T.
Why aren't such problems rejected in first place when rounds are being prepared/coordinated?
Finding out if a problem is a copy is very difficult and you can't avoid it completely.
Maybe there is an option for this kind of 'mixed' round situations?
Like, I'm not sure how possible this is, but my idea is to split it 'manually'. The last time this happened it was separated Div2-1 rounds. Later only Div1 was declared unrated since they weren't merged and the problem effect in Div2 was too brief.
Maybe in Div1+2 rounds you can split participants into 2 groups? Like it would have been for normal div2 and div1 rounds — Users with rating lower than 1900 remain rated and their scores are calculated separately from those with rating>=1900. For those with rating 1900 or more round can be declared unrated. I'm not sure how fair this is, but it probably is a lil 'fairer' than declaring round unrated for a problem F which was solved by 110 by now and probably even less on contest(not sure). Some people might have wasted time on it, but those most likely were 1900+ and they won't be affected either way.
Just an idea which might work, dunno.
It seems that this idea should be reasonable, the number of people who score more than 1900 is not large, and it is almost impossible to write F with a score below 1900
I think the solution is to stop making combined rounds.
Might be, but combined rounds are quite a good way to shift from div2 to div1.
I have never been in this situation but I envision shifting from Candidate Master to Master to be hard, since the only div option you got is Div1, where the best of CF gather together. You can participate in Div1 from 1900 but well, not sure, I haven't been there yet.
As a CM you also get to participate in Div. 2 rounds if there is no Div. 1 round at the same time. Anyway I wouldn't worry about such things. Either you have the skills to reliably get to and stay in Master or you don't.
Might be true, can't argue since I have never been there.
g
h
It's problem setter's fault, why unrated contestants?
totally agree
It's not fair. Please make the round rated again. Very few people only solved problem 'F'. Try to understand the situation of others who gave their best effort around 2 hour 15 minute.
Out of curiosity, what would've been my rating change if everything was normal? I didn't check CF before rollback since I was just competing while working (so I assume it'd be rating loss) and last time I used rating predictor browser extension, it wasn't very accurate. Is there a reliable way to find out?
I use this sometimes, it's not very accurate, but gives you an idea. It usually displays a little lower gain than you would have actually got.
F.E: the last time I used to check it, I got +137(in actuality, on CF) and the app displayed +(~110) or something similar.
I believe this is pretty accurate.
-63 by cf-predictor
No need to make round Unrated
All the participants with positive delta getting punishment for author's fault. Not Fair!
This had little to none impact on the contest, making it unrated is more punishing for those who worked hard to improve their ranking than for the one who committed the theft. It's unfair and i hope you will reconsider your decision.
g
kotatsugame was becoming LGM for the first time. I feel bad for what happened :( Hope you will reach LGM soon.
Thank you for mentioning! I will continue to participant in contests as much as possible and do my best.
I work for ten hours and then do cp just for rating. if I want skills, I can solve the problem on weekend. It's really a waste of time that the round become unrate. Doing cp then I only have 5 hours to sleep at night because of the work. It's totally no sense that someone said rating is not important.
I couldn't even reach the problem which was copied
This contest should be rated.
Because every contest is slightly unfair.In each contest I find some solution publicly before the contest ended.
If the problem setter made a mistake, then why contestant will suffer? Only about 100 people submitted the solution. I think most of them have solved it fairly.
This will not be a lesson for future authors because few weeks ago same thing happened on CF and the sufferers are contestants.
I hope, if the same thing happen in future contest and then it would be rated.
This way you punch us the author don't care about rating change ಥ‿ಥ
Why so many people care about their rating? You should solve CF contests for practice or/and for fun, not for rating.
says the guy who hasn't done a contest in 2 years to preserve his shiny yellow color.
This is a coincidence :) I have left CP (atleast, solving, I'm still a problemsetter) right after May, 2020.
For people in GMT+8, the contest runs from 1035pm to 1250am. Many people in east asian time zones have to purposely change their sleep schedules or endure 4 hours sleep to compete in contests. As others have pointed out, if you only care about practice, then virtual contests are same, if not better because I can compete at a better time.
Personally, this is why I compete so rarely on CF because I have to wait for term breaks to purposely shift my sleep schedule to something closer to MSK so that my brain will work in the middle of the night.
That said, although I am obviously unhappy about losing LGM, I support the decision to unrate the round.
Staying up all night really makes me tired, and the score is a part of my motivation.
Never thought about it, makes sense now. Thanks!
I set a small goal of 1400 points. In more than two months, the change of scores gave me the motivation to move forward. Although I was only green, I didn't miss a game. I'm not afraid of losing points, but I'm afraid I can't see my own changes. Now after this competition, I fantasize about the improvement of scores and the possibility of reaching 1400 points tomorrow. unrate, where a small number of talents are affected, is unacceptable to me.
Rating is what gives me the most fun
it seems you never wrote div2 round while being 2100+, contest is still not easy for you and should be fun on paper, but on practise it's not the same feeling as a rated round, there's much less fun when you know round is unrated for you.
I thought I would just sit back and watch. However, I decide to write this comment.
If you drive a car too fast, it is illegal. It is not about whether your car crash into other people or not. It is about law, which states the speed limit.
In our community we do not allow authors to copy problems. It is not about whether this affects the round heavily or not. Therefore, the round should be unrated.
While I understand many people got upset for the round being unrated, I'd say I want to point out that about 4-5 years ago there was another point (not about copying problems): there are connection problems where people could not access the contest page for a significant amount of time. (FYI, later it is solved by introducing m1.codeforces.com, etc.) What happened? The round is declared unrated.
Making a round unrated is not an uncommon action. People should be prepared that some problems might happen during the contest. It is not about whether you tried hard for the round or not. There is a significant about of trouble and that's enough to make the round unrated. That's why I'm so tired seeing people arguing whether this round should be unrated or not. I just accept the decision written in this post and that's already OK.
What I want to discuss about is that how can we prevent such a problem (about problem theft) again in the future. I think there should be some better protocols (i.e., plagiarism check software, etc.) to prevent cases like this.
Out of self-interest, I want to restore the score, but the more important thing is to reduce future illegal operations.
I have the feeling, many people misunderstand what a "rating" is. It is not some sort of highscore. You don't push it by solving more contests. You push it by learning and getting better. The rating is a measure about how strong of a CPler you are. As long as you participate in some contests your rating at some point will reach your skilllevel.
There are statistic fluctuations of course, and by solving more contests we expose ourselves more to them and can probably get a higher high-score-rating, but in general that is not what a rating wants to measure.
Be happy you participated in this contest and had fun with great tasks! I for one learned much from Problem E and I had much fun with C and D.
I'm genuinely sorry for the people who only seldom find time for contests though and had to make time. Better luck next time!
But that was +99 for me :')
+110 for me so sad
Good thing you guys are gonna blacklist the author. A strong message needs to be sent out that such activities will not be tolerated. I saw the author put out an apology blog yesterday but has now deleted it. Guess he doesn't feel sorry for his deed after all.
I noticed that further submissions to CFR819-F are not accepted now. here
I hope completely remove the problem (and keep the statement, ranklist... somewhere else if it's needed) rather than keep the problem slot, so that innocent future upsolvers can get the green lamp (all upsolved).
P.S. How about CFR810-D1E?
Since I would have lost a lot of rating (due to FST), I strongly agree with this decision! :)