I was just in a clash with user neohr71 in which they submitted in 25 seconds for a puzzle that was not hard, but took at least 20 seconds to understand and then minimum 30 seconds to code (even for an absurdly fast coder).
a user who pasted this in 44s :
(no doubt : the automatic code is deleted, and this one is with comments)
import re
line = input()
# Remove the comments from the line
line = re.sub(r'/\*.*?\*/', '', line)
# Count the number of digital characters
count = sum(1 for char in line if char.isdigit())
print(count)
did you have any other answers, like from the staff ?
Another likely cheater, user Derbazi_Mokhtar, submitted the below code in 50 seconds to a problem about abbreviating phrases in ā< >ā.
import re
message = input()
def replace_match(match):
words = match.group(1).split()
acronym = "".join(word[0] for word in words)
return acronym
abbreviated_message = re.sub(r"<(.*?)>", replace_match, message)
print(abbreviated_message)
Seems pretty obviously generated by ChatGPT to me, because no speed-coder is going to define a fully-named function for such a purpose, and long, descriptive variable names and still submit in 50 seconds.
Another likely cheater, user RabilliGanesh , submitted hidden code in 49 seconds to a problem about determining whether an input integer value of note could be evenly changed for made from other integer-valued notes given in the input. Some cases required you to solve for situations where none of the notes evenly divided the number, but combinations could.
Seems likely that a 49 second solution is from cheating, to me, but itās not 100% sure.
Another likely cheater, user PondZA, submitted hidden code in 43 seconds to a problem of finding the area of a polygon given all itās coordinates (shoelace formula needed, but never mentioned in the text).
I know the formula nearly by-heart now, but it still took me 2.5 minutes to code up. I highly doubt any human coded the shoelace formula in 43 seconds (not to mention the time to read the problem).
New cheater āRigwildā has been completing hard problems in less than 20s. Has to be a bot. They were here last week completing them in just slightly faster than I could (Rank 500 in CoC) which was frustrating, but now it is back and not even trying to seem like they are not a bot. Please fix this, I love CoC! But I hate this bot Lol.
Yes I agree, Bot Rigwild is actively killing the CoC (seen a submission in 3 seconds). You can see on his profile that āheās cheatingā and on its github you found the codingame-bot. Lost some dozens of places in Coc
I reported Rigwild yesterday - a user that actually ADMITS to cheating on their profile. LetĀ“s hope the moderation team do something about it. Cheating doesnāt get more blantent than this.
Iāve completely switched to code golfing ever since I noticed that people are posting solutions with full comments and function names. My favorite case is when, for some simple task (in Python, taking around 45 seconds to solve the problem, I donāt remember exactly, something like Fibonacci), someone posted a solution in 30 seconds with full variable names in C++ xd.
Hello guys. Today I came across someone who was clearly using ChatGPT in CoC to win them. Should he/she be banned? One thing is using it after the other users have finished it, another is using it to win.
P.S.: The code even had comments like ā#Separating the letters in a wordā while he finished it in like 30 seconds. I mean, if you finish everything and also comment easy stuff like this in 30 sec shout-out to you but it seems obvious that it was AI. The profile of the guy is: Coding Games and Programming Challenges to Code Better
As a suggestion for improvement/punishing behavior:
The first person who solves a clash of code - must share their code (you cannot hide if you are first)
to have a button in CoC - āI consider itās cheatingā and if 60% of participants of the clash clicked on it => ignore this person result.
introduce public metric - ātrustā level for user profile (basically if trust if low - than if anyone consider this person cheater - it will void its result, and if trust is high - you need more people to āaccuseā). Also we need idea how to recover someone, if this person once cheated but 99% time he plays fair play.
Basically we cannot protect from occasional cheating, but if someone systematically cheating - i think it should be reflected on profile (not somewhere in forum :))
Sad to see that the CodinGame-sanctioned bot JayEm94 has now started using a cheaters solution for a new puzzle. (JayEm94 is one of the siteās own bots that uses code from past clashes as its own solution, but usually isnāt submitting code that would win).
Itās bad enough for players to be submitting AI-generated code, but when the siteās bots end up winning clashes by using some past playerās AI-generated code, itās even more annoyingā¦
Three others Iāve noticed also: Anon1976 Report
36 seconds for a challenge that was about the kās complement of n. Not simple, but if you know it well itās possible to do in 2-3 minutes. Didnāt share code
Voidyy Clash Submission
27 seconds to solve printing the index (1-indexed) and the size of the longest string of 1ās in a binary input string, space-separated. Long description talked of āhillsā, took at least 15 seconds to understand the problem even skim-reading (took me ~50 seconds to read). Seems unlikely to solve the problem in the remaining 12 seconds for 27 seconds totalā¦ Didnāt share code.
FIRST Clash
This player cheated using ChatGPT most likely.
Submitted a solution in 28 seconds to a clash that took at least 1 minute to read/understand and then at least 1 minute to code. Only ranked level 2 at the time of the clash (added this as an edit to this comment on 4-Feb-24)