The one who did this is a genius

№	Пользователь	Рейтинг
1	tourist	4009
2	jiangly	3823
3	Benq	3738
4	Radewoosh	3633
5	jqdai0815	3620
6	orzdevinwang	3529
7	ecnerwala	3446
8	Um_nik	3396
9	ksun48	3390
10	gamegame	3386

№	Пользователь	Вклад
1	cry	164
1	maomao90	164
3	Um_nik	163
4	atcoder_official	160
4	adamant	160
6	-is-this-fft-	158
7	awoo	157
8	TheScrasse	154
8	Dominater069	154
8	nor	154

Hi Codeforces,

We saw an unusual amount of hacks today. Specifically, there were a lot of hacks on problem B, I did one myself. The submission I hacked looked like this

Spoiler

t = int(input().strip())
for _ in range(t):
    n = int(input().strip())
    s = str(n)
    sz = len(s)
    
    sm = "01234"
    lr = "56789"
    if sm in s and lr in s:
        print("No")
        break
    
    flag = False
    total_sum = 0
    arr = [5, 6, 7, 8, 9]
    for i in range(sz - 1):
        total_sum += int(s[i])
        if s[i] == '0':
            flag = True
    
    if s[-1] == '9':
        print("No")
    elif total_sum == 35:
        print("No")
    elif s[0] != '1':
        print("No")
    elif sz % 10 == 0:
        print("No")
    elif flag:
        print("No")
    else:
        print("Yes")

As you can see, it's a completely valid solution except there are 2 unnecessary if statements which made it hackable. That made me think of 2 things: why there wasn't any 10-digit number in pretests and why the contestant wrote that "if"?

But after the contest I saw a list of hacks and, to my surprise, there were A LOT of hacks on similar submissions. So my hypothesis is that one of the authors intentionally spread an incorrect solution in some cheating community and removed all 10digit numbers from pretests. If that's true, I pay my respect

Комментарии (21)

Написать комментарий?

null_awe

5 месяцев назад, # |

+211

there was an intentionally wrong solution uploaded by youtube that we were aware of during contest, but we have no idea who posted it. we wouldn't spread a solution online even if it would eventually WA from hacks and FST.

as for the pretests, during preparation we saw no reason to have 10 digit numbers in it (there isn't a solution that would work for 10 digit numbers but not for 18 digit numbers — solution doesn't depend on number of digits save maybe small 1-3 digit numbers)

→ Ответить

Pranav16

5 месяцев назад, # ^ |

so I hope, you are going to use this fact, to catch cheaters this time

because, cheating is at it's peak these days.

||Please say Yes||

larush

+207

To whoever spread the WA solution: if you're ever lonely or sad, remember that I will always love you. You have my respect.

capricornharsh

← Rev. 2 →

+13

I second this; we need more like him!

Everule

+57

total_sum = 35 feels super suspect, by just glancing at the code. But sz % 10 is really funny. I didn't even notice it at first, i thought it was just that other condition.

djm03178

+32

I looked through some of the other solutions that failed system test, and it looks like 35 comes from adding up all 'large digits' (5 to 9), and some people just calculated it themselves (like how arr exists but isn't actually used in that code). It's still suspicious, though.

What makes me scared is that some of them even solved up to problem D and they would've been above my rank if these solutions didn't fail. If this kind of thing has happened casually, I just can't imagine how many ranks I've been losing just because I've always participated in contests conscientiously.

Petr

+95

Given the unused variable and seemingly random checks, could this solution be the output of ChatGPT or a similar tool?

bgopc

not the exact output in my opinion, but chatgpt may be used to edit the code and make it weird, but it's a common act of cheaters to spam add function + random variables + random conditions, this causes the plag checker which uses AST to check the similarity of codes executing not be able to identify it since because e.g. in a loop they updated 10 variables which just makes the checker not be able to identify it

GroupMatrix

+78

What's even more ridiculous is 264905546, a similar code.

Can you find out what's the difference?

if (s.find(small) != string::npos && s.find(large) != string::npos) {cout << "N0" << endl;return;}

Yes, it's N0.

ZhockDen

-17

Not Only this.. There are many similar ones..

I'm sorry, but it seems thatnone of these codes have N0 instead of NO.

Oh!! Sorry I misunderstood..

shevlopmes

That's hilarious)

1moreredcoder

huge respect for authors /\

Mindeveloped

I respect this behaviour but, is there a possibility that, somebody would read the code and recognize the solution, and then implement the solution completely on his own way to escape plag check?

i dont think plag check is working anymore , they convert the code into another language , even they discard the function from code and mix with main code by chatgpt .