Problem 728

[endagorion]

So I would have expected a lot of solves now, also I'm pretty sure my answer is correct. Does the statement have the correct $N$, and system the correct answer?

[bobby_eletrica]

Have you been able to get the right values for their examples??
Under the assumption that exactly k (instead of over k) consecutive coins must be turned each go I calculated (brute force), $F(8,3)=0$, $F(9,3)=384$, $S(3)=12$, $S(10)=7990$.

[heteroing]

A friend and I are also getting correctly matching values for the test values, and the (same) incorrect value for 10^7, using at least three different programs.

[heteroing]

Have you been able to get the right values for their examples??
Under the assumption that exactly k (instead of over k) consecutive coins must be turned each go I calculated (brute force), $F(8,3)=0$, $F(9,3)=384$, $S(3)=12$, $S(10)=7990$.

Also the image included in the problem description clearly shows $F(8, 3) \geq 1$.

[bobby_eletrica]

probably you noticed I was counting unsolvable configurations

[aleksey]

Same problem here, test values match, the result is incorrect.

[user202729]

Same problem for me.

(random note: I tried submitting the value of S(10^6), S(10^8), S(2e7), S(5e7) and S(1e7) mod (10^8+7) without success. It's wrong to try random options like this, but at least if it gets accepted I can definitely prove that the result is likely wrong)


Unfortunately, does it mean that the gold medal (first place) for this problem will be meaningless (if the answer is simply updated and multiple people who solved it try to submit it at once)?

Unless Project Euler stores all the incorrect submissions (+their timestamp) for all users, and determine the fastest solver based on that.

[jaehyunp]

I'm also pretty sure that my solution is correct and yet getting wrong answers. I tried a few alternative values but didn't pass: S(1)+S(2)+...+S(10^7), S(11111111), S(12345678), ...

[utomaya]

same here...

I got S(3)=22, S(10)=10444, and S(1000)=853837042 (mod 10^9+7).
But my result for S(10^7) was rejected.

For such a simple problem, it is strange that there are no solvers til now.
[edit]:
This situation happened before.
http://projecteuler.chat/viewtopic.php? ... 9d0ea9cc27

[Mohammad_Ali_B]

Precisely the same problem here.

I get the correct value for S(3), S(10), S(1000), and I'm sure of my answer for S(10^7).

I too, tried many variations.
I even tried forcing overflows and including common bugs in my code, to see if I could match the wrong answer.

[shs.10978]

Same here... I was surprised when my answer was not accepted, and even more surprised when no one has got it correct after four hours.

[fakesson]

Same here. Downloaded the webpage and scripts and tried to see if I could find anything that appears wrong, but it looks the same as for other unsolved problem.

[rkm0959]

Darn. I was so excited to get a first solve despite being 3 hours late to the party only to find this out

[user202729]

Downloaded the webpage and scripts and tried to see if I could find anything that appears wrong

Did you know that it's possible to view the HTML source of the problem? https://projecteuler.net/minimal=728

The only thing that "appears wrong" to me is that the last sentence doesn't end with a period.

Find $S(10^7)$. Give your answer modulo $1\,000\,000\,007$

At that time I thought that it's just sloppy writing, I can't imagine that something meaningful can be put after that sentence (in base $16$? Wrapped in quotes? With spaces to separate the thousands? $+1$?)

[Mohammad_Ali_B]

I was also very excited to get a First Solve. Something seemed really off, because honestly, I had implemented $F(n, k)$ quite quickly (it took about 1 minute to figure it out). A fast solution took me about 30 minutes overall.

Anyways, I've tried far too many variations of the inputs for one night. I'm gonna come back in 8 hours hoping that my first "incorrect" submission from about 3 hours ago counted for something

[neverforget]

I also found a solution that works for the test cases but gets WA. For $S(10^8)$ I got something ending in ???, using same modulus. Is it at all possible there is a corner case we all didn't consider that first shows up between $10^3$ and $10^7$?

[Mohammad_Ali_B]

728 is 222222 in base 3, so I tried submitting the answer in base 3, and this failed.

[philiplu]

I also found a solution that works for the test cases but gets WA. For $S(10^8)$ I got something ending in ???, using same modulus. Is it at all possible there is a corner case we all didn't consider that first shows up between $10^3$ and $10^7$?

I got the same last 3 digits as you for $S(10^8) \bmod{10^9+7}$ (blocked out in case you remove your value later). Given the simplicity of the formula we probably all found for $F(n,k)$, and the fact that $S(10^3)$ is easy to verify, I'm thinking that, no, we're not missing some weird corner case that only shows up above $10^3$. It's probably a glitch in the submission system, and since it's just now getting on to 9 AM in the UK, presumably Euler or someone else will notice the problem soon.

[hk]

The answer has been corrected.
Apologies for the inconvenience.

[philiplu]

Thanks, hk.

So tempting to jump in now and finally get that top 10 Chart Topper award I've never managed, but, nah, it wouldn't feel right.

I held out for lucky #13 instead. That's still probably my best showing ever, tainted as it is.